$70B pre-money · voting-rights catch · 148x P/S · STAR Market IPO path
DeepSeek, the Chinese AI lab behind the open-weight R1 and V4 models, has restarted talks for a second funding round targeting roughly 50 billion yuan (~$7 billion) at a pre-money valuation of about 500 billion yuan (~$70 billion) — up 43% from its first round two months ago. If it closes, DeepSeek will have raised over $14 billion in under five months. This piece covers the timeline, round-by-round numbers, compute bill and voting-rights structure, peer comparison, STAR Market backdrop, and a six-step runbook. Figures for Round 2 remain media-sourced and unconfirmed by DeepSeek.
DeepSeek spent five years refusing outside capital. Then the calendar compressed: ownership restructuring, a record first round, an IPO-prep leak, a pause after a closed-door transcript circulated online, and a quiet restart in early August.
April 2026 — Cap table reset and Round 1 opens: A corporate filing shows registered capital increased and founder Liang Wenfeng raised his direct stake from 1% to 34%. Combined with an entity he controls, total control reached roughly 84.29%. DeepSeek opened its first external round and previewed V4.
June 2026 — Round 1 closes: Roughly 50 billion yuan (~$7.4 billion) raised; post-money valuation above 350 billion yuan (reported $52–59 billion across sources) — the largest first-round raise in Chinese AI history. Liang personally contributed 20 billion yuan; Tencent put in 10 billion yuan, CATL 5 billion yuan.
July 14–17 — IPO prep and Round 2 talks: Outlets reported STAR Market IPO preparation and second-round talks at roughly $71 billion pre-money (~480 billion yuan). ARR of $400–500 million (mostly API tokens) surfaced publicly for the first time.
July 25–26 — Talks pause: Bloomberg and others said Liang was unhappy that remarks from closed-door investor meetings had circulated online; some standby investors were told to hold off on signing.
August 4–5 — Round 2 restarts: Dealmakers cited by Caijing said the round restarted at ~50 billion yuan and ~500 billion yuan (~$70 billion) pre-money, with signing expected in late August. Both sides want the process kept low-profile.
Caveat: Every figure above about the second round — amount, valuation, timeline — comes from anonymous dealmakers cited by Chinese financial media, not from an official DeepSeek statement. Terms could still shift before signing.
Put the closed round next to the in-talks round and the cadence becomes obvious: same headline raise size, sharply higher valuation, and a five-month combined total that would top $14 billion if Round 2 lands.
| Item | Round 1 (closed) | Round 2 (in talks) |
|---|---|---|
| Talks opened | April 2026 | Restarted mid-July, paused, restarted again Aug 4–5 |
| Expected/actual close | June 2026 | Late August 2026 (planned) |
| Amount raised | ~50B yuan (~$7.4B) | Target ~50B yuan (~$7B) |
| Valuation basis | Post-money >350B yuan | Pre-money ~500B yuan (~$70B) |
| Valuation increase | — | ~+43% vs. Round 1 |
| Key backers | National AI Industry Investment Fund, Tencent (10B yuan), CATL (5B yuan), JD.com, NetEase, IDG Capital, Loyal Valley Capital, Shixiang Capital | Round-1 runner-up investors + some existing backers increasing stakes |
| Combined total if Round 2 closes | — | Over 100B yuan (~$14B) in under 5 months |
| Metric | Value | Note |
|---|---|---|
| Annualized revenue (ARR) | ~$400–500 million | Mostly API token usage; media-sourced, not an official disclosure |
| Gross margin | Reportedly >50% | Unverified by independent audit |
| Implied price-to-sales (P/S) | ~140–150x | Vs. OpenAI ~65x and Anthropic ~21x, per dealmaker estimates |
| Monthly active users | 100M+ (externally reported) | Methodology undisclosed |
Pricing a foundation-model company is fundamentally an options bet, not a cash-flow valuation.
Shortly after closing Round 1, DeepSeek said it would double headcount across data-center and AI-agent teams, and Reuters reported it was hiring chip-design engineers for in-house AI inference chips. Industry analysts estimate that for every 10 billion yuan raised, roughly 7 billion yuan goes into compute-related spending — chips, data centers, bandwidth, liquid cooling. The fundraising cadence is a race to keep pace with compute buildout.
In Round 1, most outside capital flowed through a limited partnership controlled by Liang Wenfeng: no voting rights and a five-year lock-up. The exception is China's National AI Industry Investment Fund, which invested directly with voting rights and no lock-up. That keeps Liang's control near 84% — and it is exactly the detail that drew scrutiny from outlets like Forbes on governance and state influence.
At a $70 billion pre-money valuation against $400–500 million in ARR, implied P/S sits around 140–150x, dwarfing OpenAI's ~65x and Anthropic's ~21x. Investors are not pricing today's revenue — they are pricing the chance that DeepSeek becomes infrastructure-level in China's compute ecosystem and enterprise agent market.
