The Long-Term Future Fund is closing down, and EA Funds is launching the Transformative AI Fund with a new full-time team.
The fund’s primary focus is technical AI safety and AI governance (including post-AGI governance), as well as supporting fields such as field-building and forecasting. We’ll also consider non-GCR implications of transformative AI such as flourishing futures and digital sentience, though we expect these to be a small minority of grants. For now, biosecurity and nuclear will only be supported at the AI intersection.
The fund will provide funding that is early-stage, fast, and tolerant of failure, to build the field’s absorptive capacity. We’re betting that a lot of philanthropic money is about to arrive, and that the field doesn’t have enough organizations to put it to work.
We aim to always be open to applications, respond quickly and reliably, and publish quarterly reports on what we fund and why.
Whether you’re an individual, a new organization, or an existing organization with a new idea, you can apply here.
We’re actively fundraising. We have a seed grant of ~$1M, but hope to raise significantly more, as we believe there are sufficient opportunities in the field to deploy several times more. The fund is designed to enable donors of all sizes to support early-stage projects that address transformative AI, with no time commitment required. Donate here or reach out at lowe.lundin@centreforeffectivealtruism.org
Background
The Long-Term Future Fund (LTFF) is closing down, as the outgoing team announced today in this post. Since LTFF’s identity has been tightly bound to the individuals running the fund, and the new fund’s scope will have substantial, but not exact, overlap, we have agreed with the team to wind down LTFF and present the Transformative AI Fund as a distinct fund with a new team. Since 2017, LTFF has made 820+ grants totaling just under $35M, and many great orgs, projects, and people have received their first check from the fund. We think LTFF has been highly impactful and are grateful to the team for their important work over the years as well as their help with the transition.
We think the niche LTFF has occupied (early-stage, small, and speculative projects) has been extremely valuable for seeding new organizations, and we aim to double down on it.
So: we’re launching the Transformative AI Fund, with a new full-time team.
Purpose
Over the coming years, this field is likely to receive far more money than it does today. In “The third wave of American philanthropy”, Nan Ransohoff estimates that AI wealth could add $37-100B per year in intended philanthropic spend, expects much of the early spending to focus on making the AI transition go well, and argues that at that scale the bottleneck shifts from money to good places for it to go. The field is not ready for that.
What’s scarce is organizations. Money gets absorbed by orgs that can hire people and run projects, and the field doesn’t have enough of them. There’s a lot of machinery for finding and developing people (fellowships, courses, research programs, upskilling grants) and much less for getting new organizations started.
Organizations need founders as much as they need money, and good founders are hard to find. But how many people attempt founding isn’t fixed: it depends partly on how costly the attempt is. For most people, starting something means months of uncertain fundraising weighed against a safe job, and reliable early capital changes that calculation. We can’t conjure founders, but we can lower the barriers to entry. Most new organizations start with small checks: a first $25k grant is often the difference between a pilot running or not. Many institutional funders don’t serve that size (for a number of good reasons).
We also think there’s room for more funding that stays open-ended about what it backs. A lot of early-stage funding starts from a funder’s view of what should exist and then looks for someone to build it, which works well for ideas already on someone’s radar. Taking applications from anyone, whenever opportunities arise, casts a wider net and finds projects nobody thought to ask for. We expect a meaningful share of our best grants to be things we wouldn’t have known to solicit.
On timing, larger checks to established organizations presuppose the initial small checks that got those organizations started, and organizations take time to build; Ransohoff argues that the next 12-18 months will disproportionately determine whether the ecosystem reaches a meaningful size in time. This is why funding now matters, even if a funding wave is coming, and it is a plausible argument for front-loading commitments. We recently saw Coefficient Giving announce an $825M funding bump for GiveWell’s recommendations for 2026 (to $1B, from the $175M they committed in December), citing expected growth in future funding:
This substantial increase is a one-off surge rather than a new steady state, driven most notably by our growing expectations of future giving from Good Ventures and other funders. While the future is uncertain, we currently expect the coming years to see significant increases in philanthropic funding for the kinds of programs GiveWell recommends. As a result, we and Good Ventures are accelerating spending now, both to fill cost-effective gaps today and to create scalable opportunities for future donors. Future funding is far from guaranteed; this is a bet on laying the groundwork for others to emerge and pick up the baton.
