Deals & IPOs

Nvidia in Advanced Talks to Buy AI Startup Inception for $500M-Plus

Nvidia is reportedly near a $500M-plus deal for diffusion-LLM startup Inception, whose Mercury API shuts down December 28 — below the $1B valuation the startup sought in May.

By Olivia Hart

4 min read

Nvidia Reported in Advanced Talks to Acquire Diffusion AI Startup Inception for $500M+
Nvidia Reported in Advanced Talks to Acquire Diffusion AI Startup Inception for $500M+StartupStockPhotos / pixabay

What's News

  • Nvidia is in advanced talks to acquire Inception for at least $500 million plus earnouts, per The Information's November 19 report.
  • The reported price sits below the >$1 billion valuation Reuters said Inception sought in May 2026.
  • Inception's Mercury API shuts down December 28, 2026, per a November 12 customer notice.
  • Nvidia's board accepted Jensen Huang's resignation as CEO on November 11, 2026; CFO Colette Kress is interim CEO.
  • Inception raised a $50 million seed round in November 2025 led by Menlo Ventures, with NVentures and Microsoft's M12 participating.

By Olivia Hart

Nvidia is in advanced talks to acquire Inception, a two-year-old startup building diffusion-based large language models, for at least $500 million plus performance earnouts, according to a November 19 report from The Information citing three people familiar with the discussions. Neither company has confirmed the deal. The terms, per the report, could still change.

The reported price carries a discount. It is roughly ten times Inception's $50 million seed round — but notably below the greater-than-$1 billion valuation Reuters said the startup sought in May 2026, when Microsoft and SpaceX were also reported to be interested. For context, $500 million is a tenth of a billion, a sobering figure for a company that reportedly aimed ten times higher six months ago.

What does Inception actually build?

Inception, founded by Stanford professor Stefano Ermon (CEO), UCLA professor Aditya Grover and Cornell professor Volodymyr Kuleshov, develops "diffusion large language models" (dLLMs) under the Mercury brand. Conventional autoregressive models generate text token by token. Diffusion models start with a rough full draft and refine all tokens in parallel over multiple steps — the same principle behind image generators such as Stable Diffusion and Sora.

Inception claims its Mercury models run 5–10× faster than speed-optimized models from OpenAI, Anthropic and Google while matching their accuracy. The company's own benchmarks show Mercury 3 exceeding 1,000 tokens per second with a 1 million-token context window. Those speed claims are vendor-reported and have not been independently reproduced across workloads — a caveat every buyer, including Nvidia, presumably has to weigh.

The startup raised its $50 million seed round in November 2025, led by Menlo Ventures with participation from Mayfield, Innovation Endeavors, Microsoft's M12, Snowflake Ventures, Databricks Investment and Nvidia's venture arm NVentures. That last name matters: the chipmaker already held a financial stake in the startup it is now reportedly trying to buy.

How does this fit Nvidia's model strategy?

The talks fit a pattern of Nvidia moving beyond hardware into model ownership. Inception's technology would feed Nvidia's Nemotron portfolio of open models, already used by Booz Allen Hamilton, Palantir Technologies and several U.S. government agencies. The Information's report, relayed by The Decoder, says Inception's team could also contribute to Nvidia's "physical AI" models for robotics, where diffusion methods are already common in research.

Nvidia has invested in AI labs and model makers across the year through NVentures and its separately named Nvidia Inception startup program. If the deal closes, Nvidia would hold a distinct architecture — diffusion — alongside its autoregressive Nemotron line, plus a research team with roots in the invention of diffusion methods.

What do the deal terms and timing reveal?

The package includes earnouts for Inception's 30-plus employees tied to performance targets. All three co-founders would join Nvidia full-time.

The timing is unusual. The Information reports that talks began before Nvidia's board accepted Jensen Huang's resignation as CEO. An SEC filing shows the board accepted Huang's resignation on November 11, 2026, effective immediately, with CFO Colette Kress appointed interim CEO while the board searches for a permanent replacement. Microsoft's M12 participated in Inception's seed round, and The Information says Microsoft also expressed acquisition interest — though whether Microsoft made a formal offer is unconfirmed.

Inception separately notified customers on November 12 that its Mercury API will shut down December 28, 2026, with no mention of a buyer. How that shutdown relates to the reported talks is unclear.

What remains unconfirmed?

Plenty. All details stem from a single paywalled report relayed through secondary outlets. No confirmation has come from either Nvidia or Inception. The report does not say whether Mercury will remain available, be folded into Nemotron or be discontinued. A half-billion-dollar Nvidia acquisition could draw antitrust attention; no filings or comments have surfaced on that front either.

The practical takeaway for Mercury API users: plan for the December 28 shutdown regardless of the acquisition outcome, export data now and line up alternatives. For builders evaluating diffusion LLMs in latency-sensitive work, vendor speed claims are not a substitute for testing on your own prompts and hardware.

If the deal closes, Nvidia gains a differentiated architecture and a founding team drawn from the inventors of the method. Whether that translates into a shipping advantage remains to be seen.

Source: epyhia.com

Share this article:

More from Olivia Hart

Olivia Hart

Show full bio

Staff writer covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.

2,535 articles

Related articles

« Previous article