The ¥80B round is closing, handing the company a top-tier stockpile of ammunition in China's AI race. The original target was just ¥50B, but investors nearly doubled it after the V4 Flash release. What matters to us is how the money gets spent, not the funding figure itself.
What This Is
According to people familiar cited by Bloomberg, DeepSeek is nearing closure of at least ¥80B (~$12B) in funding, led by Tencent and CATL. The capital will flow into three tracks: scaling compute infrastructure, accelerating model R&D, and expanding the core team.
DeepSeek has hired CITIC Securities as IPO advisor, targeting the Shanghai Stock Exchange's STAR Market with a window in early 2027. This forces a shift for a founding team built on an early "no fundraising, no commercialization, no roadshow" minimalist stance into a standardized public company.
Industry View
The mainstream read is bullish: V4 Flash validated DeepSeek's inference-efficiency roadmap, and its open-source strategy (R1, V4 series) built a differentiated edge that capital markets are voting for with real money. Tencent, Alibaba, and ByteDance have all raised capex this year (Tencent's Q1 hit ¥31.9B, Alibaba's future investment exceeds ¥380B, ByteDance revised up 25% to ¥200B). DeepSeek's round is less an end result than one move in this compute-cap allocation chain.
But the bear case holds. Raising ¥80B doesn't mean spending it well — in a cash-burn arms race, capital efficiency is the real moat. The rumored ¥400B valuation is unconfirmed, and the IPO window hinges on regulatory review and market sentiment. The more grounded risk: a quant-background closed lab crossing over to a STAR Market public company is itself a governance stress test.
Another read on why both Tencent and CATL are at the table simultaneously: two different bets. Tencent is patching its AI gap and binding the strongest tech asset to relieve Hunyuan's multi-front pressure from Doubao, Qwen, and DeepSeek. CATL is playing "compute-power synergy" — million-card cluster data centers pull hundreds of millions of kWh annually, putting energy storage players in a natural position as compute suppliers.
Impact on Regular People
For enterprise IT: Top-tier cloud vendors and IDC (Internet Data Center) providers will face a new compute procurement wave, with related orders and partnerships picking up noticeably in H2 2026.
For careers: Salary premiums for core AI roles (algorithms, inference optimization, distributed training) continue to climb, and a quant background stacked with AI experience is becoming a high-weight line on the resume.
For consumer markets: CATL's move into compute-power synergy won't show near-term impact, but over time, shifts in data center electricity structure could pass through into industrial power costs and indirectly nudge consumer pricing.