What Happened to Reflection AI?

Reflection AI, Inc. was founded in 2024 by former Google DeepMind researchers Ioannis Antonoglou and Misha Laskin, with headquarters in Brooklyn, New York. The company's core mission is to build super-intelligent autonomous systems, starting with autonomous coding agents that can automate software development processes, allowing AI to reason over and evolve code with a high degree of autonomy. They aim to automate the knowledge work done on a computer and believe that solving autonomous coding will enable broader superintelligence.

Reflection AI emerged from stealth in March 2025, announcing $130 million in early-stage funding, comprising a $25 million seed round led by Sequoia Capital and CRV, and a $105 million Series A round co-led by Lightspeed Venture Partners and Sequoia Capital. This initial funding valued the company at approximately $545 million. The company quickly gained traction, being named a 'Next Billion-Dollar Startup' by Forbes in August 2025.

A significant turning point came in October 2025, when Reflection AI closed a $2 billion Series B funding round, primarily led by Nvidia, which propelled its valuation to $8 billion. This round also saw participation from prominent investors like Eric Schmidt, Citi, and 1789 Capital. Following this, in March 2026, Reflection AI began discussions to raise an additional $2.5 billion in a Series C round, targeting a pre-money valuation of $25 billion. JPMorgan Chase was reportedly in active discussions to participate through its Security and Resiliency Initiative, highlighting the strategic importance of AI infrastructure to national security.

Despite its meteoric rise in valuation and substantial funding, Reflection AI has faced scrutiny regarding its public product releases. As of early March 2026, its frontier open-weight model had not been publicly released, and its code research and comprehension agent, Asimov, remained on a waitlist. Critics have pointed to the significant gap between the company's $25 billion valuation and its minimal public output, raising questions about the current state of AI investing. However, investors are betting on the founders' track record and the company's strategic positioning in the burgeoning 'sovereign AI' market, which emphasizes national control over AI infrastructure.

In 2026, Reflection AI has continued to forge strategic partnerships. In March, it collaborated with South Korean conglomerate Shinsegae Group to build a 250-megawatt AI factory in Korea, utilizing NVIDIA GPUs for sovereign AI cloud initiatives, with support from both U.S. and Korean governments. Most recently, on June 22, 2026, SpaceX signed a computing agreement with Reflection AI, valued at up to $6.3 billion, for the use of Nvidia GB300 chips at SpaceX's Colossus 2 facility. This deal further solidifies Reflection AI's access to critical computing resources and its role in the broader AI infrastructure landscape. The company is also establishing a lobbying and policy communications operation in Washington D.C., opening an office in August 2026, to advocate for U.S. open-source AI as a national security imperative.