GiveWell announced the largest grant in its history on 24th August, contributing $276m to the Against Malaria Foundation (AMF). This funds the distribution of 92 million bednets in the DRC over three years, and is modelled to avert 67,000 deaths.

It’s worth sitting with those numbers for a second and feeling very grateful for them. This is a staggeringly impactful grant that should prevent thousands of children from dying unnecessarily.

What does it mean for donors?

Subscribe now

How should I think about GiveWell’s grant to AMF?

Critically, this grant does not imply that AMF cannot deploy more funding effectively. The grant is for three years, committed up front, and is limited to a single country.

Indeed, it’s a useful example of how to think about increased funding during the Funding Anthropalypse. We often fall into a lazy shorthand of describing a charity or project as ‘full’. But ‘full’ actually makes no sense in this context, because almost all projects could absorb more money in some fashion.

A better model is to think of investments like this as setting a ‘price’. In this case, AMF’s work in the DRC is now fully funded at a price of roughly 16x GiveWell’s benchmark.

Does that mean AMF’s next project prices below 16x? Maybe - but even at, say, 12x, it would still be double GiveWell’s current funding bar. Furthermore, nothing in the announcement suggests AMF is out of opportunities at this level. GiveWell’s grant only concerned one particularly high-burden country, and the announcement emphasised that bednets remain “one of the most cost-effective ways we’ve found to save lives”. This grant itself shows the need was real: before it, GiveWell believed half those 92 million nets would not have been funded by anyone.

Accordingly, AMF remains a tremendous donation option, at least until we see concrete analysis to the contrary.

Should you still give to GiveWell’s Top Charities?

Yes!

First, a substantial part of this grant ($57.9m) was funded by donations to the GiveWell Top Charities Fund. It’s a great example of the extreme cost-effectiveness that donors to these charities and this fund can achieve.

Second, by definition, this means that ~$218m came from GiveWell’s other funds and Coefficient Giving’s co-funding. This shows a second real value of donating - you allow GiveWell (and Coefficient Giving) to use discretionary funds for other strong grants. At present, this will on average mean a grant at 6x GiveWell’s benchmark, which is still an extremely competitive price point.

Accordingly, for most donors, this grant alone doesn’t imply that you should shift your giving.

I’m giving millions in global health. What should I do now?

In short, it remains the case that donations to GiveWell will be spent very effectively, but it might take longer for them to be deployed.

The real variable here is time - when you can expect your money to be regranted. This usually affects larger donors more, because they are more likely to have a spending plan for their funds, giving a set amount each year. This means that they may prefer to give somewhere where money will be deployed quickly, in order to see immediate impact from each annual grant cycle.

Even with its recent influx of funding, GiveWell will continue to find outstanding donation opportunities. Indeed, as they only research grants when they expect to have the cash to cover them, a grant to GiveWell may enable them to find even more exciting projects.

However, it would be reasonable for donors to monitor GiveWell’s agreed grants versus its receipts. Following Coefficient Giving’s $1bn commitment, GiveWell should now be in a position to assess even very large grant opportunities (as this grant signifies) and has a substantial amount of funds to grant in the near term.

As it says itself, it would be reasonable to pause donations to GiveWell “if you want to see if we successfully draw down our rollover funding over time”.

One way to think about this is that dollars you donate could have one of three states. They could be uncommitted - that is, currently held back in expectation of greater future impact; they can be committed for future delivery, as is the case for some portion of this three-year AMF grant; or they may be deployed already into projects.

In practice, this grant only matters for donors who mind their money sitting uncommitted, or want it deployed fast.

Where else can I give cost-effectively in global health and development?

If the best global health options GiveWell can find are being funded with nine-figure grants, I expect casting a wider net to present significant opportunities for impact.

Finding these opportunities is very difficult, though. You either need to build the capacity to do so, or buy it. Buying it is faster and easier, especially when top-quality research talent is hard to find, and you can do that via a specialist advisory or a pooled fund. (Conflict of interest: I manage a fund and run an advisory, so discount accordingly.)

When I started Ultra Philanthropy, I didn’t build in-house research capacity. I assumed I could stand on the shoulders of GiveWell and other evaluators indefinitely. The AI windfall is making that assumption redundant.

GiveWell’s funding bar has also dropped and may drop further, so the ‘price’ of following only GiveWell’s recommendations has changed. All else being equal, this increases the chances of donors being able to find better cost-effectiveness elsewhere, particularly if they are open to different cause areas or stages of funding.

For one thing, it seems highly likely that the canonical Effective Altruism cause areas are not the only cost-effective places to give. Donors can surely make outstanding grants in areas other than just reducing under-5 mortality, animal welfare, climate, existential risk and meta donations.

In global health and development alone, my shortlist includes the following, and several are areas where we have already recommended excellent grants to donors:

1. Education in LMICs - supported by a robust evidence base from the Global Education Evidence Advisory Panel and catalytic funding from places like Prevail Fund and Livelihood Impact Fund, and an obvious complement to reducing under-5 mortality.

2. Gender-based violence reduction - Dr Akhil Bansal, a medical doctor and fellow at the Center for Global Development, laid the groundwork for this some time ago and it is already competitive, and more so as funding bars fall. It’s also an incredibly pervasive problem, which many donors have experienced first-hand, and so could attract significant co-funding. (Conflict of interest: I am chair of trustees at High Impact Athletes, which hosts the 1in3 Fund in this space.)

3. New funds focussed on a single theme, such as family planning or cataract surgery - Elevate Philanthropy identified a number of promising funds based on a single issue, which could do more field-building and shaping than general grantmakers, and also crowd in a lot of new funding.

4. Funding that tolerates different levels of risk, such as earlier-stage or hits-based giving - I wrote about this here.

5. In general, things focussed on improving wellbeing, reducing suffering or improving livelihoods, rather than predominantly preventing deaths. This seems like an obvious complement to GiveWell’s work and there are lots of options, from mental health support to economic growth.

This is without even considering opportunities in entirely new areas for effective giving, such as peacebuilding or the nascent work in democracy.

What to watch out for, following GiveWell’s AMF grant

This grant doesn’t concretely change the options for donors. However, it does indicate the direction of travel, and large donors should monitor a few key indicators to update their plans. (Actually, the $1bn grant from Coefficient Giving already signalled the new direction, but this is further evidence of how effective giving is changing.)

I will be monitoring the following for clients:

• How much GiveWell ultimately deploys in 2026, versus receipts, and whether their rollover funds are growing or shrinking.

• Further large grants to Top Charities, and what this implies about each individual project’s ‘impact price’.

• Whether GiveWell’s funding bar moves again, signalling that the marginal effectiveness of their grantmaking is falling as they fund more things with more money.

• Whether Coefficient Giving or other donors make further huge contributions to GiveWell or its recommendations.

On this last point - Coefficient Giving called its $1bn commitment a “one-off”. The Funding Anthropalypse could make these sorts of gifts the new normal. If that happens, this stops being theoretical: every serious donor will need a view on what their money buys, and when. Money deployed now and money committed for later both do enormous good. Just choose on purpose.

If you’d like help refreshing your strategy, get in touch.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and is the fund manager of its Mid-Stage Global Health Fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

I used Claude to help structure my thoughts and to suggest improvements and flag gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

I helped hundreds of unemployed people with their job search as COO of School of Hard Knocks.

We had a training module called “the difference between a dream and a plan”.

A dream is saying you want to run a bar in Spain some day. A plan is learning Spanish, understanding the licensing laws, saving up a deposit and scouting locations. Many attendees had the former and almost none had the latter.

I’m reminded of this by global health ‘scaling plans’. Ultra Philanthropy’s Mid-Stage Global Health Fund, which I manage, aims to bring a cost-effective health intervention to scale by 2029, so I assess a lot of them.

We usually talk about scale as a question of ambition and money. These organisations have plenty of ambition, and I expect the Funding Anthropalypse will soon give them enough money.

What’s missing is a proper plan to scale - one with professional management and leadership in it. We need more Managing Directors and Chiefs of Staff.

Subscribe now

What’s wrong with most scaling models?

Scaling models often sell funders a dream. A typical one looks like this:

1. 2026: team of 15 - 2 co-founders, an ops associate and 12 frontline delivery staff in-country. About as cost-effective as a GiveWell Top Charity.

2. 2027: team of 29 - 2 co-founders, an ops associate, a Director of Operations and 25 frontline staff. A bit less cost-effective, but still above the bar.

3. 2031: team of 106 - 2 co-founders, a Director of Operations, an ops associate, a finance associate and a Head of Monitoring, Evaluation and Learning (MEL), plus 100 frontline staff. Cost-effectiveness will be 10x a GiveWell Top Charity.

4. 2036: team of 314 - 2 co-founders, a Director of Operations, a Director of Finance, a Director of MEL and 3 juniors per director, plus 300 frontline staff. Cost-effectiveness will be 150x a GiveWell Top Charity.

This sets off alarm bells.

First, the GiveWell Top Charities are incredibly hard to beat. Any model predicting you’ll beat them by two orders of magnitude is probably overstating your effectiveness or understating your real costs (in my experience, usually both).

Second, I’ve seen many, many organisations whose plan reads ‘... infinity’, almost always 5+ years out. I am always suspicious of a plan where cost-effectiveness dramatically increases at a point far enough in the future that no one will check.

When those deadlines arrive, very few have hit their goal. Usually, this is because consistent delivery, team coordination and accurate monitoring all get harder as you expand, and so you need to add more bodies (and therefore more costs). There are very few roles that allow one person to oversee far more goods or services without being overburdened.

I think this copies the tech scaling model, where a single repeatable product can serve hundreds, then thousands, then millions of people. Healthcare almost never works like this. A software product can serve its millionth user with the same codebase, but you can’t serve a million people without more nurses.

Third, these models assume the co-founders can manage a much bigger staff, usually while doing their own work and most of the fundraising and strategy. The founders become the bottleneck, lacking either the time or the talent to do everything expected of them.

