Nvidia's subsidiary Reflection AI plans to raise an additional $2.5 billion, with its valuation soaring to $25 billion! Focused on open-source AI to compete with China's DeepSeek

Led by NVIDIA’s deep support, the open-source AI startup Reflection AI is currently negotiating a new funding round of $2.5 billion, with a pre-money valuation reaching as high as $25 billion. The company is dedicated to building an open-source AI ecosystem led by the United States and is seen as a key strategic move to counter China’s DeepSeek rise.
(Background: US DOJ indicts co-founder and Taiwan GM of Supermicro for smuggling billions of dollars worth of NVIDIA chips into China; SMCI stock plunges 12%)
(Additional context: Jensen Huang confirms Nvidia’s “cash flow is too large”: by 2026, 50% will be returned to shareholders via buybacks and dividends)

Table of Contents

Toggle

• Valuation surges 15-fold, JPMorgan and NVIDIA join in
• Defending US AI sovereignty, targeting China’s DeepSeek
• Strategic extension of NVIDIA’s hardware ecosystem

As the global AI arms race intensifies, a new blockbuster funding news has emerged in the open-source AI field. According to The Wall Street Journal on March 26, the highly anticipated AI unicorn Reflection AI is in talks for a new funding round, aiming to raise approximately $2.5 billion.

Notably, this startup, less than two years old, has raised its pre-money valuation from $20 billion reported in early March to $25 billion.

Valuation Surges 15-Fold, JPMorgan and NVIDIA Join In

Reflection AI’s valuation growth trajectory is astonishing. In March 2025, its valuation was only about $545 million; by October of the same year, with NVIDIA leading the investment and participation from Sequoia Capital, Lightspeed Venture Partners, and former Google CEO Eric Schmidt, it successfully raised $2 billion, skyrocketing its valuation to $8 billion—a 15-fold increase.

Now, in less than half a year, the valuation target has more than tripled again. It is reported that the potential investors remain prestigious, including existing investors Disruptive and NVIDIA, as well as traditional financial giant JPMorgan Chase, which is interested in participating through its “Resilience and Security Program.” If this round of funding proceeds smoothly (post-money valuation around $27.5 billion), it will be one of the largest open-source AI funding deals in recent years.

Defending US AI Sovereignty, Targeting China’s DeepSeek

Founded in 2024 and headquartered in New York, Reflection AI was co-founded by former Google DeepMind researcher Misha Laskin and Ioannis Antonoglou. Initially focusing on autonomous coding agents, the company later pivoted to building frontier open-source AI models.

This early team of only about 60 elite members has attracted intense capital market interest, underpinned by strategic geopolitical significance. Multiple foreign media analyses point out that, facing the rising dominance of China’s open-source AI models like DeepSeek, the US urgently needs a domestic force capable of leading global open-source AI standards. Reflection AI is seen not only as an open alternative to closed-source giants like OpenAI and Anthropic but also as a crucial line of defense to prevent other countries from dominating global AI standards.

NVIDIA’s Hardware Ecosystem Strategic Extension

As the key backer, NVIDIA provides not only massive chips and computing power but also demonstrates its strategic intent through multiple heavy investments. Market analysts generally believe that Reflection AI is an important chess piece for NVIDIA to leverage its hardware dominance to build and expand the global open-source AI network.

Although Reflection AI has not yet officially released its main “Mixture-of-Experts” model, investors are confident in its development direction. The company plans to adopt an “Open Weights” strategy, meaning it will open model weights to the public while retaining control over training datasets and underlying training pipelines. This approach of balancing an open ecosystem with a strong business moat has become one of the most favored investment targets in the current AI capital market.