Nvidia in advanced talks to acquire diffusion-AI startup Inception for more than $500 million, The Information reports
The Information, 19 November 2026: Nvidia is in advanced talks to buy Inception, a two-year-old startup building diffusion-based large language models, for more than $500 million. Neither company had confirmed a deal at the time of writing; details below are as reported, not independently verified.
The Information reported on 19 November 2026 that Nvidia is in advanced talks to acquire Inception, a startup developing diffusion-based large language models, for more than $500 million. Follow-on coverage by outlets including TipRanks, Newsquawk and Dealroom repeats the same key terms, all citing The Information. Nvidia and Inception had not publicly confirmed the talks at the time of writing.
Who Inception is
Inception is a San Francisco Bay Area startup founded about two years ago by three academic researchers: chief executive Stefano Ermon, a professor at Stanford University; Aditya Grover, a professor at UCLA; and Volodymyr Kuleshov, a professor at Cornell University. According to TipRanks' summary of the report, the company has more than 30 employees and has raised $50 million from investors including Menlo Ventures, M12, Mayfield and Nvidia itself. All three co-founders are expected to join Nvidia if the transaction completes, and the reported terms may include additional earnout payments tied to performance targets.
What diffusion-based language models do differently
Conventional large language models generate text one token at a time, in sequence. Inception's models instead refine a response through an iterative process in which many tokens are produced in parallel, a method adapted from diffusion techniques used in image generation. The company's Mercury family of models is built on this approach, and Inception claims its models can generate text up to 10 times faster than comparable models; that figure is the company's own claim, not an independent benchmark. The startup has also published work on diffusion-based coding models and on multimodal understanding and reasoning.
Why Nvidia would want the deal
The reported acquisition would give Nvidia access to new research talent and technology as it expands its own AI model work, including the Nemotron portfolio of open models that businesses and government agencies can customize for specific applications. Inception's research team is also expected to explore diffusion techniques for physical-AI and robotics models, an area Nvidia has been pushing into beyond data centers. The Information's report says Microsoft had also shown interest in a similar deal, and that the talks began before the leadership change at Nvidia earlier this month.
What is not confirmed
The reported price above $500 million, the possible earnout structure and the founders' expected move to Nvidia all come from sources cited by The Information. No company statement, signed agreement or regulatory filing had been published at the time of writing, and advanced talks do not guarantee a completed deal. We will update this story if Nvidia or Inception confirms, changes or denies the report.
- The Information: Nvidia in Advanced Talks to Buy Diffusion-Model Startup Inception for at Least $500 Million (19 November 2026)
- TipRanks: Nvidia Nears $500M Deal for AI Startup Inception to Strengthen AI Models
- Newsquawk: Nvidia in advanced talks to acquire diffusion AI startup Inception for over $500m