If you’re a donor, treat any plan that reaches huge scale without expanding its back-office staff with great scepticism. A scaling plan without new senior hires in it is a dream with a budget attached.

Instead, we want to see reasonable hiring triggers, a clear idea of the founders’ roles and a thought-through future org chart.

What should charity founders do day-to-day?

Founders have several jobs no one else can do - defining the vision and strategy, and securing the money and people to make it happen.

But the cost-effectiveness world pushes us to keep overheads as low as possible. So co-founders, often smart generalists, often end up ‘teaching themselves’ the next thing that needs doing: bookkeeping, MEL, designing their randomised controlled trial (RCT).

This is usually a false economy. You can do the basics, but the second you need to add a million contraceptive doses to your accounting inventory, you’re frantically asking Claude and hoping its advice is kosher.

Autodidact founders scaling back-office functions is usually a mistake. It costs their time, produces errors and takes them away from work that only a founder can do.

Rule of thumb: if it’s a profession and not your core skillset, buy the expertise, from a consultant, a firm or a new hire.

This shows up most often in finance, operations, HR, MEL - but also, the most neglected of all, people management.

What do Managing Directors and Chiefs of Staff actually do?

Running organisations and teams is a profession.

I’ve built staff teams twice, at School of Hard Knocks and One for the World, so I can say from experience: managing people well is incredibly difficult. It was consistently the hardest and most time-consuming part of my work.

A randomised trial in Indian textile firms also found that adopting structured management practices raised productivity by 17% in the first year. These were for-profit manufacturers, so adjust accordingly, but also see Nick Allardice’s writing for sector-relevant material.

The title varies - Managing Director, Chief of Staff, COO, Executive Director, CEO, even Founders’ Associate - but the impact is the same: it pays to hire someone to make the organisation function.

Managing Directors run the project, directing people and systems, so the founder can concentrate on strategy, fundraising and externally representing the organisation.

Chiefs of Staff focus on delivery, keeping people on the right priorities and on schedule. They specialise in seeing a plan through from conception to completion.

Most founders aren’t actually selected for their capacity to run a team or direct people’s work. They excel at articulating a vision and convincing people to back it.

So why don’t more founders hire these roles?

One reason is status. Founders want to stay at the top of the org chart. When I see a founder with 10 direct reports, it’s often because they want to remain ‘above’ all the directors and heads of functions.

That doesn’t mean they should direct everyone’s work personally. Managing Directors and Chiefs of Staff let founders keep their status without spending most of their time on line management. Hiring an MD is a ‘silent promotion’ - it frees founders to do the strategy, vision and fundraising that only they can do.

I’ve been that hire twice. School of Hard Knocks’ founder took me on as their first employee; One for the World’s volunteer founders hired me as the first full-time Executive Director while they stayed on the board, and we grew revenue 7x in 4 years. Both times, the founders kept the vision and the status, and the organisations grew because someone was paid to run them.

As a founder, ask yourself: “How will this project run when I only personally know one third of the staff?” It might seem fanciful, but it’s entirely plausible that every grantee of our Mid-Stage Fund gets to $10m revenue in the next 3 years. At that point, the founders will have hired only a fraction of the team and will probably see the whole organisation together once or twice a year. This is a significant organisational change.

A second reason is fear of bureaucracy. Tech readers will hear this as ‘manager mode’, the thing ‘founder mode’ warns against. But that failure belongs to product companies; delivery charities fail the other way, with the vision intact but the service late or substandard.

How should donors think about overheads?

The biggest blocker may be the optics of overheads. We built this problem as donors. For more than 15 years, the sector has being trying to educate donors against using low overheads as a proxy for cost-effectiveness, mostly unsuccessfully.

This is why I focus on cost-effectiveness in my own advising. A low overhead rate only tells you about costs. Cost-effectiveness assesses costs and outcomes together. If squeezing overheads also squeezes what the charity delivers - and, past a point, it usually does - the saving is false.

Most cost-effective organisations only make sense at scale. Their unit economics may not even work while small. The goal is to grow until they can compete with the most cost-effective charities in the world.

Nothing can scale without investing in its infrastructure.

An organisation may not have the budget or scale to hire a senior executor yet, but it can still set triggers for when it should. Triggers protect against both failure modes: hiring too early and overloading the budget, or overstretching the current team and hiring only when something breaks.

I’d go further: scale can solve your overhead rate, but a low overhead rate can’t take you to scale. Imagine a senior Managing Director at US$200,000 per year. At $1m revenue, that’s 20% of the budget. At $10m, it’s 2%, and no one minds any more.

But cap your overheads too brutally at $1m (or, especially, $3m-$5m revenue) and you’ll probably never reach the scale the project was set up to achieve. You’ll end up in one of the many valleys of death: burnt-out staff, failure to fundraise enough, or a serious implementation error because you couldn’t properly monitor the frontline work. If you’re really unlucky, you might discover that error during your first RCT, and disappointing findings will dog you for years.

What should we do to help organisations scale effectively?

For donors: scrutinise your grantees’ plans to scale (or hire an advisor - talk to us for a free recommendation). Ask for the org chart in 3 years’ time and what the founders aim to be doing. Look out for anyone managing more than 8 direct reports (ideally no more than 5).

Be deeply sceptical of any organisation that says it can scale dramatically with a minimalist leadership and back-office team. (Be especially sceptical of anyone claiming their cost-effectiveness will go from ‘competitive’ to ‘orders of magnitude better than New Incentives’.)

Most importantly, make it explicit that you’ll pay for overhead - and that you’re less likely to fund organisations that don’t plan for it.

For founders: turn your scaling dream into a serious plan. Set hiring triggers in advance and map your operations, finance, HR and MEL needs as you grow. Budget a line for an MD, a Chief of Staff or both before you need it. Then get someone who’s scaled something to red-team the plan.

For talented people: I hope many more of these roles are hired in the next year. If you’re organised, ruthless at delivering and can manage people, put yourself forward. A great first step is joining High Impact Professionals’ talent directory, so organisations can find you. (They also have an Impact Accelerator Program starting soon.) I think a lot of $1m organisations are about to become $10m organisations, and you can help them do it.

Sources of support

A quick non-exhaustive list:

• For advisors and evaluators who can vet scaling organisations, ask us for a recommendation or consider a scaling-focused fund (e.g. DIV, ours).

• For organisations, Mulago Foundation knows a lot about scaling in development. I’ve also seen useful, thoughtful materials from Spring Impact, though I don’t know their work well enough to endorse it.

• For prospective hires, High Impact Professionals and Ambitious Impact are great entry points for talent.

The Funding Anthropalypse can buy programmes at ten times the scale. The cheapest thing it buys is the people who turn those dreams into plans.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and is the fund manager of its Mid-Stage Global Health Fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

I used Claude to help structure my thoughts and to suggest improvements and flag gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

Most donors never find out whether their grants worked.

They model the impact before the money goes out, but don’t check if the models were accurate. Discerning impact is hard, but monitoring grant progress is simple: agree indicators in writing before you send the money, ask the grantee to put a probability on each, and score them once a year.

Charities usually send grant reports anyway. Unless you ask for something specific, they rarely contain the most useful information.

A recent report to a client touted the project’s success: it had made 20 policy recommendations. When pressed, the grantee responded that only 20% had been implemented, even partially.

A 20% implementation rate may or may not be a good outcome. Either way, it wasn’t the outcome they reported, or what they were asked to report.

Ironically, this is pervasive in ‘evidence-based giving’.

How much do donors know about what their grants achieved?

When it comes to individual donors, often nothing.

‘Evidence-based giving’ typically applies before the grant. We are very good at modelling what a grant might achieve. We’re surprisingly bad at checking what it did.

This varies by sector. Frontline global health interventions routinely monitor outputs - clinic visits, vaccines distributed - and typically push further to monitor outcomes and impact, because the whole sector expects it.

In harder-to-measure areas, with no strong standard for self-monitoring, it’s the Wild West. I see policy, advocacy and governance interventions struggle the most.

Berlin-based think tank Future Matters recently found that 13 out of 22 major earning-to-give donors couldn’t tell whether AI governance work achieves anything, including four out of six who work at frontier AI labs.

If even the people closest to the technology aren’t sure what their grants are achieving, what hope do the rest of us have?

Future Matters blames legibility. I’d go further: illegibility is a choice funders make, by not asking the right questions at the right time.

The effective giving space currently inverts best practice: monitoring is often most rigorous where uncertainty is lowest (bednets), and often absent where it’s highest (policy and advocacy).

Why does monitoring and evaluation of donations matter?

Monitoring, evaluation and learning (MEL) is how we understand what is working, to improve both projects and funders.

A useful MEL plan starts with its purpose, and what questions it seeks to answer.

As a project, you might want to know whether your services are effective and reach the right people; how to improve them; or what you can use to fundraise.

As a funder, you have a stake in whether your individual grantees are working, but you should also be interested in improving your funding decisions; managing risk across your portfolio; and making a meaningful difference to the problem (even more than a single project, as you hold a variety of bets). You might also want to share lessons with the wider ecosystem or even build the field.

Defining which of these questions matters focuses our activities and makes our MEL plans useful.

MEL is the only way to answer these questions systematically. Without it, we risk wasting funding and time. In the absence of evidence, projects can go astray and donors can withhold grants. Consequently, we will struggle to solve the life-or-death problems we are working on.

Why don’t donors ask for more reporting from their grantees?

One motivation is noble: not overburdening the project. This has been a positive shift away from intensive, performative reporting that almost nobody read anyway.

This change has been driven by two very different camps: cost-effective giving at one end and trust-based philanthropy at the other.

The problem is that each risks having the right diagnosis and the wrong cure. The fact that traditional reporting is too burdensome and not useful doesn’t mean that you shouldn’t ask anything at all.

Any strong organisation should actively want to know whether its work is effective. If it is happy to run purely on anecdotes and intuition, that’s a red flag.

