Nvidia in talks to acquire Reflection AI or increase investment, FT reports

Nvidia, the leading chipmaker, is reportedly in talks to acquire artificial intelligence startup Reflection AI or bolster its investment, according to the Financial Times. Sources close to the matter suggest several potential outcomes, such as a full takeover, increased financial support, or an agreement to hire Reflection's employees and license its technology.

This type of arrangement, known as an acqui-hire, could enable Nvidia to obtain the startup's talent and technology without facing the lengthy regulatory hurdles typically associated with a traditional acquisition. While negotiations are still in the preliminary stages, a deal could be finalized within weeks, though it remains uncertain if these discussions will materialize.

Reuters has not independently verified the reported talks. Nvidia has already been a substantial investor in Reflection AI, committing $800 million to the company in 2024. Reflection AI was founded in 2024 by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou, and the startup focuses on developing AI tools aimed at automating software development.

Positioned in the burgeoning market for AI-powered coding assistants and autonomous software development agents, Reflection AI is aiming to compete with lower-cost alternatives such as DeepSeek and Kimi. In April, Reflection AI's CEO, Misha Laskin, disclosed to CNBC that the company was in talks to raise additional funding at a $25 billion pre-money valuation.

Launching its first open-weight AI model, Beam, in April, the company is targeting coding and agentic applications, aiming to provide a more affordable option for developers tackling complex programming tasks. If the rumored deal comes to fruition, it would expand Nvidia's engagement in AI software development, complementing its dominant position in the processor market for training and running artificial intelligence models.

The discussions also highlight the increasing interest in AI coding technology as chipmakers and software companies explore opportunities beyond infrastructure spending.