Nvidia Weighs Buying Open-Weight Start-Up Reflection AI
Trending Topics Jakob Steinschaden ● Covered by 2 sources
Nvidia is talking about buying Reflection AI or putting in more money. It could give the chip giant a bigger grip on the models themselves, not just the hardware.
Based on reporting by Trending Topics, Jakob Steinschaden — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Nvidia may be trying to move up the stack, from selling the picks and shovels of AI to owning more of the gold mine. The company is in talks to acquire Reflection AI or raise its stake in the start-up, according to the Financial Times, as relayed by Reuters. The discussions are early, a deal could happen within weeks, and they could still collapse.
The FT says three paths are being discussed: a full buyout, an acqui-hire, or a bigger investment. That middle option would let Nvidia hire Reflection’s staff and license the technology without buying the company outright. It would also likely be easier to clear than a classic merger review. Nvidia has done something similar before, when it brought in the team behind Groq and licensed its technology in a deal CNBC valued at about $20 billion.
Nvidia is already deeply involved with Reflection. The FT says it has committed $800 million, and the start-up has raised close to $2.6 billion overall from backers including Nvidia, Sequoia Capital and Lightspeed. In April, chief executive Misha Laskin told CNBC that Reflection was raising fresh money at a $25 billion pre-money valuation.
Reflection itself is young, having been founded in 2024 by Laskin and Ioannis Antonoglou, both former Google DeepMind researchers. Its first open-weight model, Beam, is aimed at coding and AI agents and is meant to take on lower-cost Chinese models like DeepSeek and Kimi. A deal would also give Nvidia a bigger model business, alongside its own open Nemotron models. But the antitrust angle is the one to watch. Nvidia already dominates AI chips, and it keeps investing in the companies that need them.
And the cash side is not the problem. Over the last four quarters, Nvidia produced about $127 billion in free cash flow. The real question is whether regulators decide that owning more of the AI stack looks too much like one company trying to hold both the faucet and the plumbing.
My take — AI-written commentary, not fact-checked reporting
This is what happens when the biggest chip seller decides chips are not enough. Nvidia keeps reaching for the layer above, and regulators should treat that as more than a cute acqui-hire trick with a nicer name. The market loves vertical control until one company starts collecting the whole stack like baseball cards.
Read more about this at: Trending Topics
Related stories
Nvidia nears $12.9 billion deal to buy open-source AI platform, Hugging Face, report says
Fortune ·
11
Open-weight AI companies are the Valley’s hottest acquisition targets
TechCrunch · 1 month ago ·
26
Report: Nvidia to acquire AI model repository Hugging Face for $13 billion
Ars Technica · 1 month ago ·
9