Is policy work too hard to measure?

A second reason to forgo monitoring and evaluation is the belief that some interventions are too hard to measure.

This is half right. Evaluating the impact of policy work is extremely difficult. You can’t usually run RCTs, so counterfactuals are tricky. Results can take years to arrive. Several groups will usually work on the same issue, without a clean way to assign the credit. As Future Matters points out, publicly claiming attribution can burn your most important relationships.

This is particularly acute in AI safety and governance, where it isn’t even obvious what makes a good regulation. If reasonable people disagree about whether slowing down AI development is net positive, no monitoring and evaluation plan will solve that. It’s just the risk you take by funding this work.

There are nonetheless proxy indicators of effectiveness. For example, bills sometimes use language drawn directly from a project’s publications; lawmakers sometimes acknowledge conversations with particular groups; and it can be surprisingly effective to ring congressional staff and ask how influential an advocacy project was.

In addition, monitoring policy work is really not that complex, even when impact measurement is hard. Any MEL professional would find it routine: ask an organisation what it plans to do and how likely it is to succeed, and then track whether it did.

How can I monitor my policy and advocacy donations?

A strong monitoring plan includes a handful of indicators, with the project assigning each a probability of success. These can be scored once a year, with any failures left in. It should take the project around half a day to complete and track things it wants to know anyway.

My advisory, Ultra Philanthropy, has an in-house MEL specialist with 12 years’ experience advising development programmes. We have worked successfully with some admirably responsive, transparent organisations to produce these sorts of plans, even in thorny areas of advocacy.

For example, we worked with a climate advocacy group in 2025 to define the following:

1. Defend climate provisions in reconciliation and agency reorganisation

   1. Protect Loan Programs Office credit subsidy at Department of Energy and defend other DOE funding - 25% [chance of success]

   2. Protect advanced geothermal and nuclear access to tax credit through at least 2029 with viable Foreign Entities of Concern restrictions - 60%

2. Achieve trade legislative objectives

   1. PROVE IT Act on emissions intensity to pass the House (20%) and Senate (30%) in 119th Congress

   2. Pass reworked carbon border adjustment mechanism in the House (20%) and Senate (25%) during 120th Congress

3. Success in Energy Acts

   1. Pass an Energy Act with meaningful regulatory reform in 2027 (House, 15%; Senate, 30%) or 2029 (House, 20%; Senate, 30%)

4. Success in permitting reform

   1. Pass bipartisan bill in 119th Congress (65% of slim or full package passing)

   2. If 4a fails, groundwork laid to pass slim or full package in 120th Congress (75%, provided Democrats control the House)

It would have been very easy for the project to focus on controllable activities (‘we’ll have 50 meetings with lawmakers’), or to claim that their work was powerful magic that defied tracking.

Instead, its plan gave our client (and the organisation itself) a clear, measurable monitoring framework. Even though the impact of these legislative achievements remains uncertain (laws still have to be enforced, might have unintended consequences, take a long time to have an effect, etc.), it’s easy to see whether or not they moved the ball down the field.

Unfortunately, this clarity is extremely rare in the advocacy and policy work that I encounter.

Why ask grantees to put a probability on each target?

One aspect of our monitoring that goes beyond the norm is assigning probabilities to each output.

One benefit is that this exposes overconfident fundraisers. Often, fundraising teams send donors several impressive milestones that a grant could enable. Ask their programme team to put a success percentage against each, though, and it’s clear how likely you are to see them.

It also prevents sandbagging. If a project suggests indicators that it has a 95% chance of achieving, we look more closely at its ambition and risk-tolerance, and check whether those outcomes are actually meaningful.

Sometimes success is both highly likely and meaningful - ‘we’re 85% confident of reaching 10,000 people with oral rehydration solution this year’. But this can also be playing it safe, or not focusing on the most important changes.

Wielded together, these opposing concerns create something exciting: a portfolio that blends solid gains with stretching, ambitious goals, which tell the donor and the project how close they are to getting the hard thing done.

Donors then need to focus on calibration, not merely success rate. An ambitious miss is usually more useful than a 100% hit rate. It’s essential if we are going to embrace more risk in our giving.

What should I agree with a grantee before I send the money?

For most donors, it makes sense to engage a MEL consultant or an advisory with in-house expertise to do this work, since it usually involves negotiating with grantees, and a reasonable amount of data collection and analysis across a portfolio.

For donors wanting to implement this themselves, however, there are some things to watch out for:

1. Start by defining your purpose and related questions. If you know what you want to learn and what you want to do with the information, you increase your chances of asking the right questions, focusing on what is useful over what you might find interesting.

2. Make sure to agree targets in writing before sending the money. It is much harder (and sometimes unfair) to negotiate the goals of a grant after sending the funds.

3. Don’t accept activities and outputs dressed as outcomes. Activities are things a project does (write policy briefs, convene conferences). Outputs are things produced by the activities (number of policy briefs, number of conferences). Neither is meaningful unless it achieves an outcome, i.e. a positive change (such a bill being passed into law). It is extremely common for projects to substitute activities and outputs for outcomes - don’t let them.

4. Agree a suitable range of targets. One way to look great is to list many possible outcomes and then report only the one or two that landed. A written framework makes projects focus on things they can actually affect, and acknowledge the things they tried that didn’t work out. Hits-based giving is fine - you might only need one success in eight - as long as both sides knew that going in.

5. Include confidence levels. This prevents overclaiming and sandbagging, and lets you gauge how ambitious the project is, and how much risk you have across your whole portfolio.

6. Take refusal as a signal. If the project won’t agree to ambitious and clearly defined indicators (or, even worse, tries to explain why this doesn’t apply to its work because it’s unique), that’s a huge red flag.

7. Make decisions based on the results. Look at the monitoring data and update your beliefs when making future decisions. It’s perfectly fine to continue funding difficult work, even if it missed every objective - presumably most successful vaccines came after tens or even hundreds of failed candidates. What’s harmful is to burden projects with reporting and then not to use it to improve your grantmaking.

I’m expecting an AI windfall. What should I do?

Establishing a consistent process to monitor your grants and learn from them is essential if you are about to become a significant philanthropist.

The Funding Anthropalypse will create many large donors who have little or no experience in grantmaking or MEL. This makes feedback loops a high priority, so that you can improve your grants, risk-tolerance and general understanding over time.

You may also give a significant amount to AI safety and governance work, which is much more uncertain and harder to measure than many cause areas, with greater risk of accidental harm. This makes effective monitoring even more important.

Use the pointers above to shape your monitoring strategy, and talk to a professional if you can. You can also defer to evaluators, advisories or funds you trust to do this work for you.

Julian Hazell (formerly Coefficient Giving, now Astralis Foundation) estimates that only 30-60 people investigate and recommend AI safety grants full-time.

I know several of them, and I’m happy to make an introduction - just drop me a line.

Illegibility is a choice. Make a different one.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and is the fund manager of its Mid-Stage Global Health Fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

Thanks to Olivia Kaye and Samuel Verbi for feedback on the draft. I used Claude to help structure my thoughts and to suggest improvements and flag gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

“What is the most common mistake donors make?”

This is one of the questions I get asked most as a donation adviser. Often, I say that it’s risk tolerance.

Many donors have risk backwards. They hold the money they plan to give at close to maximum risk, often in a single company’s stock, then give it away at close to minimum risk, backing only the most heavily evidenced organisations.

Getting it the right way round means working at three levels: de-risk the wealth that funds your giving, diversify your grant portfolio, and take more risk inside it.

What do most donors get wrong about risk?

Candidly - almost everything. The majority of individual donors are distributing risk carelessly across multiple axes.

Take the distribution of risk between wealth and grantmaking, for example. Many donors are taking extremely large risks with their capital, whilst only giving to massive organisations backed by 24 randomised controlled trials (RCTs).

That’s two risk decisions taken with the same money, sometimes without noticing either.

AI equity holders are a clear case. Many are committed to the most rigorously evaluated forms of giving, but hold their wealth in a single stock.

More than $1trn came off semiconductor and AI hardware stocks in July.

If your wealth sits in one AI company - public or private - the market that prices these companies just moved a long way. You should therefore be thinking actively about how much risk you’re willing to take with these funds, and, as I’ll argue, consider demanding less certainty before giving them away.

How much risk should I take with my charitable giving?

In short: more risk than you (probably) take now at the grant level, and less with the wealth itself.

Most donors should give their core funds to low-risk, evidence-backed giving; commit a genuinely risky bucket to funds that make enough grants to survive a high failure rate; and keep a small slice for direct bets. As a starting point, this could be 70% to the core, 20-30% through risk-tolerant funds and 0-10% directly to moonshots.

On the wealth side, Anthropic staff had the chance to sell some of their shares in April. However, buyers were left unable to purchase their target amounts, as staff chose to hold on to their equity. In this instance, the bet has paid off handsomely - the company has since been valued at $965bn post-money.

This is, however, an extremely risk-tolerant approach. Many staff at frontier labs hold most of their wealth in the undiversified volatile stock of a single company.

Single stocks are a wild ride. Hendrik Bessembinder studied every US common stock since 1926 - around 26,000 of them. Most returned less over their lifetime than one-month Treasury bills, and the single most common lifetime outcome was a complete loss. Just 4% of companies account for the entire net gain of the US stock market in the same time period.

Anthropic may well be one of the exceptions that does extremely well. But you don’t need a failure to get hurt - as July showed, concentration means the money you’ve earmarked for giving could move a long way in a few days.

Much of this concentration isn’t chosen - lock-ups and sale windows can put constraints on donors. At the margin, however, staff are choosing: whether to sell into a tender, how much to diversify when they can, and when to start giving.

What would happen, meanwhile, if a donor took these sorts of risks in their grantmaking? In my experience, most would quit after a year.

