Nvidia is in advanced talks to acquire Inception, a two-year-old Palo Alto startup developing diffusion-based large language models. The proposed transaction is valued at more than US$500 million. It would bring the company's three co-founders into Nvidia full-time.
Funding so far
Inception raised a US$50 million Seed round in November 2025. Menlo Ventures led the round. Mayfield, Innovation Endeavors, Microsoft's M12, Snowflake Ventures, Databricks Investment, and Nvidia's NVentures participated.
What Inception builds
Inception develops diffusion-based language models that generate and refine multiple tokens in parallel. Its Mercury models use iterative refinement to speed up text generation. The company claims Mercury is 5 to 10 times faster than speed-optimized models from OpenAI, Anthropic, and Google.
Nvidia's wider AI push
The potential deal fits Nvidia's push beyond chips. Nvidia agreed to acquire Hugging Face for about US$12.9 billion in September 2026. It also signed a reported US$20 billion licensing deal for Groq's inference technology in December 2025.
The talks remain at an advanced negotiation stage. Neither company has publicly confirmed an agreement. The price and terms could still change.
Read more: theinformation.com