We think a similar surge is warranted in the AI space and that creating “scalable opportunities” requires funding that is early-stage, fast, and tolerant of failure. And the need won’t expire when the wave arrives: new projects will keep starting small.
As always when funding early-stage work, we should expect power-law-distributed payoffs: most of these grants will yield no or modest returns; what pays for the portfolio is the occasional home run (hits-based giving).
This niche needs many more funders, evaluators, and organizations, and the Transformative AI Fund exists to help fill the gap.
Focus areas
The fund will support work within transformative AI and its implications:
The primary focus will be on reducing global catastrophic risks through grants in technical AI safety and AI governance (including post-AGI governance), as well as areas that support these, such as field-building and forecasting.
Beyond catastrophic risk, we’re open to a wider set of areas that matter for the AI transition going well, including Flourishing Futures, Digital Sentience, and others such as those highlighted in EA in the age of AGI, though we expect these to be a small minority of grants. This has been informed by recent developments in the fields and in discussions, such as those highlighted during the EA Forum debate weeks.
For now, the fund will not fund biosecurity and nuclear beyond AI x bio and AI x nuclear, as we currently see fewer opportunities in these areas and have limited in-house expertise.
Team
We have been very impressed by the recent trajectory of the Animal Welfare Fund, which set an ambitious strategy and has raised significant funds to execute it. We believe that the primary reason for their success was Karolina Sarek stepping into a full-time role and hiring a kick-ass team.
We aspire to take the Transformative AI Fund on a similar trajectory and are committing significant resources to the project, with a new full-time team:
Lowe Lundin as full-time Head of Fund. He previously ran his own impact consultancy organization, led the AI grantmaking program at Astralis Foundation, co-founded Impact Academy (now SteadRise), and led the Meta Charity Funders. Before turning to impact work, he worked on AI in venture capital and finance.
Loïc Watine as Director of EA Funds since January 2026. He brings 16 years of experience in the NGO sector, most recently being Chief Research and Policy Officer at Innovations for Poverty Action, where he led teams working across research, policy, and grantmaking.
We’re hiring a full-time Associate Program Officer to work alongside Lowe. The primary (>80%) focus will be on grantmaking. The two things we care about most are agency and knowledge of the field: you take initiative without being asked, and you know the AI landscape well enough to have your own views on it. Read more and apply here.
We’re also in the final stages of hiring for a Senior Associate to run operations across all of EA Funds.
If we’re happy with the trajectory of the fund, we hope to make more hires in the not-too-distant future.
Caleb Parikh, the former LTFF lead, will serve as an advisor to the fund.
Fund principles
Early-stage funding
Given our purpose, we hope to be a go-to fund for anyone starting something new in the transformative AI space to get their project off the ground. This can be funding for an individual, a new organization, or an existing organization with a new project.
We will primarily fund early-stage projects, with grants typically below $150k. We expect grants above $300k to be rare and don’t think we’ll be making any grants above ~$1M for the foreseeable future, as we don’t think our comparative advantage will be with established organizations that are already well served by institutional funders. They serve a different part of the funnel and operate at a scale where evaluating smaller grants or replying within weeks is rarely practical.
Transparent by default
We’ll keep using the grants database for public grants. We’ll also publish regular public grant reports on our website and here on the Forum, with aggregated breakdowns by cause area, grant size, impact assessment, and more. We’ll also include representative write-ups of what we funded and why we chose to do so. We hope this is a good way to build trust and provide a forum for feedback and dialogue with donors and grantees. Initially, we plan to publish the reports on a quarterly basis.
Some impactful work requires privacy. When we conclude, in collaboration with the grantee, that their work should not be publicized, we’ll refrain from listing them in the database, but will include them in the aggregated reports.
Open applications with quick, reliable responses
Opportunities rarely arrive on a given schedule. We want to be able to pounce when they arise, as some opportunistic projects might not happen at all if funding can’t be secured quickly.
Applicants who can’t secure funding in time might need to choose safer options, e.g., pursuing a PhD or a secure but possibly less impactful job. People need to eat and keep the lights on.
We’re not going to put numbers in this post before we can stand behind them and recognize that nailing this will require time and effort. However, this is one of our top priorities, and we believe that a full-time team and dedicated work to building a well-oiled grantmaking pipeline will help achieve this. We’ll report concrete metrics in a future quarterly report.
