Nvidia is in advanced negotiations to acquire the AI startup Inception for more than $500 million. The deal would bring the startup's three co-founders and its team of over 30 employees into the company. The agreement may also include performance-based earnout awards for staff.

Inception, based in Palo Alto, specializes in diffusion-based large language models. Its Mercury model family utilizes an iterative refinement process to generate text. The company claims these models perform 5 to 10 times faster than existing solutions from Google, OpenAI, and Anthropic while maintaining similar accuracy levels.

The startup previously raised a $50 million seed round in November 2025, with participation from investors including Menlo Ventures, Microsoft M12, and Nvidia’s own venture arm, NVentures. Individual angel investors included Andrew Ng and Andrej Karpathy.

This potential acquisition aligns with Nvidia's strategy to expand its influence beyond hardware into AI models and software. The company aims to integrate Inception’s technology into its Nemotron portfolio, which is currently utilized by organizations such as Palantir Technologies and Booz Allen Hamilton.

Nvidia has recently pursued several major deals to strengthen its AI ecosystem. These include the planned $12.9 billion acquisition of Hugging Face and a $20 billion licensing agreement with Groq, both initiated in 2026. The Inception deal remains subject to change as negotiations continue.