Set Trending Topics as a preferred source on Google.
After Zhipu and MiniMax, China’s two best-known A.I. labs now want to go public as well: DeepSeek is preparing a listing in Shanghai, while Moonshot AI, the company behind the Kimi models, is aiming for Hong Kong. Both have grown to valuations above $50 billion within a few months. A look at the benchmarks, however, shows that their models are currently not at the very top, not even among their Chinese rivals.
Moonshot AI: Hong Kong in the First Quarter of 2027
Moonshot AI was valued at nearly $50 billion (about 42.7 billion euros) in its latest funding round, Bloomberg reported. Over the summer, its valuation stood at $31.5 billion; in May, in a round led by Meituan, it was around $20 billion, and at the end of 2025 only about $4.3 billion. Its backers include Alibaba and Tencent. A group of mostly French investors also bought in ahead of the listing through an investment vehicle.
The initial public offering is planned for Hong Kong in the first quarter of 2027, and early meetings with investors could begin as soon as this month. According to Bloomberg, Moonshot is considering raising up to $5 billion (about 4.3 billion euros). Bank of America, CICC, Deutsche Bank and Goldman Sachs are working on the deal. The business is growing rapidly as well: Annualized revenue is around $1 billion and is expected to rise to $2 billion by December. The driver is the open-weight model Kimi K3, which triggered a “new DeepSeek moment” on stock markets in July.
DeepSeek: A Second Multibillion-Dollar Round and a Shanghai Listing
Until a few months ago, DeepSeek had no outside investors at all. In early summer, the Hangzhou-based lab closed its first funding round, raising about $7.4 billion at a valuation of more than $50 billion. The founder, Liang Wenfeng, contributed the largest share himself, joined by Tencent, the battery maker CATL and China’s state-backed A.I. fund.
According to The Information, a second round is now close to being finalized: another 50 billion yuan (about $7.5 billion, or 6.4 billion euros) at a valuation of 500 billion yuan, roughly $75 billion (64 billion euros). Annualized revenue has doubled to about $1 billion after DeepSeek raised its prices in August. For its listing on Shanghai’s tech-focused STAR Market, DeepSeek has brought in Citic Securities and three other banks as underwriters, the South China Morning Post reported. The I.P.O. process is set to begin this year, with the listing itself expected in 2027.
Where Their Models Rank on Artificial Analysis
On the Artificial Analysis Intelligence Index, which combines results from numerous benchmarks, U.S. labs currently dominate. Here is how the best models from Moonshot and DeepSeek compare:
| Model | Developer | Intelligence Index | Cost per Task |
|---|---|---|---|
| Claude Opus 5.5 (max) | Anthropic | 58 | $5.98 |
| Gemini 4 Argon (high) | 53 | $1.99 | |
| GPT-6 Astra (max) | OpenAI | 53 | $3.26 |
| MiMo-V2.6-Pro | Xiaomi | 46 | $0.13 |
| GLM-5.3 (max) | Z.ai | 45 | $2.01 |
| Qwen3.8 Max | Alibaba | 45 | $5.41 |
| Kimi K3 (max) | Moonshot AI | 44 | $2.00 |
| DeepSeek V4.1 Flash (max) | DeepSeek | 39 | $0.27 |
| DeepSeek V4 Pro 0813 (max) | DeepSeek | 36 | $0.67 |
Kimi K3 scores 44 points, 14 points behind the leader. At its launch in July, the model was playing in the same league as Gemini 3.1 Pro; since then, not only U.S. labs but also Chinese rivals have moved ahead: Xiaomi’s MiMo-V2.6-Pro tops the ranking of open-weight models with 46 points, just ahead of Z.ai’s GLM-5.3 and Kimi K3.
DeepSeek trails even further. Its best model, V4.1 Flash, scores 39 points, and its flagship V4 Pro just 36. DeepSeek’s strength is price: V4.1 Flash costs only about 27 cents per task, a fraction of what Claude, GPT or Kimi K3 charge.
Strong Competition on the Stock Market
On the stock market, DeepSeek and Moonshot will meet Chinese rivals that are already listed. Zhipu, which operates under the name Z.ai, and MiniMax have gone public and were most recently valued at $46.6 billion and $13 billion. Then there are tech giants like Alibaba and Xiaomi, whose models now rank ahead of Kimi and DeepSeek on the benchmark. For investors, the top spot in the ranking may therefore matter less than the question of who can turn rapid revenue growth from cheap open-weight models into profits.