This isn’t because donors don’t understand variance - many donors built their money through judicious risk-taking, either in companies they founded or by directly trading volatile assets. Instead, I think it’s because a failed grant hurts more and is more embarrassing than a failed investment. It can be very difficult to explain that a grant did not succeed, whether to your spouse, your co-founders, the friend who recommended the nonprofit or to yourself. The reputational and emotional cost often outweighs the financial loss.

In addition, donors don’t swiftly cut their losses. Venture funds drop failing bets quickly, guided by clear pricing signals. Donors don’t get any signal without robust grant monitoring - and they don’t want to admit a mistake or let down a favoured nonprofit anyway. So the failing grant keeps its funding.

It’s less ‘costly’ overall to keep giving money to a charity that isn’t succeeding than it is to change direction. Accordingly, most donors act in a way that is inconsistent with rational risk tolerance.

How should I spread risk across my giving?

Donor confusion often starts with conflating different types of risk.

Risk at the grant level can sit in at least three different places:

1. Intermediary risk. Using an intermediary (a grantmaker, evaluator or fund) that has a limited track record, even if it works in an established area.

2. Project risk. Backing an unproven intervention, even if you or your intermediary is an established grantmaker.

3. Team risk. Backing people who haven’t executed before, even if the thesis of the intervention is solid.

Risk increases as you go down this list. You can therefore manage your exposure by making grants that are spread between these tiers, or even avoiding some of them altogether.

For example, you could separate a chunk of your capital for risk-tolerant giving, and then put most of it into backing a couple of new-ish funds in a well-established area, where newer teams can lean on best practice from established peers. (Conflict of interest disclosure: I run Ultra Philanthropy’s Mid-Stage Global Health Fund, exactly this kind of vehicle, so discount appropriately. I can confidently recommend several others without conflict, though, like DIV Fund, the EPIC Air Quality Fund and D-Prize.)

You could then use the final, say, 10% of your giving to back a range of experienced operators trying new moonshots - maximally risky grants, not diversified at all at the grant level and with little information to predict the outcomes, but at least following a tried-and-tested venture model. Coefficient Giving’s hits-based giving rule is to back people with excellent records, even when the approach itself is unproven.

When I talk to donors about their approach, they often say that they have extremely high risk tolerance in their giving. They then say that they made fewer than ten grants in the past year, all directly to frontline projects.

These things are incompatible. Roughly two-thirds of venture financings return less than the capital invested, and only about 4% return more than 10x, which seem like reasonable proxies for grantmaking. If you make a single-digit number of grants per year, at the failure rate of true hits-based giving, you’ll probably fail with most of them.

Five direct-to-project grants is not a portfolio - at a 4% hit rate, you have an 82% chance of achieving no big hit at all.

You might not have the time or the expertise to vet large numbers of direct grants per year. That’s where intermediaries like evaluators and pooled funds can help - you essentially ‘buy’ their diversified risky portfolio, rather than building your own. Large enough funds can make tens of grants per year - even well over a hundred - without losing their nerve or running out of vetting capacity.

Capital matters too. AI staff may be tied to preset sales schedules and extended lock-up periods, forcing them to build a portfolio over time, as money arrives in a steady drip of (relatively) smaller amounts.

This makes the riskiest kind of grant - seeding a new venture from scratch with meaningful start-up capital - hard to do alone, even if the most senior staff could still free up substantial amounts each year.

One solution is again to pool your funds with other donors, either via a fund or a donor circle. This allows small donors to club together to create a bigger funding lever.

How should I balance evidence and risk in my donations?

Recent events should tilt the balance towards risk. Effective giving has historically pushed the other way - towards evidence-backed opportunities that give donors much more confidence than the average grant.

These opportunities are likely to be amply supported in the Funding Anthropalypse, though. Indeed, as far back as 2016, Coefficient Giving (then Open Philanthropy) made this point when talking about hits-based giving:

If an idea is backed by strong evidence, expert consensus, and obvious appeal, it’s probably already well-funded.
Coefficient Giving

If that was true then, the evidence suggests it is much truer now. GiveWell has lowered its funding bar from 8x to 6x its benchmark, which it estimates may result in $90m in counterfactual grants. This is the direct result of more available funding for higher-confidence global health bets, and that was before its commitment from Coefficient Giving increased to $1bn.

Many donors should respond by taking more risk.

Although GiveWell still believes that it will find many unfunded opportunities, and Coefficient Giving warned that its increased commitment shouldn’t be seen as a steady state, money is currently arriving at the safer end of grantmaking faster than opportunities are being discovered.

The answer isn’t necessarily to switch entirely to hits-based giving - it’s to increase your risk tolerance to take account of the new facts on the ground. If you score even one hit, the impact can be enormous - see the polio vaccine. You just won’t know in advance which bet will win, and so need to spread your bets appropriately.

Do charities need RCTs before I fund them?

Usually not - look for an intervention that has been tested by an RCT, but don’t necessarily require that an organisation has measured its own implementation this way.

Donors who want to take more risk need to match the evidence appropriately to the speed and type of intervention. Although RCTs are excellent evidence of effectiveness, they have significant drawbacks, including prohibitive costs (often several million dollars), years-long delays in producing results and sometimes being the wrong way to measure the intervention, full stop.

Jonathan Jackson is CEO of Dimagi, a large and respected global tech-for-good social enterprise (disclosure: I recommended that a donor make a five-figure grant to Dimagi earlier this year). He told me recently about an RCT of an AI-driven mental health intervention - by the time the programme was halfway through the data collection period, the LLM it relied on had been retired, making the study impossible to complete.

This doesn’t mean we have to give with no evidence at all. The alternative is proper monitoring, evaluation and learning plans, calibrated to the size and type of intervention. These include outputs, outcomes and accurate costs, measured against targets set in advance, alongside probabilistic bets on success. If housed within a clear theory of change, this shows whether a project is on track without a randomised study - and lets donors accept a grant’s risks with clarity, and offset those risks elsewhere.

I’m expecting an AI windfall. What should I do?

First, you should seriously consider diversifying your wealth. At present, the overwhelming majority of your philanthropic funds might be held in a single volatile asset. Worse, it could be the same asset class (and likely the very same asset) that the biggest players in effective giving are already exposed to. This is a serious ongoing risk to the whole effective giving space.

Diversifying could well mean less philanthropy in total - if the stock keeps appreciating, you’ll give away less than if you’d held your position. However, the whole point of hedging financially is that you accept swapping assets that prove more lucrative in hindsight for less lucrative ones, in order to improve your risk-adjusted returns.

To be clear, this isn’t financial advice - speak to a professional for that. It’s philanthropic advice.

It’s not possible to build a thriving effective giving ecosystem, and in particular to seed and scale new ventures, if your wealth evaporates in a market or company downturn. This problem is worse because the rest of the effective giving space is exposed to the same asset class. If AI valuations crash, most of the movement’s funding crashes at once - and a philanthropic dollar will be worth the most at exactly that moment. Only the donors who diversified will have one to give.

Creating consistent funding even for risky ventures is one of the biggest opportunities of the coming windfalls. By contrast, writing one cheque and being unable to back it up the next year is often worse than not writing a cheque at all.

Second, adjust towards more risk-tolerant philanthropy (unless you’re already there). Keep your core in relatively low-risk bets, but build a risky bucket that writes enough cheques to survive a high failure rate. Allocate it yourself, if you have the time and expertise; otherwise, buy into a risk-diversified portfolio from a small number of funds that you trust.

For the most risk-tolerant donors, you can then also assign a small amount to direct giving, focusing on completely new ventures with a credible team behind them. But only do this if you’re happy to fail with almost all this giving and to take a punt with minimal information before making a decision.

(This doesn’t solve issues of power concentration - I’ve written about that separately - but you can tune your allocations until the balance feels sufficiently ambitious.)

Third, I recommend committing to three disciplines in advance:

1. Define what failure and success mean, and distinguish ‘failures’ from ‘mistakes’. If you expect 65% of your grants to fail, and they do, that’s fine. If you find that 65% of your grants look like active mistakes in hindsight (for example, they do more harm than good, or the founder turns out to be a fraud), that’s something you need to learn from. It will help to define each before starting.

2. Work out how much you can securely commit over a two-year horizon. This will give you enough to double down on things that look good after the first year. It may also influence your wealth management strategy, so you can be confident of staying the course.

3. Be open about your failures. The leader in this is GiveWell, and we can educate the whole field by talking more about what didn’t work and why. It’s currently less painful to keep funding something that doesn’t work than to pull your support and publicly admit to a failure. That’s worth changing.

If you would like advice specific to your circumstances and goals, from me or from someone I would recommend, please get in touch.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and is the fund manager of its Mid-Stage Global Health Fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

I used Claude to help structure my thoughts and to suggest improvements and flag gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

Coefficient Giving (cG) increased their 2026 commitment to GiveWell from $175m to $1bn this week. Both organisations have reduced their funding bars this year, and cG says they are spending against anticipated giving from “extremely volatile assets”. Where should other donors give in light of this news?

I think the answer depends on size. For donors giving thousands, not much has changed. For donors giving millions, a lot has.

We should expect much more of this as the Funding Anthropalypse unfolds.

What just happened with GiveWell and Coefficient Giving?

Coefficient Giving announced a $1 billion gift to GiveWell on 23rd July (just two days after I posted about power concentration in effective giving). This increases a previous commitment of $175m for 2026. The announcement, from Alexander Berger and Otis Reid, says they made the decision essentially based on uncertain predictions about future funding. In particular, they think donating now will do more good than donating later, which is reflected in the fact that both organisations have lowered their funding thresholds this year (GiveWell from 8x their benchmark to 6x, Coefficient from 2,000x their benchmark to 1,000x - these are different scales, but cG reckons its 1,000x is roughly 4.5x in GiveWell’s units).

This is a huge increase in cG’s support for GiveWell, nearly doubling the amount they have given to GiveWell’s recommendations over the last decade. That’s ten years of donations, from GiveWell’s single biggest funder, nearly doubled in one gift. I expect this to do a great deal of good.

