Nvidia, which dominates the AI chip market, may be moving from selling the "picks and shovels" of AI to owning the models themselves. The Financial Times reported on Saturday, Oct. 10, 2026, citing people with direct knowledge, that Nvidia is in early talks to acquire Reflection AI or deepen its investment in the startup. Reflection is a US developer of open-weight AI models, last valued at $25 billion, and Nvidia is already one of its largest shareholders. However, the $25 billion figure reflects Reflection's earlier funding round, not a price agreed between the two companies, and sources said the talks are at an early stage and could still fall apart. Neither Nvidia nor Reflection has commented publicly.
Nvidia's Talks With Reflection AI Could Take Three Forms
According to the report, the deal could take several forms. The first is a full acquisition. The second is a larger stake, possibly combined with a bigger supply of chips and computing power. The third is an acqui-hire, in which Nvidia would hire Reflection's staff and license its technology without taking control of the company, a structure that would likely avoid a lengthy regulatory review. Nvidia used that approach with AI chip developer Groq in December, licensing its technology and hiring its CEO in a deal CNBC valued at about $20 billion. Sources said an agreement with Reflection could be reached within weeks.
Who Is Reflection AI, and Why Is It Worth $25 Billion?
Reflection AI was founded in 2024 by Misha Laskin and Ioannis Antonoglou, former Google DeepMind researchers who worked on the Gemini model. The company builds open-weight models for coding and AI agent tasks, one of the fastest-growing areas of AI. Nvidia led Reflection's $2 billion round in October 2025 and has invested $800 million in total. In April, Laskin told CNBC that the latest round closed at a $25 billion pre-money valuation. Reflection has raised about $4.7 billion in total from investors including Nvidia, Sequoia Capital, Lightspeed and DST Global. It also has partnerships with US defense and energy agencies and agreements to develop AI models for countries including South Korea.
Beam, Its First Open-Weight Model, Takes On Chinese AI
The talks come just days after Reflection launched Beam, its first open-weight model, on Monday, Oct. 5, aiming to compete with low-cost Chinese models such as DeepSeek and Kimi in coding and agentic tasks. Beam has 501 billion total parameters but activates only 23 billion per task, making it faster and cheaper to run. Reflection says Beam is competitive with Z.ai's GLM-5.2 and close to Alibaba's Qwen3.8-Max. The model was pretrained on 23.8 trillion tokens, supports a 1-million-token context window, and used 10,500 Nvidia GB300 GPUs for its reinforcement learning run. The launch matters for policy too, since parts of the US government see American open-weight models as an important alternative to cheaper Chinese AI.
Why Nvidia Wants Reflection
For Nvidia, turning a financial stake into ownership or deeper control of Reflection would extend its reach from chips into the models that run on them. CEO Jensen Huang said on the company's latest earnings call that nearly all open models run on Nvidia and that open models are doing incredibly well. The more capable, low-cost open models there are, the more developers build on them, and most of that work runs on Nvidia chips. Reflection is also a major buyer of Nvidia compute through deals with SpaceX and Nebius worth more than $7 billion to secure GB300 chips through 2029. A deal would therefore both support chip demand and align with US efforts to export American AI.
What It Means for Investors: Nvidia Can Afford It, but the Structure Is the Signal
Financially, a deal of this size is not a heavy lift for Nvidia, which generated about $127 billion in free cash flow over the past four quarters. A $25 billion check equals roughly 20% of that, or about 10 weeks of cash generation. What investors should watch more closely is the structure. A full buyout would be a clear signal that Nvidia intends to enter the model business, potentially competing with major customers such as OpenAI, which Nvidia also funds, and inviting tougher antitrust scrutiny. An acqui-hire could avoid some merger-review requirements but would not eliminate competition scrutiny entirely. Key things to watch are official announcements from either company, the final deal structure, and how regulators respond.
