What is the real per-recipient economic value?

USD 2,000,000 maximum. Award minimum: 100000. Dilution/equity: none.

Which jurisdictions and applicants are eligible?

Academic labs, nonprofits and individual researchers could request $100k–$2m; graduate students could apply for the separate one-year $150k fellowship. No prior alignment experience was required.

Technical research on weak-to-strong generalization, interpretability or scalable oversight; A tractable project capable of progress quickly; Researchers new to alignment were explicitly welcomed

What burden or follow-on should I expect?

Research-output and award terms were in the historical application; no current reporting contract is public. No current renewal or successor pathway published.

OpenAI Superalignment Fast Grants

USD 2,000,000 maximum; see notes for structure · OpenAI (with Eric Schmidt)

$10M grant program (2023–2024) supporting technical research toward alignment and safety of superhuman AI systems — weak-to-strong generalization, interpretability, scalable oversight, and more. Current status: closed; One historical call launched December 2023 and closed 18 February 2024; no successor call is published. Award economics: USD 2,000,000 maximum; dilution/equity: none.

Tier 2 · ranked #1aigrantai-safetyalignment

Apply Official site Open in the live directory

Key facts

Funder	OpenAI (with Eric Schmidt)
Amount	USD 2,000,000 maximum; see notes for structure
Decision time	OpenAI said applicants would hear within four weeks after applications closed.
Eligibility	Academic labs, nonprofits and individual researchers could request $100k–$2m; graduate students could apply for the separate one-year $150k fellowship. No prior alignment experience was required.
Restrictions	Research-output and award terms were in the historical application; no current reporting contract is public. Dilution/equity: none.
Status	closed

Always confirm on the official page before applying — dates and terms change.

Application process

Choose a listed or adjacent technical superalignment direction
Submit the simple historical application by the closed deadline
Program review
Decision within four weeks after close
Execute the funded research or fellowship

Background

OpenAI (with Eric Schmidt) operates OpenAI Superalignment Fast Grants. The official-source pass separates the currently actionable opportunity from historical cohorts and separates per-recipient value from headline program envelopes. The program committed $10m total; the archived page names OpenAI and Eric Schmidt as partners.

How the application really works

Choose a listed or adjacent technical superalignment direction → Submit the simple historical application by the closed deadline → Program review → Decision within four weeks after close → Execute the funded research or fellowship

Tips

Demonstrate: Technical research on weak-to-strong generalization, interpretability or scalable oversight
Demonstrate: A tractable project capable of progress quickly
Demonstrate: Researchers new to alignment were explicitly welcomed

Watch out for

Current status is closed; do not use an archived form.
Dilution/equity treatment: none.
Research-output and award terms were in the historical application; no current reporting contract is public.

Track record

The program committed $10m total; the archived page names OpenAI and Eric Schmidt as partners.

No applicant or award count is published on the checked official page.

Frequently asked questions

Is this opportunity actionable now?: closed. One historical call launched December 2023 and closed 18 February 2024; no successor call is published.
What is the real per-recipient economic value?: USD 2,000,000 maximum. Award minimum: 100000. Dilution/equity: none.
Which jurisdictions and applicants are eligible?: Academic labs, nonprofits and individual researchers could request $100k–$2m; graduate students could apply for the separate one-year $150k fellowship. No prior alignment experience was required.
How does selection work?: Choose a listed or adjacent technical superalignment direction; Submit the simple historical application by the closed deadline; Program review; Decision within four weeks after close; Execute the funded research or fellowship. OpenAI said applicants would hear within four weeks after applications closed.
What appears to win?: Technical research on weak-to-strong generalization, interpretability or scalable oversight; A tractable project capable of progress quickly; Researchers new to alignment were explicitly welcomed
What burden or follow-on should I expect?: Research-output and award terms were in the historical application; no current reporting contract is public. No current renewal or successor pathway published.

Sources

openai.com/index/superalignment-fast-grants — Official historical amount, eligibility, topics, decision promise and deadline (checked 2026-08-10)