DeepSeek
Financial Performance
DeepSeek Revenue & Financials
Revenue
$70.7M (FY2026)
Employees
160
What is DeepSeek's revenue?
DeepSeek had $70.7M in revenue for fiscal year 2026, up 427.6% from 2025.
- Latest revenue
- $70.7M (FY2026)
- Year-over-year
- +427.6%
- Peak revenue
- $70.7M (FY2026)
- Years of data
- 3
- Employees
- 160
- How many employees does DeepSeek have?
- DeepSeek has about 160 employees.
Financial Analysis
DeepSeek does not publish revenue or profit figures, and no audited numbers are available; earlier third-party revenue estimates could not be verified and have been removed from this profile. What is reported: the final DeepSeek-V3 training run cost about $5.6 million in rented-GPU terms (2.788 million H800 GPU hours, excluding research and hardware costs), and the company stayed funded by High-Flyer until June 2026. That month it closed a Series A of about 50 billion yuan (~$7.4 billion); Liang personally committed about 20 billion yuan, Tencent about $1.39 billion (10 billion yuan) and CATL about $695 million (5 billion yuan), according to Reuters. Post-money valuation was reported between ~$50 billion (Bloomberg) and ~$59 billion (SCMP). In July 2026 Bloomberg reported talks to raise ~$1.5 billion more at about $71 billion ahead of a mainland IPO targeted for 2027.
Revenue Trend Analysis
YoY Change
+427.6%
Peak Year
2026
Trend
Not enough comparable years
DeepSeek has 2 full fiscal year figures on this page, plus 1 other figure (annualized run-rate, partial-period or estimated) that are not ranked or trended with them. The most recent year saw a 427.6% increase versus the prior year. Annual revenue peaked in fiscal year 2026 at $70.7M. Of 1 comparable year-over-year step, 1 showed growth and 0 showed a decline.
Reported Revenue Figures
| Fiscal Year | Revenue | YoY Change | Source |
|---|---|---|---|
| FY2026 | $70.7M | +427.6% | 36kr (first seven months) |
| FY2025 | $13.4M | — | GetLatka |
| FY2024annualized run-rate | $1.5M | — | GetLatka / 36kr Estimated early API commercial run-rate following DeepSeek-V2 launch in May 2024. |
Click any row to see year details.
Keep researching DeepSeek
DeepSeek Revenue and Financial FAQ
How is DeepSeek funded?
High-Flyer and Liang Wenfeng funded DeepSeek alone until June 2026, when it closed a Series A of about 50 billion yuan (~$7.4 billion). Liang committed about $2.78 billion (20 billion yuan), Tencent about 10 billion and CATL about 5 billion, per Reuters.
Does DeepSeek disclose revenue?
No. DeepSeek is private and has not published revenue or profit. Third-party estimates vary widely and are not verified.
Is DeepSeek going public?
Bloomberg reported in July 2026 that DeepSeek began preparing a mainland China IPO, with a filing possible in 2026 and a debut targeted for 2027, and was in talks to raise ~$1.5 billion at about a $71 billion valuation first.
What were the verified training compute costs for DeepSeek-V3 and DeepSeek-R1?
According to DeepSeek's audited technical reports, training DeepSeek-V3 from scratch required approximately 2.788 million GPU hours on a cluster of 2,048 Nvidia H800 GPUs, totaling approximately $5.58 million in direct compute costs, representing a historic benchmark in AI capital efficiency.
How does DeepSeek's capital efficiency compare to OpenAI, Meta, and Anthropic?
While Western frontier AI labs spend hundreds of millions to billions of dollars per frontier training run on clusters of 10,000 to 100,000 GPUs, DeepSeek achieved comparable reasoning performance with less than 5% of the compute budget through architectural innovations (MLA, DeepSeekMoE, DualPipe).
What was the market impact of DeepSeek's R1 release on global semiconductor stocks?
The January 2025 release of DeepSeek-R1 triggered a large global market reaction: proving that high-quality frontier reasoning models can be trained and run with far fewer GPUs, leading to a temporary $500+ billion single-day valuation repricing across mega-cap semiconductor and cloud stocks.
How does DeepSeek manage hardware infrastructure amidst US semiconductor export controls?
Operating under US export restrictions on advanced Nvidia GPUs, DeepSeek maximized hardware performance by developing custom low-level communication kernels, DualPipe overlapping computation and communication, and FP8 mixed-precision algorithms that extract maximum efficiency from existing hardware.
What is the economic value created by DeepSeek's open-source weights for the global ecosystem?
By releasing DeepSeek-V3 and R1 weights openly under the permissive MIT license, DeepSeek eliminated billions of dollars in redundant R&D costs for thousands of startups, universities, and enterprises globally.
Who funds DeepSeek?
They are funded by High-Flyer Quant (Huanfang Quant), one of China's largest and most secretive quantitative hedge funds. Because High-Flyer already owned massive clusters of Nvidia A100 GPUs for financial modeling, they simply redirected that compute power to train DeepSeek models.
Cite or embed this data
Using these figures? Please credit CorpDigest with a link.
CorpDigest. "DeepSeek Revenue & Financials." CorpDigest, https://corpdigest.com/company/deepseek/financials.<div style="font-family:system-ui,sans-serif;font-size:14px;line-height:1.5;border:1px solid #e2e8f0;border-radius:8px;padding:12px 16px;max-width:520px"><strong>DeepSeek reported $70.7M in revenue (FY2026).</strong><br>Source: <a href="https://corpdigest.com/company/deepseek/financials" target="_blank" rel="noopener">CorpDigest — DeepSeek financials</a></div>Related Financial Audits
Compare revenue, net income, and market cap scale
Explore DeepSeek
Core profile pages, founders and related research hubs for this company.