Response timelines will be longer initially than in the steady state. As we’re onboarding a new team and building new processes through iterative experimentation, we appreciate your patience during the first few months, when things will be bumpier. You can expect the first quarterly report before the end of the year, which will provide an update on metrics, what we’ve done, and the remaining challenges. Once our quarterly reports are released, we think you should judge us by those reports rather than this post.
Why give to the Transformative AI Fund?
Why give through a fund?
We believe that for most individual donors, donating to a fund will achieve more with the same amount of money than they could on their own. Grantmaking has advantages that scale, which are hard for individual funders to replicate:
Deal flow. An individual is less likely to hear about many of the smaller opportunities that make up the fat tail. It’s more likely that they would see more visible organizations with comms teams and established networks, and miss things that have not yet gained traction or have not happened at all. Thus, solo donors without good deal flow are more likely to fund what’s easiest to find, which is disproportionately what least needs a marginal dollar.
Expertise. Seeing thousands of applications and tracking outcomes provides a better overview of the field and its people than most individuals can achieve. This allows coordination with other funders, seeing gaps, and RFPs to fill said gaps. Grantees benefit too: a funder who has seen many similar projects can make introductions and tell you what usually goes wrong.
Operations and tooling. A pipeline built for due diligence, payments, and reporting enables grants at relatively low marginal costs. Solo donors need to pay near-full overhead on each grant.
To be clear, some well-informed, well-connected donors with a good knowledge of the space may be in a position to identify opportunities that may be more promising than fund recommendations.
Why this fund?
We hope that giving to funds like the Transformative AI Fund can be a fairly attractive proposition for people who:
Want to support early-stage projects, at a size where funding is often all-or-nothing and many funders struggle to operate
Buy our argument on field absorption: if a funding influx is coming, the organizations that absorb it have to exist first
Accept power-law payoffs: most grants will return little or nothing; what pays for the portfolio is the occasional home run.
If you disagree with the above, that might be a good reason not to donate to us.
How to get involved
If you’re a donor
We’re actively fundraising. We have a seed grant of ~$1M, but hope to raise significantly more, as we believe there are sufficient opportunities in the field to deploy several times more. The fund is designed to enable donors of all sizes to support early-stage projects that address transformative AI, with no time commitment required. Donate here or reach out at lowe.lundin@centreforeffectivealtruism.org
If you’re starting something new
We’re especially excited about early-stage and small-but-promising work that falls below the threshold at which larger funders usually operate. You can apply here.
EA Funds is launching the Transformative AI Fund
TL;DR
The Long-Term Future Fund is closing down, and EA Funds is launching the Transformative AI Fund with a new full-time team.
The fund’s primary focus is technical AI safety and AI governance (including post-AGI governance), as well as supporting fields such as field-building and forecasting. We’ll also consider non-GCR implications of transformative AI such as flourishing futures and digital sentience, though we expect these to be a small minority of grants. For now, biosecurity and nuclear will only be supported at the AI intersection.
The fund will provide funding that is early-stage, fast, and tolerant of failure, to build the field’s absorptive capacity. We’re betting that a lot of philanthropic money is about to arrive, and that the field doesn’t have enough organizations to put it to work.
We aim to always be open to applications, respond quickly and reliably, and publish quarterly reports on what we fund and why.
Whether you’re an individual, a new organization, or an existing organization with a new idea, you can apply here.
We’re actively fundraising. We have a seed grant of ~$1M, but hope to raise significantly more, as we believe there are sufficient opportunities in the field to deploy several times more. The fund is designed to enable donors of all sizes to support early-stage projects that address transformative AI, with no time commitment required. Donate here or reach out at lowe.lundin@centreforeffectivealtruism.org
Background
The Long-Term Future Fund (LTFF) is closing down, as the outgoing team announced today in this post. Since LTFF’s identity has been tightly bound to the individuals running the fund, and the new fund’s scope will have substantial, but not exact, overlap, we have agreed with the team to wind down LTFF and present the Transformative AI Fund as a distinct fund with a new team. Since 2017, LTFF has made 820+ grants totaling just under $35M, and many great orgs, projects, and people have received their first check from the fund. We think LTFF has been highly impactful and are grateful to the team for their important work over the years as well as their help with the transition.
We think the niche LTFF has occupied (early-stage, small, and speculative projects) has been extremely valuable for seeding new organizations, and we aim to double down on it.
So: we’re launching the Transformative AI Fund, with a new full-time team.
