IT之家 · 智能时代 · 10/11/2026, 06:46:30
Nvidia Negotiates Acquisition or Investment in Reflection AI to Secure Open-Weight Model Beam via Talent Acqui-hire
Nvidia is in advanced negotiations to acquire or further invest in open-source startup Reflection AI, aiming to secure its newly released open-weight model Beam and core technical team. The deal may utilize a 'talent acqui-hire' structure to bypass antitrust scrutiny, with terms still being finalized. This move signifies Nvidia's strategic deepening of efforts to build a domestic open-weight ecosystem through capital means, compensating for its lack of proprietary frontier LLMs.
SOURCE COVERAGEFull Original Report
On October 11, the Financial Times reported that, according to sources familiar with the matter, Nvidia is in advanced negotiations regarding a potential acquisition of or additional investment in US startup Reflection AI.
Reflection AI specializes in developing open-weight models and released its first open-weight AI model, Beam, on October 5 (local time).
The sources added that the talks are still in their early stages and the potential deal could take various forms—including what is known as an "acqui-hire." Under this structure, Nvidia would absorb Reflection’s core team and obtain technology licensing rights, thereby bypassing the lengthy antitrust reviews typically associated with full-scale mergers and acquisitions.
Nvidia is already a major shareholder in Reflection and may choose to deepen ties through a new round of equity investment or by providing additional chips and compute resources. Previously, Nvidia invested $800 million in Reflection (approximately RMB 5.365 billion at current exchange rates, per IT Home).
According to multiple sources, the two parties are expected to finalize terms within the coming weeks, though they emphasized that the negotiations could still fall apart. Specific deal terms remain unclear, but during its previous funding round in March of this year, Reflection was valued at $25 billion (approximately RMB 167.662 billion at current exchange rates). Both Nvidia and Reflection declined requests for comment.
Amid strict antitrust scrutiny by regulators over M&A activities in the AI sector, "acqui-hires" have increasingly become the preferred strategy for large tech companies. In December last year, Nvidia employed a similar structure in a $20 billion deal involving chip design firm Groq (approximately RMB 134.129 billion at current exchange rates); this transaction also recently sparked legal action from former Groq employees who were not absorbed into the deal.
Over the past two years, Nvidia has embarked on an aggressive expansion of investments aimed at boosting the overall AI industry ecosystem and consolidating its dominant position at the core of the compute economy. Although Nvidia has launched its own family of open-weight AI models called Nemotron, it still lacks a top-tier frontier large language model developed entirely in-house.
In July this year, Nvidia CEO Jensen Huang led calls for establishing a domestic US open-weight "ecosystem," emphasizing its importance for "sparking innovation and prosperity across the nation." In September, the tech giant—with a market capitalization of 13 billion (approximately RMB 87.184 billion at current exchange rates).
Nvidia has cumulatively channeled tens of billions of dollars into leading LLM labs like OpenAI, emerging GPU cloud providers such as Nebius, and numerous smaller teams. These investments, in turn, further drive demand for Nvidia’s AI chips—which remain the critical lifeline for any enterprise building or running large models.