After DeepSeek: Moon’s Dark Side and Jumps Star Raise New AI Funding

From the start of 2026, the financing rhythm of China's large‑model market accelerated sharply. DeepSeek was reported to be negotiating a first external round led by the China National Integrated Circuit Industry Investment Fund (the "National Big Fund"), with a tentative valuation of about $45 billion (≈¥3,065 billion). Although the exact terms remain under discussion, the involvement of the state fund adds significant credit, enhancing DeepSeek's access to compute resources, policy support, and industry partnerships.

“Not short of money” rebels open doors

According to the Financial Times, four sources confirmed the National Big Fund's lead‑investor role. Tencent and Alibaba are also said to be in talks, indicating a joint push by state capital and internet giants. Founder Liang Wenfeng, who originally vowed never to accept external financing, now embraces the capital injection, reflecting DeepSeek's transition from an "open‑source outlier" to a "national strategic core".

Investors note that the primary purpose of this round is to establish a fair market value for employee stock options and to stabilize the technical team amid fierce talent competition and soaring compute costs.

20 billion refinancing, valuation over $200 billion

On May 7, HuaFeng Capital announced that Kimi (Moon’s Dark Side) completed a new financing round of roughly $20 billion, lifting its post‑money valuation beyond $200 billion (≈¥1,400 billion). The round was led by Meituan Longzhu, with participation from ShuiMu Capital, China Mobile, and CPE Yuanfeng. Meituan contributed over $2 billion, underscoring its confidence in Kimi's AI‑driven local‑service and instant‑retail scenarios.

China Mobile's involvement brings strong compute, network, and enterprise‑customer advantages, while CPE Yuanfeng (CITIC Industrial Fund) signals mainstream capital approval of Kimi's business model. Within six months, Kimi has completed four financing rounds, accumulating more than $3.9 billion in total funding. Its ARR surpassed $200 million in April, driven by subscription and API revenue, and the K2.6 model released in April 2024 boosted code‑generation and autonomous agent capabilities.

Despite these achievements, Kimi faces challenges: narrowing technical moats, pressure on consumer‑user growth, reliance on API/subscription revenue, and high compute and talent costs.

25 billion financing record, sprinting for Hong Kong IPO

On May 8, media reported that Jumps Star (Jie Yue Xing Chen) is set to close a financing round of nearly $25 billion, surpassing Kimi's latest round and positioning it as the largest recent funding in China's large‑model arena. The round attracted a cluster of industry‑chain investors—Huawei’s ODM partner HuaQin, phone‑ODM leader Longqi, sensor‑chip leader OmniVision, and telecom giant ZTE—reflecting Jumps Star's focus on embedding models into terminal devices.

Hong Kong Investment Management Co. (HKIC), dubbed the "Hong Kong Temasek," also joined the shareholder list. The company has recently converted from a limited liability company to a joint‑stock company, a prerequisite step for a Hong Kong IPO. If successful, Jumps Star would become the third Chinese large‑model firm listed in Hong Kong after Zhipu and MiniMax.

Founded in 2023 by former Microsoft Global VP Dr. Jiang Daxin, Jumps Star has released the Step series of general‑purpose models, including the open‑source Step 3.5 Flash and its optimized version Step 3.5 Flash 2603, which approach closed‑source performance on agent and math tasks. However, its 2025 revenue is projected at ¥5 billion, rising to ¥12 billion in 2026, indicating an early commercial stage.

Overall, the financing surge—DeepSeek's state‑backed round, Kimi's massive $20 billion raise, and Jumps Star's $25 billion funding—marks a pivotal shift in China's AI industry from pure technology breakthroughs to a multi‑dimensional competition involving capital, ecosystem, and policy. While abundant capital can accelerate growth, over‑financing risks misallocation of resources, and only firms that convert funding into sustainable technical and commercial barriers will prevail.