For builders: Faster fundraising usually tracks hotter API demand and quota risk. If you run Agents on DeepSeek, plan multi-model routing and isolated remote Mac nodes. See our V4 Flash benchmark review and custom inference chip analysis.
Still-private frontrunners carry a steeper premium than peers that already face public-market scrutiny.
| Company | Listing status | Latest valuation/market cap | Reported ARR | Recent funding pace |
|---|---|---|---|---|
| DeepSeek | Private, preparing STAR Market IPO | ~500B yuan pre-money (~$70B, in talks) | ~$400–500M | 2 rounds in 4 months, targeting >$14B combined |
| Moonshot AI (Kimi) | Private | ~$20B (May 2026); reportedly seeking $30B later | ~$200M | 4 rounds in 6 months, ~$3.9B total |
| Zhipu AI (Z.ai) | Listed (Hong Kong) | ~350B yuan market cap (May 2026) | Undisclosed | ~8.3B yuan raised pre-IPO |
| MiniMax | Listed (Hong Kong) | ~210B yuan market cap (May 2026) | Undisclosed | ~11B yuan raised pre-IPO |
DeepSeek and Moonshot both sit near 140–150x P/S — well above listed Zhipu and MiniMax. Private-market investors are, for now, paying more for the two labs that have not yet faced secondary-market pricing.
A leaked closed-door transcript stalled the deal: The July pause was reportedly driven by Liang's frustration that investor-meeting remarks spread online — a reminder that a larger investor base makes a low profile harder to keep.
Voting-rights structure draws outside scrutiny: Most external investors have no vote and a five-year lock-up; only the state-backed National AI Industry Investment Fund gets direct voting rights with no lock-up. DeepSeek has not publicly settled the governance questions this raises.
The valuation-to-revenue gap remains unresolved: A 140–150x P/S is extreme even for high-growth SaaS (typically 30–50x). Whether the mark holds depends on converting technical lead into scaled enterprise revenue after a STAR Market listing — still untested.
Again: Deal size, valuation, and ownership details above come from anonymous-sourced reporting (Caijing, Reuters, Bloomberg, Forbes, among others). Treat specific Round 2 numbers as reported-but-unconfirmed until a formal announcement.
Funding headlines move faster than quotas and inference costs. Use this checklist to keep production Agents resilient while Round 2 remains in talks.
Separate "in talks" from closed facts: The $70B pre-money and ~$7B target are negotiation marks — label any external briefings as media-sourced until DeepSeek confirms.
Watch the late-August signing window: If close slips or terms step down, revisit token budgets and fallback models immediately.
Model the compute bill, not just API list prices: Analysts put ~70% of raised capital into compute. Re-run your three-year TCO for API vs local/cloud Mac inference.
Configure multi-model fallback: Pair DeepSeek with open-weight alternatives (Kimi K3, Qwen family) so a single vendor's funding or policy narrative cannot stall you.
Track STAR Market and governance signals: Year-end IPO filing progress and any clarification on voting rights/lock-ups matter for long-term partnership risk.
Isolate Agents and CI on remote Mac nodes: Run inference experiments, Xcode builds, and 24/7 Agents off your laptop and shared cloud accounts.
Teams chasing DeepSeek's funding and compute narrative often default to public-cloud GPUs or purchased Mac fleets. Pure cloud GPUs leave little buffer against quota throttling, regional latency, and pricing swings; owned fleets bring long procurement cycles, fast depreciation, and 24/7 ops overhead. For production environments that need stable iOS CI/CD and AI Agent automation, VpsMesh Mac Mini cloud rental is usually the better fit — scale remote Mac nodes on demand and run Agent pipelines plus local inference experiments in isolation. See Mac Mini M4 rental pricing and the cloud order page.
Not yet. As of this writing, the round is still in negotiation, targeting a close by late August 2026. The final amount and terms could differ from what's currently being reported.
The company is funding a rapid buildout of data centers, in-house AI chips, and headcount across its agent and infrastructure teams — capital expenditure that is outpacing what its first raise covered, according to multiple reports.
No. It comes from dealmakers cited anonymously by Chinese financial media (primarily Caijing), not from an official DeepSeek statement, and it could change before any agreement is signed.
Not necessarily — and that is part of the controversy. In Round 1, most outside investors received no voting rights and a five-year lock-up, while only China's National AI Industry Investment Fund got direct voting rights. Those governance questions remain unresolved.
DeepSeek is reportedly preparing to file for a STAR Market listing in Shanghai by the end of 2026, targeting a 2027 debut. The STAR Market is a mainland China exchange, so retail access for international investors would likely be indirect (e.g., through connect programs) rather than direct participation in this private round.
Configure multi-model routing and run Agents plus local inference experiments on isolated remote Mac nodes. See the help center for VpsMesh Mac Mini cloud deployment, check rental pricing, or use the order page to provision nodes on demand.