PRIV-DEEP

DeepSeek

Business overview

Chinese open-weights frontier lab whose R1 release in 2025 disrupted assumptions about training cost and US compute moat.

Most consequential proof point that frontier-adjacent capability is reproducible at a fraction of prior cost; resets the AI compute investment debate.

Products

DeepSeek-V3
DeepSeek-R1

Customers (1)

Open-source community

Competitors (3)

Composite score

Higher = stronger overall research case

Fundamental54

Technical50

Valuation60

AI relevance88

Risk-adjusted (higher = safer)33

Segment-relative fundamental percentile: 100

Financial summary

Curated, illustrative

Revenue

$50.00M

Growth (YoY)

1000.0%

Gross margin

40.0%

Operating margin

-150.0%

FCF

$-100.00M

Capex

$75.00M

Net cash / (debt)

$0.00

ROIC

-30.0%

Customer conc.

Medium

Valuation

Market cap

$10.00B

Enterprise value

$10.00B

P/E (TTM)

—

Forward P/E

—

EV / Sales

—

EV / EBITDA

—

P / FCF

—

PEG

—

Rule of 40

—

Risks

Severity 1 (low) to 5 (severe)

5
Chinese export-control exposure
4
Limited commercial monetization

Score components

What drove the composite

Growth100

Profitability20

Cash generation20

Balance sheet78

Moat / AI relevance88

Capital intensity (higher = lighter)20

Customer diversification60

Valuation discipline (dilution)42

Trend (MA50/MA200)50

Momentum (3/6/12m)50

Relative strength50

Realized vol (higher = lower vol)50

Segment peers

Other companies in Models