TL;DR: MATS Winter 2027 applications are now open. It's a fully-funded, 12-week research fellowship (Jan 19 – Apr 10, 2027) in AI alignment, security, and governance, with mentorship from researchers at Anthropic, Google DeepMind, OpenAI, Redwood Research, AI Futures Project, and more. Fellows receive a $6,400/month stipend and a compute budget of up to $8,000/month for non-technical tracks and up to $16,000/month for technical tracks, as well as housing, meals, and travel.
MATS Research is an educational research nonprofit dedicated to solving the talent pipeline bottleneck in AI alignment and security research. We believe reducing risks from powerful AI is one of the world's most urgent and talent-constrained challenges, and that ambitious people from a wide range of backgrounds and career stages can meaningfully contribute to this work. That's why we're training the next generation of AI safety researchers and founders.
Program details
The Winter 2027 cohort runs from January 19 to April 10, 2027, based primarily in Berkeley, CA and London, UK. Fellows receive:
$6,400/month stipend + compute budget of up to $8,000/month (non-technical streams) or up to $16,000/month (technical streams)
Access to office spaces in Berkeley or London (depending on mentor preference)
Housing, meals, and travel covered
J1 visa support, if needed
Mentorship from world-class researchers and a dedicated research manager
A close-knit cohort, regular seminars and workshops with industry experts, and an active global alumni network
Opportunity to receive a 6–12 month extension with continued mentorship, support, and funding ($7,680/mo stipend and continued compute support, secured by more than 80% of candidates who apply.
Research tracks
Applicants can apply to one or more of the following tracks. Each track page describes the research agenda, the mentors involved, and what we're looking for in applicants. Before submitting an application, prospective fellows are encouraged to review the track pages below:
The Founding and Field-Building track for founders, field-builders, and high-agency generalists looking to launch new AI safety initiatives.
The Biosecurity track focused on preventing catastrophic biological risk from AI.
The Policy and Governance track focused on supporting research on how advanced AI is governed and how it should be governed.
Our track record
631 alumni, 100+ mentors
215+ publications with 17,000+ citations
75% of alumni who graduated before 2026 are working in AI safety/security
10% have co-founded active AI safety startups
30+ initiatives have been founded by alumni
Who we're looking for
MATS explicitly looks for talent that traditional pipelines might overlook. We welcome technical researchers without prior ML experience who can demonstrate strong reasoning and research potential. We also encourage applications from policy professionals with strong writing skills, familiarity with governmental processes, and the technical literacy to engage with AI systems–particularly those with backgrounds in national security, cybersecurity, US-China relations, biosecurity, and/or nuclear policy.
Remote participation
While we prefer fellows to participate in-person from Berkeley or London, we understand this may not always be feasible and are open to remote participation for exceptional candidates on a case-by-case basis.
TL;DR: MATS Winter 2027 applications are now open. It's a fully-funded, 12-week research fellowship (Jan 19 – Apr 10, 2027) in AI alignment, security, and governance, with mentorship from researchers at Anthropic, Google DeepMind, OpenAI, Redwood Research, AI Futures Project, and more. Fellows receive a $6,400/month stipend and a compute budget of up to $8,000/month for non-technical tracks and up to $16,000/month for technical tracks, as well as housing, meals, and travel.
Please apply by September 6, 2026 AoE at matsprogram.org/apply.
About MATS
MATS Research is an educational research nonprofit dedicated to solving the talent pipeline bottleneck in AI alignment and security research. We believe reducing risks from powerful AI is one of the world's most urgent and talent-constrained challenges, and that ambitious people from a wide range of backgrounds and career stages can meaningfully contribute to this work. That's why we're training the next generation of AI safety researchers and founders.
Program details
The Winter 2027 cohort runs from January 19 to April 10, 2027, based primarily in Berkeley, CA and London, UK. Fellows receive:
Research tracks
Applicants can apply to one or more of the following tracks. Each track page describes the research agenda, the mentors involved, and what we're looking for in applicants. Before submitting an application, prospective fellows are encouraged to review the track pages below:
Featured tracks for this cohort
Our track record
Who we're looking for
MATS explicitly looks for talent that traditional pipelines might overlook. We welcome technical researchers without prior ML experience who can demonstrate strong reasoning and research potential. We also encourage applications from policy professionals with strong writing skills, familiarity with governmental processes, and the technical literacy to engage with AI systems–particularly those with backgrounds in national security, cybersecurity, US-China relations, biosecurity, and/or nuclear policy.
Remote participation
While we prefer fellows to participate in-person from Berkeley or London, we understand this may not always be feasible and are open to remote participation for exceptional candidates on a case-by-case basis.
📖 Learn more about the MATS program
🔗 Apply by September 6, 2026 AoE
Please share this post with people you know who'd be a strong fit.