Purpose
Over the coming years, this field is likely to receive far more money than it does today. In “The third wave of American philanthropy”, Nan Ransohoff estimates that AI wealth could add $37-100B per year in intended philanthropic spend, expects much of the early spending to focus on making the AI transition go well, and argues that at that scale the bottleneck shifts from money to good places for it to go. The field is not ready for that.
What’s scarce is organizations. Money gets absorbed by orgs that can hire people and run projects, and the field doesn’t have enough of them. There’s a lot of machinery for finding and developing people (fellowships, courses, research programs, upskilling grants) and much less for getting new organizations started.
Organizations need founders as much as they need money, and good founders are hard to find. But how many people attempt founding isn’t fixed: it depends partly on how costly the attempt is. For most people, starting something means months of uncertain fundraising weighed against a safe job, and reliable early capital changes that calculation. We can’t conjure founders, but we can lower the barriers to entry. Most new organizations start with small checks: a first $25k grant is often the difference between a pilot running or not. Many institutional funders don’t serve that size (for a number of good reasons).
We also think there’s room for more funding that stays open-ended about what it backs. A lot of early-stage funding starts from a funder’s view of what should exist and then looks for someone to build it, which works well for ideas already on someone’s radar. Taking applications from anyone, whenever opportunities arise, casts a wider net and finds projects nobody thought to ask for. We expect a meaningful share of our best grants to be things we wouldn’t have known to solicit.
On timing, larger checks to established organizations presuppose the initial small checks that got those organizations started, and organizations take time to build; Ransohoff argues that the next 12-18 months will disproportionately determine whether the ecosystem reaches a meaningful size in time. This is why funding now matters, even if a funding wave is coming, and it is a plausible argument for front-loading commitments. We recently saw Coefficient Giving announce an $825M funding bump for GiveWell’s recommendations for 2026 (to $1B, from the $175M they committed in December), citing expected growth in future funding:
We think a similar surge is warranted in the AI space and that creating “scalable opportunities” requires funding that is early-stage, fast, and tolerant of failure. And the need won’t expire when the wave arrives: new projects will keep starting small.
As always when funding early-stage work, we should expect power-law-distributed payoffs: most of these grants will yield no or modest returns; what pays for the portfolio is the occasional home run (hits-based giving).
This niche needs many more funders, evaluators, and organizations, and the Transformative AI Fund exists to help fill the gap.
Focus areas
The fund will support work within transformative AI and its implications:
The primary focus will be on reducing global catastrophic risks through grants in technical AI safety and AI governance (including post-AGI governance), as well as areas that support these, such as field-building and forecasting.
Beyond catastrophic risk, we’re open to a wider set of areas that matter for the AI transition going well, including Flourishing Futures, Digital Sentience, and others such as those highlighted in EA in the age of AGI, though we expect these to be a small minority of grants. This has been informed by recent developments in the fields and in discussions, such as those highlighted during the EA Forum debate weeks.
For now, the fund will not fund biosecurity and nuclear beyond AI x bio and AI x nuclear, as we currently see fewer opportunities in these areas and have limited in-house expertise.
Team
We have been very impressed by the recent trajectory of the Animal Welfare Fund, which set an ambitious strategy and has raised significant funds to execute it. We believe that the primary reason for their success was Karolina Sarek stepping into a full-time role and hiring a kick-ass team.
We aspire to take the Transformative AI Fund on a similar trajectory and are committing significant resources to the project, with a new full-time team:
Lowe Lundin as full-time Head of Fund. He previously ran his own impact consultancy organization, led the AI grantmaking program at Astralis Foundation, co-founded Impact Academy (now SteadRise), and led the Meta Charity Funders. Before turning to impact work, he worked on AI in venture capital and finance.
Loïc Watine as Director of EA Funds since January 2026. He brings 16 years of experience in the NGO sector, most recently being Chief Research and Policy Officer at Innovations for Poverty Action, where he led teams working across research, policy, and grantmaking.
We’re hiring a full-time Associate Program Officer to work alongside Lowe. The primary (>80%) focus will be on grantmaking. The two things we care about most are agency and knowledge of the field: you take initiative without being asked, and you know the AI landscape well enough to have your own views on it. Read more and apply here.
We’re also in the final stages of hiring for a Senior Associate to run operations across all of EA Funds.
If we’re happy with the trajectory of the fund, we hope to make more hires in the not-too-distant future.