Although cG says that this could be a one-off surge, it has implications for other donors either way. GiveWell’s announcement says that they think their grantmaking “is likely to approach or even exceed $1 billion this year”. However, even if GiveWell meets that aspiration, cG’s gift is roughly the size of everything GiveWell expects to grant this year, before you count gifts from other donors or funds already rolled over from previous years. (cG frames this as their allocation “for 2026”; GiveWell calls it a “near-term” commitment, so some may be deployed over more than one year. I take cG at their word and treat it as this year’s money.)

Why did Coefficient Giving give $1 billion to GiveWell?

cG’s reasoning is multifaceted. In short, they anticipate large increases in future funding and are spending now against that expectation. This is both to buy marginal impact at current rates (because an influx of future funding would probably fill the most cost-effective opportunities) and also to prepare GiveWell and their grantees for this funding surge.

They don’t specifically mention AI wealth, but it’s fairly clear what they mean: “much of the future funding we’re anticipating depends on the valuations of some extremely volatile assets.”

Intriguingly, part of their thinking rests on the possibility that advances in AI could make the world so much richer that it reduces the need for global health and development giving altogether.

This is the first major example of how the Funding Anthropalypse will reshape the world. In essence, AI wealth and short AI timelines are now setting budgets for bednets, vitamin-A distribution and vaccination incentives.

Should I still give to GiveWell?

There is still an extremely compelling case to give to GiveWell. They remain the most rigorous, transparent funder in cost-effective global health giving; even with this increase, their budget will still be roughly ~⅛ that of the Gates Foundation; and 6x GiveWell’s benchmark is still some of the best marginal cost-effectiveness a normal donor can buy.

It’s also notable that GiveWell has itself committed to tell donors if it thinks their money should go elsewhere. Given their market-leading commitment to transparency, I think we should cautiously trust them on this.

Accordingly, if you are giving GiveWell thousands each year, I would stay as you are. You’ll get excellent outcomes from your giving - and, practically speaking, you don’t have a big enough lever to substantially alter the field, or to justify in-depth, bespoke advice on how to do that.

Does my donation still count now that big money is arriving?

Yes, it does. At the sector level, this amount of money is nowhere near enough to end extreme poverty or preventable under-5 mortality. The United States alone gave $37.3 billion less in overseas development assistance in 2025, compared to 2024, so this extra donation doesn’t even touch the sides. Even if the Funding Anthropalypse materialises, we are orders of magnitude short of the money needed to achieve a truly just world.

Even within cost-effective global giving, as cG themselves argue in their announcement, donations now help GiveWell to fund great projects and build their own capacity to do more good in the future. Although their bar has fallen, projects that are 6x GiveWell’s benchmark remain outstandingly good.

The bar hasn’t fallen because the things we can fund got worse - it fell because money is arriving more quickly than we can find opportunities.

What should large donors do following GiveWell’s $1 billion gift?

If you are giving millions of dollars this year, the highest impact move is probably not adding to GiveWell’s margin. Instead, it is to build the capacity that will help to absorb next year’s likely windfall.

Indeed, cG themselves say that they are “betting that other funders will pick up the baton for future giving.” They even concede that, with GiveWell’s marginal cost-effectiveness now lower, “some funders might be put off by that and want to allocate their funding elsewhere instead” - a reaction they call, to some degree, “perfectly rational.” cG has now given GiveWell an amount roughly equal to its expected grantmaking for the year, so the marginal large gift to GiveWell buys less than it used to. In my view, the highest-leverage way to pick up that baton is not to add to GiveWell’s expected surplus, but to build the absorption capacity that lets next year’s windfall land well - the very thing cG says it is racing to create.

This is a topic I specifically advise donors on, including AI wealth. It’s not a question of marginal cost-effectiveness. Instead, it’s about how much future capacity a dollar today will unlock. The following are options I found in my own research, in rough order of their ability to create new absorption capacity next year:

1. Fund operating costs for capacity-constrained funds. As more funding arrives, most of the great projects we already know about will be funded up to their marginal cost-effectiveness. We therefore need to be able to unearth more opportunities, at every stage of the pipeline. The binding constraint is likely to shift from available funds to available projects. It is therefore highly effective to fund more research capacity for smaller funds. An example is the Happier Lives Institute, which is currently raising operating costs to expand its research. My own analysis suggests that modest operating grants here could unlock very large sums in cost-effective wellbeing interventions; I’m happy to share the figures with donors on request. There are also fledgling evaluators in livelihoods, gender-based violence, humanitarian response and global health policy, and each one should unearth new giving opportunities.

2. Give regranting capital to smaller cost-effectiveness funds. How much a fund receives next is likely to be correlated with how much it regrants this year - donors don’t typically give tens of millions to funds that have only managed single-digit millions to date. For example, the DIV Fund has so far raised around $48m to deploy over four years; helping it move materially more than that in its first year will give it a stronger track record with which to raise much larger sums in the future. Likewise, my own Mid-stage Global Health Fund is on track to move $6m this year, but increasing that to $10m+ would move us to a different level. (Please note the obvious conflict of interest in me recommending my own fund.)

3. Fund incubator and seed prizes, to create lots more organisations that could scale. Large grants, which GiveWell expects to move towards, are ultimately reliant on a healthy pipeline of startups. Ambitious Impact has already launched numerous cost-effective charities, many of which are now GiveWell funded (Suvita, Ansh, Taimaka, Lafiya, Family Empowerment Media etc.). D-Prize seeds more ventures than anything else in the ecosystem and has a strong historical record of those interventions scaling (8% go on to reach >100,000 people). They could both do more, with more.

4. Fund things that GiveWell doesn’t specialise in. At present, the ecosystem often looks to GiveWell to fund everything - seed grants, mid-stage grants, USAID-response grants, massive scale grants. However, no one can specialise in all things, and it would be better for grantees (and less risky) if more specialist funders existed. That includes specialist seed grants from places like AIM, specialist mid-stage grants from places like Ultra Philanthropy and DIV, catalytic funding from places like Founders Pledge, and so on. Each specialist funder can then design its processes, evidence requirements, risk tolerance, turnaround time etc. to suit its specific tranche. I strongly encourage donors to support these varied specialists. You can also invest in important things that fall outside GiveWell’s predominant focus on under-5 mortality, such as education or wellbeing.

5. Take a risk on something new that could grow rapidly. These grants require the highest risk tolerance, but there are implementing organisations with ambitious plans to reach serious scale, such as Evidence Action’s internal accelerator or Sanku, whose new TED talk sets out a vision to reach a billion people (it currently reaches around 70 million, having grown from one million after a single unrestricted gift from MacKenzie Scott). Likewise, there are several interesting proposals for deploying AI in global health. Most are not public (get in touch if you want to know more), but one that is public is the AI Access Initiative, incubated by Evidence Action.

Each of these options builds the absorption capacity of the field as a whole and de-risks the ecosystem, ensuring a richer pipeline of opportunities for payers at scale like GiveWell. And now that GiveWell is practically guaranteed to receive more this year than it can re-grant at its own bar, large donors should take these options seriously.

Of course, you should note throughout that I have a conflict of interest in recommending some of these options, as an operator in this space myself via the Mid-stage Global Health Fund. Discount accordingly.

What does this mean for power concentration in effective giving?

Well, it didn’t make it better.

The biggest (or second biggest, depending on where you draw the boundary) funder in cost-effective global health just roughly tripled its expected grantmaking this year, from around $418m in 2025 to a likely $1bn-$1.5bn in 2026. The next biggest funder in the space gives in the double-digit millions. If you took this risk seriously last week, you should take it even more seriously now.

It’s also interesting to map who depends on whom. cG and GiveWell are the two largest players in cost-effective giving, and they are not independent: cG’s new commitment is roughly the size of everything GiveWell expects to grant this year. So the field’s two biggest institutions are now even more closely linked, and to a great extent their outlooks are dependent on the same set of “volatile assets”. This isn’t diversification. It is the opposite: the two largest funders in effective giving are now exposed to the same volatile source at the same time.

cG themselves acknowledge that this could be a one-off surge, and that the future funding they are counting on might not materialise. So the sector’s two largest flows now depend on volatile asset prices, on cG’s ability to predict the future, and on how it sets its own funding bars from one year to the next.

I expect the $1 billion gift to GiveWell to be an outstanding grant at the object level, funding highly cost-effective programmes that do an enormous amount of good for very little money. This excellent grant does nothing to address the fragile architecture of this space. In fact, it makes the risks significantly worse.

I am expecting a windfall this year. What should I do?

My advice has not substantially changed from the post I wrote about power concentration just a few days ago. The only difference is that I believe my thesis has been somewhat validated by this news.

Accordingly, any donor giving seven figures or more, and even some donors in the six-figure category, should at least consider splitting their donation between GiveWell (where they will get excellent marginal cost-effectiveness) and building the ecosystem, exactly as cG called for in their announcement post.

This is a rare chance for donors to improve the ecosystem for grantees and, in doing so, to do more good on the ground.

If you would like advice specific to your circumstances and goals, from me or from someone I would recommend, please get in touch.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and fund manager of its mid-stage global health fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

I used Claude to help structure my thoughts and to suggest improvements and gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

Imagine you could live in one of two worlds in 2027.

In the first, Coefficient Giving is twice as big.

In the second, there are two Coefficient Givings of the same size.

Which seems better to you?

Having considered it for a while, I would choose the second world. Put plainly, two Coefficient Givings beat one twice as big. I think the potential $37bn+ AI windfall is a once-in-a-lifetime opportunity to achieve this, with significant long-term benefits.

Is effective giving too concentrated?

In my view, yes.

Coefficient Giving (and to a great extent GiveWell) have enormous influence and power within the ecosystem, by virtue of their size and track records. Coefficient Giving is also the predominant or only funder in some of its domains, such as funding the effective giving ecosystem and farm animal welfare.

