Less than six months after closing a $2 billion funding round that valued it at $8 billion, Nvidia-backed Reflection AI is already circling back to investors with ambitions that would nearly triple its worth, according to a report Wednesday.
The Wall Street Journal, citing people familiar with the matter, said the fast-rising artificial intelligence startup is in discussions to raise about $2.5 billion at a $25 billion pre-money valuation — the price tag before the new capital comes in. If completed, the deal would rank among the largest single funding rounds in the open-source AI sector and underscore the ferocious investor appetite still gripping frontier AI companies.
Founded in 2024 by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou, Reflection has positioned itself as a Western counterweight to Chinese AI models, particularly those from DeepSeek. The company focuses on developing powerful open-source large language models and tools that automate software development — helping engineers write, test, debug, and maintain code at unprecedented speed.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
It has also pursued partnerships for “sovereign AI” infrastructure, including a high-profile plan to build a 250-megawatt AI data center factory in South Korea with local conglomerate Shinsegae Group, leveraging Nvidia’s GPU technology.
Nvidia, the dominant force in AI hardware, has been a major backer. It led or participated heavily in Reflection’s previous $2 billion round in October 2025 and had earlier invested roughly $800 million when the company was valued at around $8 billion post-money. Other notable investors include Sequoia Capital, Lightspeed Venture Partners, DST Global, Disruptive, Eric Schmidt (former Google CEO), and Zoom founder Eric Yuan.
The latest talks include potential participation from JPMorgan Chase, which is considering investing through its newly launched Security and Resiliency Initiative. That $1.5 trillion, 10-year program aims to bolster critical U.S. industries and frontier technologies, including AI, cybersecurity, and supply-chain resilience, with up to $10 billion earmarked for direct equity investments in strategic companies.
The rapid valuation jump, from roughly $545 million in early 2025 to $8 billion last fall and now a potential $25 billion pre-money, reflects both the blistering pace of progress in AI coding assistants and the broader market’s willingness to bet big on open models that could challenge closed systems from OpenAI, Anthropic, and Google.
Reflection’s strategy blends aggressive scaling of compute resources with a commitment to open-source development, appealing to governments and enterprises wary of relying solely on proprietary U.S. or Chinese models. Its push into sovereign AI clouds and localized systems has drawn interest from allies seeking greater control over their AI infrastructure amid rising geopolitical tensions.
Still, the AI funding environment remains selective. While top-tier infrastructure and model companies continue to command sky-high valuations, many observers warn of growing scrutiny over unit economics, energy demands, and the risk of commoditization as open-source alternatives proliferate.
However, Reflection appears to be riding the wave for now. A successful $2.5 billion raise at the reported terms would give the company substantial dry powder to expand its model training efforts, deepen hardware partnerships with Nvidia, and accelerate international infrastructure plays. The talks come as the AI boom shows little sign of cooling, with Nvidia itself maintaining its position as the world’s most valuable company on the strength of insatiable demand for its chips.