Caleb Parikh, the former LTFF lead, will serve as an advisor to the fund.
Fund principles
Early-stage funding
Given our purpose, we hope to be a go-to fund for anyone starting something new in the transformative AI space to get their project off the ground. This can be funding for an individual, a new organization, or an existing organization with a new project.
We will primarily fund early-stage projects, with grants typically below $150k. We expect grants above $300k to be rare and don’t think we’ll be making any grants above ~$1M for the foreseeable future, as we don’t think our comparative advantage will be with established organizations that are already well served by institutional funders. They serve a different part of the funnel and operate at a scale where evaluating smaller grants or replying within weeks is rarely practical.
Transparent by default
We’ll keep using the grants database for public grants. We’ll also publish regular public grant reports on our website and here on the Forum, with aggregated breakdowns by cause area, grant size, impact assessment, and more. We’ll also include representative write-ups of what we funded and why we chose to do so. We hope this is a good way to build trust and provide a forum for feedback and dialogue with donors and grantees. Initially, we plan to publish the reports on a quarterly basis.
Some impactful work requires privacy. When we conclude, in collaboration with the grantee, that their work should not be publicized, we’ll refrain from listing them in the database, but will include them in the aggregated reports.
Open applications with quick, reliable responses
Opportunities rarely arrive on a given schedule. We want to be able to pounce when they arise, as some opportunistic projects might not happen at all if funding can’t be secured quickly.
Applicants who can’t secure funding in time might need to choose safer options, e.g., pursuing a PhD or a secure but possibly less impactful job. People need to eat and keep the lights on.
We’re not going to put numbers in this post before we can stand behind them and recognize that nailing this will require time and effort. However, this is one of our top priorities, and we believe that a full-time team and dedicated work to building a well-oiled grantmaking pipeline will help achieve this. We’ll report concrete metrics in a future quarterly report.
Response timelines will be longer initially than in the steady state. As we’re onboarding a new team and building new processes through iterative experimentation, we appreciate your patience during the first few months, when things will be bumpier. You can expect the first quarterly report before the end of the year, which will provide an update on metrics, what we’ve done, and the remaining challenges. Once our quarterly reports are released, we think you should judge us by those reports rather than this post.
Why give to the Transformative AI Fund?
Why give through a fund?
We believe that for most individual donors, donating to a fund will achieve more with the same amount of money than they could on their own. Grantmaking has advantages that scale, which are hard for individual funders to replicate:
Deal flow. An individual is less likely to hear about many of the smaller opportunities that make up the fat tail. It’s more likely that they would see more visible organizations with comms teams and established networks, and miss things that have not yet gained traction or have not happened at all. Thus, solo donors without good deal flow are more likely to fund what’s easiest to find, which is disproportionately what least needs a marginal dollar.
Expertise. Seeing thousands of applications and tracking outcomes provides a better overview of the field and its people than most individuals can achieve. This allows coordination with other funders, seeing gaps, and RFPs to fill said gaps. Grantees benefit too: a funder who has seen many similar projects can make introductions and tell you what usually goes wrong.
Operations and tooling. A pipeline built for due diligence, payments, and reporting enables grants at relatively low marginal costs. Solo donors need to pay near-full overhead on each grant.
To be clear, some well-informed, well-connected donors with a good knowledge of the space may be in a position to identify opportunities that may be more promising than fund recommendations.
Why this fund?
We hope that giving to funds like the Transformative AI Fund can be a fairly attractive proposition for people who:
Want to support early-stage projects, at a size where funding is often all-or-nothing and many funders struggle to operate
Buy our argument on field absorption: if a funding influx is coming, the organizations that absorb it have to exist first
Accept power-law payoffs: most grants will return little or nothing; what pays for the portfolio is the occasional home run.
If you disagree with the above, that might be a good reason not to donate to us.
How to get involved
If you’re a donor
We’re actively fundraising. We have a seed grant of ~$1M, but hope to raise significantly more, as we believe there are sufficient opportunities in the field to deploy several times more. The fund is designed to enable donors of all sizes to support early-stage projects that address transformative AI, with no time commitment required. Donate here or reach out at lowe.lundin@centreforeffectivealtruism.org
If you’re starting something new
We’re especially excited about early-stage and small-but-promising work that falls below the threshold at which larger funders usually operate. You can apply here.
If you’re interested in working with us
Apply to the Associate Program Officer role here.
Express your interest in future roles here.