The situation seems analogous to a monopsony in economics:

monopsony, in economic theory, is a market situation in which there is only one buyer... Although cases of pure monopsony are rare, monopsonistic elements are found wherever there are many sellers and few purchasers. (Encyclopedia Britannica)

You could argue that there is in fact a diversity of funding in some cost-effective cause areas. For example, in AI safety, there is Longview Philanthropy, the Survival and Flourishing Fund, Astralis, the AI Safety Tactical Opportunities Fund etc.

In global health, there is the Gates Foundation (granting roughly 20x more than GiveWell and 8.5x more than Coefficient Giving each year); the remaining bilateral and multilateral funders; and a large number of smaller foundations (Mulago, CRI, WAM Foundation, DRK etc.). You could therefore also claim that the global health landscape is sufficiently diverse.

However, Coefficient Giving outsources a substantial amount of its global health grantmaking to GiveWell - around $1bn to date, and $175m for 2026 so far. It’s therefore not really accurate to describe GiveWell as a counterweight to cG - it could even be considered an extension of it.

More to the point, I think the true test of over-concentration is not just the number of grantmakers in a field - it’s also their size and, more importantly, the amount grantees rely on them. In AI safety, for example, Coefficient Giving gives substantially more, and in almost all cases at least an order of magnitude more, than any other actor. In combating factory farming, it is widely estimated to be the largest funder globally, and in supporting the effective giving ecosystem it is essentially the only funder of note.

Across all cause areas, it is often true that there is no other funder who could step in and become a grantee’s majority funder if the main one walked away.

This, to me, is the true test of over-concentration, and it seems true of both Coefficient Giving and GiveWell, for a large number of their grantees. It is also not just theoretical - it is something that Coefficient Giving’s grantees themselves are concerned about.

In short, Coefficient Giving and GiveWell are exceptionally important to the majority of their grantees, in ways that cause significant risks to the grantees, the ecosystem and the problems we are trying to solve.

What’s wrong with one dominant funder?

To some extent, the answer to this is just ‘read about power concentration’. However, I observe at least five major issues with the current situation.

The first is how dependent grantees are on one organisation for their financial health. The decision of a single program officer, formed under uncertainty, can virtually dictate the grantee’s survival. This is unduly risky for the grantee and also has unwanted second order consequences. For example, I have literally been in strategy meetings where organisations have asked “what will Coefficient Giving think of that?”

The second is that the blind spots of the organisation and individual grantmakers can become significant failure points. Even the very best program officers (and Coefficient has many outstanding ones) are also human, with biases, limited hours in the day and imperfect judgement. Especially in areas of high uncertainty, it is incredibly helpful to have grantmakers who can step in if another can’t or won’t. And, while dominant organisations may try to hedge internally against their own missteps, I think it is much easier to do this when the decision-makers are separated structurally. That is, several smart people thinking independently about an issue in different organisations are more likely to hedge effectively than the same smart people in the same team or office.

(As an aside, an ecosystem with a diverse set of large funders is also more attractive to top talent, versus one where there is a ‘party line’ and power is concentrated in a single funder. This will likely be relevant if we want to attract a lot more people to work on causes we care about and increase our ability to deploy money well in these fields, something I intend to post about in the future.)

Third, a single dominant funder is also a single point of failure. We are currently all-in on Coefficient Giving and GiveWell’s reputations and assets. We already saw how incredibly damaging it was when the FTX Foundation collapsed and it was only a relatively small fraction of the ecosystem at the time. I am in no way suggesting any malpractice at either organisation - but I am uncomfortable with relying so completely on their continued good judgement and lack of scandal.

Fourth, it is bad for grantees if a single funder tries to do the majority of grantmaking in a given field, because they are very unlikely to be well-suited to every type of grant. Trying to assess $100m grants alongside $100k seed grants is almost never the right option, because the decision timelines, risk tolerance, due diligence and opportunity costs are completely different at different scales.

At the moment, we expect GiveWell to fund seed grants, mid-stage grants, USAID-collapse-response grants and things at massive scale - but this leads to numerous issues, such as small organisations reporting substantial delays in being evaluated, when they need a rapid review for a grant before their runway expires.

You can instead imagine a pipeline of seed funders and incubators, followed by mid-stage funders who can underwrite growth, before GiveWell funds the best things at serious scale. And this is without even considering that some grantees will have constraints that stop them taking money even from a funder perfectly suited to their stage of life, for example if they want money that is politically-aligned with their other supporters.

Finally, monopsonies also have a dominant function in shaping the market in which they operate. This creates a dominant view of what gets evaluated in the first place (cause areas and organisations), and smaller organisations usually follow their assessments. Unless we assume that their judgement is perfect, and that big organisations can adapt at speed when the situation evolves, this is sub-optimal.

Can GiveWell and Coefficient Giving absorb the windfall?

The two most dominant organisations in the space are making concerted efforts to grow quickly. Both are hiring fast and both are specifically focusing on high-absorption options to deploy more money quickly. They are also outstanding grantmakers and some of the only places with track records of handling hundreds of millions in grants. You can see the attraction for time-poor donors, looking for a safe place to give.

However, the scale of the Funding Anthropalypse is likely to exceed their capacity, even if that capacity increases significantly. Last year, GiveWell regranted $418m, and Coefficient Giving over $1bn. Even with the most aggressive scale-up plan imaginable, it seems unlikely to me that either organisation can even 5x its grantmaking in the next 24 months, and we may be talking about sums far in excess of the $7bn in grants that this would imply ($2bn for GiveWell and $5bn for cG). As an incredibly rough rule of thumb, I am sceptical of any large organisation claiming it can more than double year-on-year, which would imply a ceiling of $2.8bn.

Coefficient themselves acknowledged this bottleneck recently and others have made compelling arguments about an upcoming logjam. It’s also notable that even the Gates Foundation, arguably the best-resourced foundation in history, took roughly 25 years to ramp up to its current annual payout. Scaling up grantmaking is simply very hard to do quickly.

If the Funding Anthropalypse gives more to these two organisations than they can deploy, the most likely outcome is that this money sits on the sidelines for a long time. This is, in my view, a very bad outcome, given that global health has seen a global retrenchment; meat consumption per capita is rising; and many in and around the frontier labs expect very short AI timelines.

Of course, if you are completely ‘time-neutral’, this may not matter to you. I do also think it’s very difficult to judge exactly when and how much to spend now, versus saving capital for potential future opportunities, especially if you think you would need to lower your bar to make grants today. I expect we can all agree, however, that impact shouldn’t be constrained solely by a lack of grantmaker capacity.

Zooming out, this sort of difficult question is the whole point: I think capital sitting in a queue for years would be very bad, but I might be wrong, which is exactly why we want a diversity of decision-makers!

Lastly, even if both organisations defeat my predictions and scale up hugely, this somewhat solves the absorption problem; it only makes the concentration problem worse. So two Coefficient Givings would still beat one, even if the original is twice as big.

Could doubling Coefficient Giving be better?

Although I have come to the firm conclusion that we need more Coefficient Givings, there are strong arguments against my position.

The process of establishing new massive grantmakers will be really messy. It will likely start a talent war, as senior grantmakers are rare and reportedly hard to hire, so new organisations may just poach from existing ones.

Equally, setting up new foundations might not help epistemic diversity much if they just hire from a homogenous talent pool with the same views. However, I think the contention that the only smart, talented people are those already known to or within Effective Altruism doesn’t survive scrutiny, so there is plenty of potential to hire excellent people with diverse viewpoints and significant expertise. For one thing, about 260,000 aid professionals lost their jobs recently, and some of them are already doing excellent things that look very cost-effective.

We should also expect the quality of grantmaking at less experienced foundations to be worse in the short-term. On a pure cost-effectiveness basis, it is likely that GiveWell and Coefficient Giving can deploy money today with the most expertise, experience and analysis of anyone in the space. You also lose efficiencies of scale if you set up multiple organisations to move the same amount of money. The trade-off in both cases is between short-term and long-term consequences.

The coordination costs to multiple actors can also be steep. For example, it is not unheard of for two grantmakers to write large cheques to the same organisation, thus double-funding it, and avoiding this is a cost to both funders.

Maybe most concerning, there is a danger of empowering harmful grantmakers. This is a version of the unilateralist’s curse: when several funders can each act alone, it only takes one - the most optimistic - to back a project. The more independent funders there are, the higher the chance a harmful project finds its single most enthusiastic backer, even if most funders would have said no. This is particularly concerning in AI safety, where you could easily imagine someone accidentally supporting something harmful.

This objection is real and serious. However, this seems to argue for a handful of large, professionally-run funders, but not just one. It’s not obvious to me at all that a single dominant grantmaker is immune from making a harmful call - indeed, the risk of this seems quite acute.

In addition, in a practical sense, we are probably talking about fewer than fifteen peer competitors (say, a foundation for each of Anthropic’s co-founders and significant growth from Longview, AISTOF etc.). There simply isn’t enough money to found a hundred unilateral Coefficient Givings, and a small number of high-quality grantmakers is better-insulated from the curse than hundreds would be.

Lastly, the original paper lays out several remedies, none of which is to empower single decision-makers. Instead, it argues for a principle of conformity (pages 8-9) when dealing with unilateral decisions, and also for judicious information-sharing. A handful of funders who compare notes are barely unilateralists at all; a lone funder, with no peer to check it, is the one actor who can never be talked down.

Ultimately, I do not believe that any of these concerns outweigh the downsides and risks of the status quo, and I think this is a once-in-a-generation chance to fix it.

How has this played out before in global health?

Probably the most famous example of consolidation is Warren Buffett. When he made his 2006 pledge, he committed the bulk of his giving to the Gates Foundation, rather than setting up his own thing. This was widely lauded as an example of putting impact before ego, as it gave the Foundation significantly more resources, without any extra overhead.

