Nvidia is in talks to buy Reflection AI or invest more in the US open-weight model developer, the Financial Times reported on Saturday, citing people with direct knowledge of the matter. TNW has not independently verified the report.
The talks are at an early stage and could still fall apart, according to the FT. The two sides could reach an agreement in the coming weeks, the paper said. Arash Massoudi, George Hammond, Michael Acton and George Steer wrote the report.
An acqui-hire is one option
A deal could take several forms, the FT reported. One is a so-called acqui-hire. Nvidia would hire Reflection’s staff and license its technology instead of buying the company outright. That route could avoid a lengthy regulatory review.
Other options include supplying the startup with more chips or raising Nvidia’s stake. Nvidia has already invested $800m in Reflection, the FT reported.
The Trump administration hopes Reflection will rival cheap Chinese open models such as DeepSeek, according to the FT.
Bloomberg reported separately that Nvidia is in talks to boost its investment in Reflection or offer it computing power, citing a person familiar with the matter. The discussions are early and the two sides have not agreed terms, that person said.
Weeks after Beam
Misha Laskin and Ioannis Antonoglou, both former Google DeepMind researchers, founded Reflection in 2024. On 5 October it released its first model, Beam, an open-weight model aimed at coding and agent tasks.
The startup has also signed compute deals with Elon Musk’s SpaceX and cloud provider Nebius.
Nvidia’s other deals
In September, Nvidia agreed to buy Hugging Face for $12.93bn. It also struck a licensing deal with the chip startup Groq, whose chief executive and chief operating officer joined Nvidia.
The Justice Department is investigating the Groq deal, The New York Times reported in September. It is checking whether Nvidia structured the deal to avoid antitrust review. An Nvidia spokesman told the Times that the Groq deal showed “the American system working as designed”.
Published
Back to top