
What happened
Nvidia is in advanced talks to buy Inception, a two-year-old diffusion-based language-model startup, for more than $500 million plus earnouts tied to post-close hiring targets, per The Information.
Why it matters
Inception's parallel-refinement architecture, which it claims runs 10x faster and at 1/10th the cost of traditional LLMs, is expected to strengthen Nvidia's Nemotron open-weight models and robotics AI, though neither company has confirmed the deal.
WHO IT HITSThe deal would bring Stanford's Stefano Ermon, UCLA's Aditya Grover and Cornell's Volodymyr Kuleshov — and their diffusion methods — inside Nvidia, affecting teams building open-weight and robotics AI on the Nemotron stack.
Summaries like this, every morning.
The target is young and small but technically distinctive. Inception is described as a two-year-old startup whose architecture refines an entire response in parallel rather than one token at a time — the approach behind its flagship Mercury model. That is a different bet from the token-by-token generation that powers products such as ChatGPT and Claude, and Inception claims it runs 10x faster and at 1/10th the cost of traditional LLMs.
The founding team carries weight beyond the startup. CEO Stefano Ermon is a co-inventor of the diffusion methods widely used in modern image and video generation, and he is expected to join alongside UCLA's Aditya Grover and Cornell's Volodymyr Kuleshov. Inside Nvidia, the technology is expected to feed the company's Nemotron open-weight models and its AI systems for robotics.
The structure of the deal is itself notable. The earnouts are pegged to whether new employees meet hiring targets after close, an arrangement unusual enough to flag, and the $500 million figure is a floor rather than a final price. Neither Nvidia nor Inception has publicly confirmed the talks.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Oxide Computer Co., which raised $445 million in October to expand rack production, built its business on the…
OpenAI now requires gateways and routers that provide third-party access to its AI models to be reviewed and a…

Anthropic shares traded as low as $83.15 on Friday, down more than 14% on their second day of trading, leaving…

Running the official harness, Claude Opus 5.5 wrote an ARC model with no gradient descent and no forward calls…

A Zenn article by Hiromi lays out Continuous Reasoning Acquisition (CRA), a design for capturing veteran judgm…

Major US tech companies are refocusing generative AI competition from answer quality toward agent capability…