I think this is really a cautionary tale, however. If Buffett had set up a similar foundation, of similar size, it would have reduced concentration and increased epistemic diversity. It would also have insulated the ecosystem better against scandal, potentially avoided the frustrations some grantees have with the single foundation, and the Gates Foundation could still have achieved its grantmaking to date from Gates’ own fortune.

It’s also notable that Buffet has since changed his mind, first excluding the Foundation from his legacy, and earlier this month excluding it from his mid-year giving for the first time. Whilst I am not claiming that this was a move driven principally by concerns about power concentration, it’s interesting that the arrangement did not survive as intended.

Where can the next Coefficient Giving come from?

The good news is that even some of the strongest objections to diversity can be mitigated by choosing the right donations.

First, there’s no need to invent new mega foundations from whole cloth. There are many good mid-tier organisations that donors can contribute to, across a variety of cause areas. Most are deploying in the tens of millions and would provide significant value if escalated to the hundreds of millions. To name just a few already directing >$10m/year:

• Longview Philanthropy (existential risk)

• AI Safety Tactical Opportunities Fund

• EA Animal Welfare Fund (animal welfare)

• DIV Fund (global health)

• Founders Pledge’s Global Health and Development Catalytic Impact Fund

• Navigation Fund (multi-cause)

(This list is not exhaustive and is very loosely vetted on general reputation and places deploying >$10m/year. For more in-depth analysis, please get in touch.)

Indeed, although this isn’t a developed model, I think there is roughly a bellcurve-shaped distribution of size and risk-adjusted effectiveness. That is, at the lower end, funds need to grow or merge to have a significant impact - but once they reach the hundreds of millions per year scale, diversification is a defensible alternative, and at the billions per year, it’s essential.

My own mid-stage global health fund is an example of this. At our present run rate of only $3m per year, we are hardly ‘solving’ the mid-stage valley of death in global health. Indeed, we can’t support even one late mid-stage organisation in a significant way. At $50m-$100m, however, we could be a major force for good, but we wouldn’t want to be the only player of that size. A world where we operate at that level alongside DIV Fund, Livelihood Impact Fund and others seems optimal.

I also want to be clear: I definitely do not want Coefficient Giving or GiveWell to shrink. They are tremendously successful grantmakers who have done an enormous amount of good, shaping entire fields for the better.

My answer is instead to create counterweights and coordinating peers, who can take the opposite side of 60:40 calls, diversify grantees’ risk, make grants on asynchronous timelines and fund things the others can’t, won’t or shouldn’t (such as 501(c)(4) activity, for example).

As Coefficient Giving put it themselves: “when external funders come to us for advice on giving, we are typically able to recommend funding opportunities that we believe are 2-5x as cost-effective as Good Ventures’ marginal AI safety funding.”

What should I do with my AI windfall?

If you are someone who expects to come into AI wealth, you can be an active part of the solution (provided you can see past the incessant noise about your philanthropy).

My strongest piece of advice is to diversify your own grantmaking plans. You have the luxury of being able to use outstanding grantmakers like GiveWell and Coefficient Giving, while also supporting a handful of other actors to grow and de-risk the ecosystem. You can do this whether you have $10m to give or $100m - and, indeed, at this scale, you might be particularly well-suited to supporting smaller funds and grantmakers alongside other donors. You might also effectively unlock matched giving, to augment your resources.

If you are a particularly large philanthropist, you have an even bigger opportunity to build something that might be a genuine counterweight to the biggest existing organisations. This applies most obviously to the companies’ co-founders, the OpenAI Foundation and the Long Term Benefit Trust at Anthropic - mathematically, these are the only places that can rival the resources of Good Ventures as actual peers. However, I think anyone contemplating over $100m in giving should give serious thought to seeding or building a new vehicle, at least with part of their resources and especially if they can do it with a few other donors of the same size. This is admittedly a marginal call at the nine-figure scale, and I would recommend reading these existing resources and sourcing bespoke advice before committing.

In general, deciding the right split between existing, scaling and new allocators is very difficult, and this is what advisors are for. As someone who runs an advisory and has a fledgling fund, I obviously have conflicts of interest here. Accordingly, I would discount the above appropriately, stress test the arguments and consider a range of advice.

For free advice on where to get help, including pointing you to others in the space who are better-placed than me, please get in touch.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and fund manager of its mid-stage global health fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

Thanks to Jakob Graabak, Olivia Kaye and Patrick Kaczmarek for their review, suggestions and feedback. I also used Claude to help structure my thoughts and to suggest improvements and gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

Short version: if you want a share of the coming Anthropic and OpenAI windfall - the $37bn+ that could be in play next year - the way in is to become 'legibly excellent', so the evaluators and donors that frontier lab staff already trust point them to you. The way to blow it is to cold email them. Here's why, and what to do instead.

Someone in a WhatsApp group recently asked for donation recommendations that they could pass to Anthropic employees. The subsequent exchanges reminded me of the episode in The Last of Us when the infected start pouring out of the ground. Every person in the chat started sending their own charity/fund/pet project (including me, in fairness) and the questioner was immediately overwhelmed.

As someone who has thought a lot about the Funding Anthropalypse (hat tip to Pablo Melchor for the name), I want to suggest ways for projects to access some of the $37bn+ in donations that could be available next year. In short - become ‘legibly excellent’ to ants.

Do not cold email them.

Should I cold email people at Anthropic and OpenAI?

Frontier lab employees have told me they are each fielding up to 20 cold emails a week from charities and wealth managers, touting for business. They believe their emails are being scraped and even reverse engineered. This at least implies that charity leaders believe it’s an effective fundraising strategy to pitch anyone with an @anthropic email address.

I apologise in advance if this sounds patronising, but please do not do this. It has next to no chance of working, it has harmful side effects, and it actively damages your personal chances of receiving funding.

How can I get noticed by frontier AI lab donors?

In the past year, donors working inside frontier AI labs have started routing large global health gifts to me to place. I’m currently advising on a multi-year, eight-figure global health allocation across a set of vetted funds. My advice for success with these donors?

Make it your mission to be ‘legibly excellent’ to them.

I think Justin’s point about legibility is right - genuinely good work can be illegible and lose out - which is why you have to work at legibility rather than assume your quality speaks for itself.

In terms of excellence: the most credible way to demonstrate this is to be highly rated by evaluators and grantmakers already trusted by AI staff: Coefficient Giving, GiveWell, Longview Philanthropy, the EA Animal Welfare Fund etc.

A happy side effect of this is that it will probably make you a better organisation if you find ways to score more highly with these intermediaries, as they prioritise evidence-based, cost-effective organisations. (Indeed, by becoming more legible to AI employees, you will also become more legible and higher-scoring with these evaluators.)

If getting onto the radar of these organisations isn’t feasible, develop your internal legibility through a thorough theory of change, solid monitoring and evaluation to track your outputs and outcomes, and a strong understanding of the evidence base for your intervention. That way, you’ll have a compelling suite of things to show any prospective donor.

I also recommend having a credible and thought-through plan for scaling. Absorption capacity is going to be a major challenge of such a huge philanthropic windfall - with so many donors concentrating in the same areas, it is likely that most projects will quickly be funded at their current scale. You can differentiate yourself by having a credible plan for growth, including the extra non-delivery staffing you might need (CFO, HR function, an MD to keep the engine running smoothly, enough monitoring capacity and sophistication that you can accurately tell what’s happening on the ground).

Many donors want to support things now that can absorb 2-3x more funding next year. Many organisations have a dream for how to scale, but a serious, costed, red-teamed plan will really make you stand out. Some funds and evaluators specialise in scalability, such as the DIV Fund, Mulago, the Prevail Fund and Ultra Philanthropy’s Mid-Stage Global Health Fund (which I myself run, to declare a conflict of interest). Each of these may be willing to offer advice.

How much time do I have to build awareness?

Actually more than you think. The two major IPOs in the pipeline are Anthropic and OpenAI. Each has confidentially filed a draft S-1, which is the first step to going public, but neither is expected to float until Q4 2026 at the earliest.

Even after an IPO, there will be delays before philanthropic capital comes online. Each equity holder needs to decide how much (if any) of their holdings to cash out, and the average staffer will have a 180-day lockup before they can sell anything. Accordingly, money is likely to land in donor-advised funds around Q2 2027 at the earliest, which gives plenty of time to build credibility and legibility, and to stress-test a plan for scale.

What’s so bad about emailing them anyway?

You might think that people can just ignore emails they don’t like. Or even that receiving incessant pitches is a small price for being a newly minted millionaire.

However, I still urge you not to start cold emailing people because it almost never works. If you put yourself in the shoes of a donor, would you fund someone who contacted you unsolicited? How much less likely would you be to fund the twentieth person who emailed you unsolicited?

More than just being annoying, unsolicited outreach suggests you don’t respect boundaries, which is disqualifying for someone deciding whether to trust you with their money.

At scale, this approach also creates negative side effects. Feeling overwhelmed by demands makes every potential donor feel defensive and raises barriers to giving to anything.

It also makes it impossible to assess opportunities on a practical level. As an advisor, I receive multiple pitches each week, and it’s my full-time job to assess them - and yet the most common unread email in my inbox is ‘that project that I haven’t got around to yet’.

This does require some collective discipline - it’s hard to do the slow work of building an amazing organisation and generating awareness of your work. But there is a shared responsibility not to make this an exhausting, unpleasant experience for donors, in ways that could really damage the ecosystem in the medium term.

I’m locked out, so why shouldn’t I just email them anyway?

I am aware that many nonprofit leaders are desperate for funding, and that it would be professional malpractice not to react to an incoming surge of funding. Posts like this will also compound issues of unequal access to wealth and wealth holders - if you don’t know someone who knows someone, how can you raise awareness?

I think it’s possible for anyone to develop relationships with AI equity holders in non-transactional ways. “Develop relationships” doesn’t require you to know anyone at the company already. As you build your excellent and legible organisation, you can present it where AI staff and advisors will find it: via evaluators; at philanthropic conferences; in think pieces; on podcasts.

