Skip to main content
Back to News
DeepSeek and Moonshot AI compete in $10B+ Chinese large model funding race
Funding
2 min read
CN

DeepSeek and Moonshot AI compete in $10B+ Chinese large model funding race

The AMW Read

DeepSeek's first external round after years of self-funding overturns a core case-study claim; both rounds total >$10B, a structural capital event reshaping China's LLM landscape.
NoveltySignificance
Foundation Models · Player MapFoundation Models · Case StudiesCapital Cycles
Moonshot AI
Moonshot AI

Foundation Models / LLMs

View Company Profile

DeepSeek and Moonshot AI compete in $10B+ Chinese large model funding race

In a 48-hour window in early May 2026, China's large-model sector unleashed over $10 billion in funding signals. Moonshot AI (月之暗面) closed roughly $2 billion in a Series D round at a $20 billion post-money valuation. The following day, sources disclosed that DeepSeek is pursuing its first-ever external equity round, targeting up to RMB 50 billion (~$7.3 billion), which would value the company at approximately RMB 350 billion ($51.5 billion). Simultaneously, Step-2 (阶跃星辰) raised nearly $2.5 billion, pushing the aggregate capital wave past $10 billion.

Why it matters: This concentrated capital event embodies the fastest-ARR-ramp pattern as China's top labs shift from R&D-centric narratives to strategic-asset pricing. The two companies exemplify divergent routes inside the same capital-compression arc: DeepSeek is positioning itself as the open-source infrastructure layer, betting on developer ecosystem lock-in and domestic-chip cost advantages to justify a $51.5 billion valuation with opaque revenue. Moonshot is pursuing a vertically integrated, agentic-AI commercial path, using its trillion-parameter MoE architecture and operator-backed compute network to expand enterprise ARR, which reached ~$200 million annualized. The outcome will test whether China's frontier models can sustain independent ecosystems or will converge toward hyperscaler-distribution consolidation.

Grounded expert take: The funding wave resolves an open debate about whether Chinese labs could attract global-scale capital without US chip access. DeepSeek's validated domestic-chip inference stack — Huawei Ascension (昇腾) full-stack adaptation with 10ms latency — provides a structural cost moat that opens a new frame: compute sovereign pricing. Meanwhile, Moonshot's shift from consumer subscription to enterprise API revenue (overseas revenue now exceeding domestic) signals that China's agent layer may command SaaS-like multiples. Both companies now face the same reckoning: at these valuations, the substrate's capital-cycle discipline demands verifiable unit economics within 12-18 months or a severe multiple correction.

#DeepSeek #MoonshotAI #ChinaAI #FoundationModels #AgentAI #DomesticChips #FundingRound

#DeepSeek#Moonshot AI#Chinese large models#funding race#open-source infrastructure#enterprise AI#domestic chips

How This Connects

Based on Foundation Models · Player Map

  1. 1d agoOpenAI cuts GPT-5.6 Sol developer API pricing by more than 20% for three monthsOpenAI
  2. 2w agoHugging Face CEO warns China may win the open AI race due to its open-source strategyHugging Face
  3. 1mo agoMoonshot AI plans final funding round at up to $50 billion valuation before Hong Kong IPO. Chinese A...Moonshot AI
  4. 1mo agoMoonshot AI launches Kimi K3, a 2.8 trillion-parameter open-weight model, claiming performance near US frontier labsMoonshot AI
  5. 1mo agoDeepSeek reportedly raising $1.5B at $71B valuation, planning 2027 IPODeepSeek
  6. 3mo agoDeepSeek and Moonshot AI compete in $10B+ Chinese large model funding race · THIS ARTICLE

Related News

More news from Moonshot AI

Stay updated with the latest news and announcements from Moonshot AI.

View all Moonshot AI news

Discover AI Startups

Explore 5,000+ AI companies with VC-grade analysis, funding data, and investment insights.

Explore Dashboard