AudaciousMake intelligence open and accessible to all.
Reflection AI is building open frontier models, releasing open weights, publishing the research behind them, and open-sourcing the tools to customize and build with them. Founded by former Google DeepMind researchers, including a core developer of AlphaGo, the company is positioning itself as America’s open frontier AI lab.
Misha Laskin
Former Google DeepMind researcher, working on reinforcement learning and large-scale model training before starting Reflection.
Ioannis Antonoglou
Former Google DeepMind researcher and a core developer behind AlphaGo.
Introducing Beam.
Beam is an agentic open model with 501B total parameters and 23B active, trained end to end from scratch.
Reflection is pitching it on frontier reasoning efficiency, and on advancing the Western open frontier for coding and agentic tasks. Full weights release this month.
Open frontier AI, already landing sovereign and enterprise deals.
Reflection is a research lab making intelligence open and accessible for everyone to use, customize, and build on, the open alternative to closed frontier labs, with model weights, research, and tooling all released openly so anyone can control their own intelligence. That openness is already translating into large, strategic commercial and government partnerships.
Open beats closed
Open systems evolve faster, attract broader participation, and are safer, since more researchers can inspect and improve them.
Dell, the DOE, and Shinsegae
Delivering AI factories for enterprises with Dell, powering U.S. national labs under the Genesis Mission, and building Korea’s sovereign AI cloud with Shinsegae.
From $545M to $25B in a year
Emerged from stealth at a ~$545M valuation in March 2025; raised $2B at an $8B valuation by October 2025; and reached a $25B valuation with a ~$2.5B raise by March 2026.
Backed by top-tier investors.
A $25B valuation as of March 2026, with Nvidia and SpaceX among the strategic backers and compute partners.
~$545M valuation
Emerged from stealth as America’s open frontier AI lab.
$8B valuation
Raised $2B, backed by strategic investors and compute partners.
$25B valuation
A ~$2.5B raise, roughly a 3× step-up in five months.
Recent coverage.
Reflection AI and the Genesis Mission energy partnership
Read on Axios →
Reflection’s compute deal to challenge China’s open-source dominance
Read on Forbes →
SpaceX lands a $6.3B compute deal with Reflection, with Nvidia on both sides of the trade
Read on TechFundingNews →
Reflection raises $2B to be America’s open frontier AI lab, challenging DeepSeek
Read on TechCrunch →
What they’re hiring for.
Reflection is hiring across its Applied AI team and its Engineering, Foundations, Platform, Runtime, and Security orgs. Here is the full list of open roles.
- Engagement ManagerSan Francisco / New York · On-site
- Forward Deployed Engineer — AI EngineerNew York / San Francisco · On-site
- Forward Deployed Engineer — AI EngineerSeoul · On-site
- Forward Deployed Engineer — LLM Post-trainingSan Francisco / New York · On-site
- Forward Deployed Engineer Lead, Post-Training (LLM)New York / San Francisco · On-site
- Forward Deployed Engineer, Lead — AI EngineerSeoul · On-site
- Forward Deployed Engineer, Lead — AI EngineerNew York / San Francisco · On-site
- MTS — ApplicationsNew York / London / San Francisco · On-site
- MTS — Data PlatformNew York / London / San Francisco · On-site
- MTS — Distributed Systems EngineerNew York / London / San Francisco · On-site
- MTS — Engineering Lead, Compute PlatformSan Francisco / London / New York · On-site
- MTS — Engineering Lead, Platform FoundationsSan Francisco / London / New York · On-site
- MTS — Engineering Lead, Shared ServicesSan Francisco / London / New York · On-site
- MTS — Platform FoundationsNew York / London / San Francisco · On-site
- MTS — Compute PlatformNew York / London / San Francisco · On-site
- MTS — Engineering Lead, ApplicationsSan Francisco / London / New York · On-site
- MTS — Engineering Lead, Compute PlatformSan Francisco / London / New York · On-site
- MTS — Research Software Engineer, Safety Evaluations InfraNew York / London / San Francisco · On-site
- MTS — Mid-Training InfraSan Francisco / London / New York · On-site
- MTS — Pre-Training InfraSan Francisco / London / New York · On-site
- MTS — IT EngineeringNew York / London / San Francisco · On-site
- MTS — Infrastructure SecuritySan Francisco · On-site
- MTS — Security EngineerSan Francisco · On-site
MTS = Member of Technical Staff · all roles full-time, on-site.
Build the deployments behind open intelligence.
A frontier AI lab going from stealth to a $25B valuation in about a year, already landing sovereign and enterprise partnerships with Dell, the U.S. Department of Energy, and Shinsegae. Forward Deployed Engineering is how those deployments happen.