(You could even, say, start a Substack and post a lot of things that aren’t in your personal financial interest. But probably don’t do that, I doubt it’ll work.)

This is more work than asking Claude to shoot off 50 cold emails, but it has a much higher chance of being effective, and it avoids the harmful downsides of firing emails into the void.

I’m an AI equity holder and I’m overwhelmed. Who can help me with my philanthropy?

My intention with this Substack is to offer useful advice and perspective on a major philanthropic event. Accordingly, I’m going to highlight organisations doing good work and offer the best advice I can, as far as possible without promoting my own interests.

To that end, there are many mature advisories already recommended by people at Anthropic and OpenAI. These organisations are values-aligned and thorough, with formidable track records: GiveWell, Coefficient Giving, Longview Philanthropy, Founders Pledge, Regeneration Group. I would recommend them to anyone.

For a second, personalised opinion on your giving, or to promote a plurality of views across the ecosystem, consider a boutique advisory or subject matter specialist: Senterra Funders, EA Animal Welfare Fund, Bedrock Philanthropy, (my own) Ultra Philanthropy. I’d argue the ecosystem needs more of these, not fewer: two Coefficient Givings beat one that’s twice as big.

Neither list is exhaustive, and the right fit depends on your cause area, risk tolerance and the services you need. If you’d like free, neutral advice or an introduction to someone who can help, drop me a line.

Jack Lewars is the founder of Ultra Philanthropy, an independent advisory that helps major donors give for maximum impact, and fund manager of its mid-stage global health fund. He advises donors giving up to nine figures a year, and is Chair of Trustees at High Impact Athletes. Talk to him about your giving.

I used Claude to help structure my thoughts and to suggest improvements and gaps, as well as for proofreading; all views, primary drafting and final edits are mine.

The Funding Anthropalypse is the expected wave of philanthropy from equity holders in frontier AI labs (not just Anthropic). This is likely to change the face of philanthropy because of its sheer size, expected to be in the tens or hundreds of billions. The philanthropic ecosystem is not yet ready for this windfall, but we can give it a fighting chance with urgent action.

I’m Jack Lewars, the founder of Ultra Philanthropy. UP is an independent advisory that helps major donors - increasingly from tech and AI - give for maximum impact. I’m also fund manager of its mid-stage global health fund and Chair of Trustees at High Impact Athletes.

The Funding Anthropalypse is what I call an expected wave of philanthropy from people who founded or work for frontier AI labs.

To start off with, this will most likely be Anthropic and OpenAI, because each of them has confidentially filed a draft S-1 with the SEC, the first step to going public. In the next few years, however, AI has the potential to mint millionaires and billionaires at an unprecedented rate, so we may see many more donors and much more money coming online.

Will Anthropic and OpenAI staff donate?

At least in the case of Anthropic, we pretty much know for certain that large donations are coming. Early staff were incentivised with a 3:1 match from the company if they put private shares into a Donor Advised Fund (DAF) (and later staff still receive 1:1 matching). There is much speculation as to the size of this DAF, but it is undoubtedly in the billions of dollars, as the company’s valuation has skyrocketed since those early days.

Money or equity held in a DAF cannot be withdrawn; it can only be granted to registered 501(c)(3) nonprofits in the US, or something the DAF provider accepts as equivalent overseas (e.g. a charity registered in England & Wales that meets the IRS requirements of a nonprofit). This means that, one way or another, this money is being donated at some point. There is no legal way (or, at least, nothing that isn’t extremely sketchy) to use the money except for charitable grants.

In addition, many staff joined these companies specifically because they were effective altruists, and a key tenet of EA is giving away money you don’t need. They are therefore highly likely to give even more than is already in this DAF, with many staff pledging 80% of any money they make to charity.

And that’s without even mentioning the founders, each of whom owns tens of billions of dollars in equity, and has pledged to give away 80% of their wealth; or OpenAI’s Foundation, which owns 26% of the company, currently worth hundreds of billions.

OpenAI is similarly stacked with Effective Altruists. In general, I expect the actions of these market leaders to influence a wide range of AI companies, and to set a precedent for sharing the immense wealth that this field will probably generate. I also think the speed at which AI companies can generate wealth to increase the openness of founders and staff to philanthropy.

Why is this a Funding Anthropalypse?

This is philanthropy at historic levels. Nan Ransohoff gives some useful perspective, assuming $50bn per year comes online:

$50B/year could fully fund the annual budgets of the following organizations:

• 6 Gates Foundations (~$9B/yr), or

• 67 Coefficient Givings (~$1B/yr), or

• 100 GiveWells (~$500M/yr), or

• 333 Arc Institutes (~$150M/yr), or

• 5000 Institutes for Progress (~$10M/yr)

Ransohoff’s figures don’t completely add up (for example, 67 Coefficient Givings at ~$1bn is nearer $67bn than $50bn) and there is a lot of doubt about the exact amounts of money at stake. However, even if you doubt the exact figures, even 10% of this money would completely rewrite the rules. The biggest organisations in cost-effective philanthropy, like Coefficient Giving and GiveWell, deploy only a fraction of the expected windfall (as do the largest foundations in the whole world). Even if the biggest grantmakers in EA grow substantially, there is going to be more funding available than most of us have ever seen in this corner of philanthropy.

Where should Anthropic employees give their windfall?

This is the centibillion dollar question. We strongly expect the ‘traditional’ effective altruist causes to get enormous support: global health and wellbeing; animal welfare; existential risk, whether from AI or biorisk or something else; plausibly climate change.

The conventional wisdom is that this will be almost all of the employees’ donations, but I disagree. I expect many employees to have or develop a reasonably long list of causes of interest. Some of this will be because they suddenly have $100m in their DAF, and so a few ‘warm fuzzies’ donations seem fine. Some of it will be because there are many things of value that are not easily captured by EA’s focus on measurability and cost-effectiveness, like justice, racial justice, the arts, democracy protection, abortion access and so on. Some (potentially a lot) will be to offset the predicted near-term harms of the AI revolution, such as widespread job losses.

And some will be because, at least at the moment, the cost-effectiveness giving ecosystem is not ready for this amount of money.

According to unpublished data from Giving What We Can, cost-effective giving moved about $2bn last year. This means that, even if every organisation in the space grew 5x next year, it would still only handle a fraction of the expected windfall. And, as wise people have pointed out, growing 5x is incredibly hard, and sometimes actually impossible.

Accordingly, unless urgent action is taken, I expect a fair chunk of the philanthropy to go elsewhere, if the donors are eager to donate now, rather than waiting for the ecosystem to catch up.

What can we do to help Anthropic and OpenAI equity holders donate well?

We need several types of action to help equity holders donate effectively, if we accept two things:

1. In general, giving according to cost-effectiveness is a good way to decide where to give, given the overwhelming number of causes we could support.

2. The traditional recipients of EA donations will quickly have more money than they can distribute.

This funding wave is epoch-defining. We therefore need action to match its scale. For example:

• We should invest now in researchers and evaluators. I generally believe that you can find something cost-effective in almost any cause area. I believe particularly strongly that there are thousands of highly cost-effective interventions that are undiscovered in global health and similar areas. No sane person would argue that this amount of money can ‘end poverty’ (given that USAID was managing $35bn each year before it was dismantled). More researchers and evaluators can find these hidden gems and direct money to them next year.

• We should invest in projects that can actually scale. Notwithstanding the difficulties of scaling, some projects have the potential to move billions of dollars. Even if they are not currently ‘above the bar’, we should expect the bar to fall as the supply of money dramatically increases. Absorption capacity and scalability are now almost as important as cost-effectiveness (something I focus on deliberately in the mid-stage of global health).

• We should work to avoid creating or reinforcing damaging monopolies in cost-effective giving. I will write more on this in future posts, but in short: monopolies (or, in this case, monopsonies) are almost always bad. They are bad for the charities that need grants from them, because they become the only opinion that matters, expose the charities to risks from receiving all their money from one place and cause enormous harm if they collapse; and they are bad for donors because they have blind spots, restrict donor choice in finding a second opinion on their giving and make it harder for competitors to enter the market. Put simply, Two Coefficient Givings beat one twice as big. This is a once-in-a-lifetime opportunity to create a thriving, diverse ecosystem of mega funders, in a way that will de-risk and improve cost-effective giving for generations. We need to take it, while respecting those organisations that already have formidable track records and expert teams. Thankfully, there is enough money to do both with room to spare.

This list is not exhaustive but it is a start. This Substack is intended to bring these discussions into the open, and offer advice to grantmakers, evaluators, advisors and donors as we enter the Funding Anthropalypse.

Who can advise on giving after a tech or AI windfall?

Part of my goal is to discuss openly and honestly the challenges of such a huge philanthropic windfall. At the same time, however, I aim to give everyone involved the best advice I can, striving to be impartial. Accordingly, if this is a live issue for you, I suggest contacting one of the following. For free, impartial advice, or help choosing, please reach out to me.

Not listed in any particular order, and tiered by size:

The biggest current organisations

• GiveWell (global health and wellbeing)

• Coefficient Giving (multiple cause areas)

• Longview Philanthropy (existential and long-term risk)

• Founders Pledge (multi-cause)

Slightly smaller, but with a strong history

• EA Animal Welfare Fund

• Regeneration Group (multi-cause)

• Senterra Funders (animal welfare)

• AI Safety Tactical Opportunities Fund (AISTOF)

• Animal Charity Evaluators

• The Life You Can Save (global health and wellbeing)

Boutique advisors (especially if you want to diversify the ecosystem)

• Ultra Philanthropy (multi-cause, but principally global health - my own advisory, so conflict of interest warning)

• Ellis Impact (multi-cause)

• Bedrock Philanthropy (multi-cause)

I used Claude to give feedback and for proofreading; all views, primary drafting and final edits are mine.