- DeepSeek founder Liang Wenfeng declared an explicit AGI goal as a $10 billion funding round for the Chinese AI lab advances, Bloomberg reported on May 22, 2026.
- The disclosure marks the most direct AGI commitment from a major Chinese AI lab.
- DeepSeek’s previous V4 Pro model was recently evaluated by the US Commerce Department’s CAISI; the $10B round would put it in the same capital tier as Anthropic and OpenAI.
- The fundraise advances against the backdrop of US export controls limiting Chinese access to Nvidia’s most capable chips.
What Happened
DeepSeek founder Liang Wenfeng declared an explicit artificial general intelligence (AGI) goal as a $10 billion funding round for the Chinese AI lab advances, Bloomberg reported on Thursday. The framing places DeepSeek alongside OpenAI, Anthropic, and Google DeepMind in publicly committing to AGI as an end-state objective rather than narrower commercial AI capability.
Why It Matters
DeepSeek has been the most prominent Chinese AI lab since the V3 model release in late 2024 and the V4 series in 2025. The lab’s V4 Pro model was recently evaluated by the U.S. Commerce Department’s Center for AI Standards and Innovation (CAISI), reflecting its position in the global AI capability tier. A $10 billion funding round would put DeepSeek in the same capital band as Anthropic’s recent rounds, OpenAI’s $122 billion Series G, and xAI’s $230 billion valuation.
The AGI framing is also politically significant. Chinese AI strategy through 2024-2026 has been articulated by the State Council primarily in terms of commercial deployment and economic growth. An explicit AGI goal from DeepSeek’s founder shifts the framing toward the same long-horizon capability-development trajectory that Western frontier labs publicly pursue.
Technical Details
Bloomberg’s reporting did not disclose the specific lead investors, valuation, or close timing for the $10 billion round. Earlier reporting has suggested participation from major Chinese tech investors and possibly sovereign wealth participants. DeepSeek’s existing investor base includes Alibaba and other major Chinese tech holding companies.
The fundraise advances against the backdrop of US export controls. The Biden administration’s chip-export restrictions, continued under the Trump administration, limit Chinese access to Nvidia’s most capable H100, H200, and B200 chips. DeepSeek’s earlier V3 release demonstrated the lab’s ability to train competitive models on constrained Nvidia hardware (PTX-level optimisation rather than higher-level CUDA work). The Alibaba Zhenwu M890 chip announcement on May 20 and Huawei’s Ascend roadmap provide alternative compute substrates for Chinese frontier-AI development.
Who’s Affected
DeepSeek gains the capital base and AGI-aligned narrative to compete more directly with Western frontier labs. Chinese AI competitors — Alibaba (Qwen series), Baidu (Ernie), ByteDance (Doubao), Moonshot AI (Kimi), Zhipu AI (GLM) — face a higher-capitalised lab with explicit long-horizon framing. Western frontier labs — OpenAI, Anthropic, Google DeepMind — face renewed competitive pressure on AGI-tier capability. The U.S. AISI and policy community will read the AGI declaration as evidence supporting export-control rationale. Chinese policymakers gain a national-champion narrative for the next leg of capital deployment.
What’s Next
Specific deal-close details have not been disclosed. Expect formal funding announcements with named investors, valuation, and use-of-proceeds in coming weeks. DeepSeek’s next major model release (V5 or successor) will be the technical milestone that translates the new capital into capability disclosure. Industry watchers should track parallel funding announcements from Moonshot, Zhipu, and other Chinese frontier labs.
