{"id":"1e9bc5ec-550c-4b27-9033-4ea00f8e33aa","name":"DeepSeek","slug":"deepseek","description":"DeepSeek is a Hangzhou-based AI model developer spun out of the Chinese quantitative hedge fund High-Flyer in July 2023 by founder Liang Wenfeng. Unlike the US neocloud pack (CoreWeave, Lambda, Crusoe) that rents GPUs by the hour to third parties, DeepSeek runs its own fleet as a captive research resource: the Fire-Flyer 2 cluster built out from 2021 onward, roughly 10,000 Nvidia A100 GPUs plus a later top-up of H800 accelerators procured before US export controls tightened. The lab does not sell compute; it sells inference on its own frontier models through a low-priced API.\\n\\nThe company vaulted into global view in January 2025 when DeepSeek-R1 briefly displaced ChatGPT as the top free iOS download and Chinese researchers claimed a training run for the underlying V3 base model costing roughly $5.6 million, an order of magnitude below Western frontier labs. That efficiency claim, whether taken at face value or read as a narrow marginal-cost figure, triggered a one-day $600B drop in Nvidia's market cap and reframed the \"how much compute do you need for a frontier model\" debate. Successive releases (V3.1, V3.2, the V4-Flash and V4-Pro tiers in 2026) have kept the cadence.\\n\\nDeepSeek is privately held and until 2026 was funded entirely off High-Flyer's balance sheet. In May 2026 it closed a $7B Series A at a $52B post-money valuation and began IPO preparation targeting a 2027 listing. Its distinctive posture in the neocloud/AI-infra map is that it is an operator of its own inference stack rather than a rentable cloud, and it is doing frontier model work under Chinese export-restricted silicon, which has made it both a proof point for compute-efficient training and a target of US and Australian government use bans.","status":"private","founded":2023,"hq":"Hangzhou, China","hqLat":30.2489634,"hqLng":120.2052342,"hqGeocodedAt":"2026-10-10T04:21:20.562Z","hqAddress":null,"hqPrecision":null,"website":"https://www.deepseek.com","fundingTotal":"7000000000","type":"brain","roles":[],"reviewStatus":"reviewed","heatIndex":0,"lifecycleState":"active","supersededByCompanyId":null,"sources":[{"url":"https://en.wikipedia.org/wiki/DeepSeek","title":"DeepSeek — Wikipedia"},{"url":"https://www.deepseek.com","title":"DeepSeek official site"},{"url":"https://arxiv.org/abs/2412.19437","title":"DeepSeek-V3 Technical Report"},{"url":"https://news.bloomberglaw.com/capital-markets/deepseek-to-raise-at-least-12-billion-in-tencent-backed-funding","title":"DeepSeek to Raise at Least $12 Billion in Tencent-Backed Funding","sourceName":"Bloomberg Law","publishedAt":"2026-10-06"},{"url":"https://www.cnbctv18.com/business/chinas-deepseek-to-raise-12-billion-in-tencent-backed-funding-20005498.htm","title":"China's DeepSeek to raise $12 billion in Tencent-backed funding","sourceName":"CNBC TV18 (Bloomberg syndication)","publishedAt":"2026-10-06"},{"url":"https://qz.com/deepseek-funding-round-tencent-catl-ipo-100626","title":"DeepSeek raising $12 billion ahead of 2027 IPO","sourceName":"Quartz (Bloomberg paraphrase)","publishedAt":"2026-10"},{"url":"https://x.com/NextBillionLab/status/2108709096274952493","title":"DeepSeek $12B / Huawei DC note","sourceName":"X/@NextBillionLab","publishedAt":"2026-10-09"}],"keyFacts":[{"label":"Parent company","value":"Spun out of Chinese quantitative hedge fund High-Flyer in July 2023; founder Liang Wenfeng also runs High-Flyer."},{"label":"GPU inventory","value":"Fire-Flyer 2 cluster of ~10,000 Nvidia A100s (625 nodes of 8) built from 2021, later supplemented with H800s, connected over InfiniBand."},{"label":"Silicon","value":"Nvidia-only stack (A100, H800); no disclosed AMD, Cerebras or Chinese-domestic accelerators, and no public H100/H200/B200 numbers."},{"label":"Positioning","value":"A training-focused frontier lab with a captive research fleet; sells low-priced hosted inference through its API and app, not a rentable GPU cloud."},{"label":"Series A valuation (May 2026)","value":"The $7B round closed at a $52B post-money valuation; earlier April 2026 talks had targeted $70B pre-money."},{"label":"IPO preparation (Jul 2026)","value":"Began IPO preparation in July 2026, targeting a 2027 listing."},{"label":"V3 training cost (company-stated, Dec 2024)","value":"$5.576M for the DeepSeek-V3 base-model training run; widely disputed as a marginal cost that excludes prior R&D and cluster capex."},{"label":"Latest releases (Aug 2026)","value":"DeepSeek-V4-Pro (1.6T parameters) released August 13, 2026, after V4-Flash (284B) on July 31, 2026."},{"label":"R1 launch impact (Jan 2025)","value":"R1 launched January 20, 2025, briefly displaced ChatGPT as top free iOS app by January 27, and coincided with a one-day ~$600B drop in Nvidia's market cap."},{"label":"Government restrictions","value":"The US FY2026 NDAA bars DeepSeek from Department of Defense systems; Australia issued a government-wide removal directive in 2025."},{"label":"Huawei DC ~160k accelerators (reported, Bloomberg)","value":"Bloomberg-cited plan to deploy at least about 160,000 Huawei top accelerators at a massive Inner Mongolia data center build the round is meant to fund. No Ulanqab GW figure used here."},{"label":"Early-2027 IPO prep (reported, unconfirmed)","value":"Bloomberg sources say after the round closes DeepSeek plans to restructure toward an early-2027 IPO. Timing and listing venue not company-confirmed in these clips."}],"aliases":[],"collisionRisk":"low","reviewNote":null,"atsProvider":null,"atsSlug":null,"factoryLocations":null,"manufacturingCapacity":null,"orgType":null,"burnSignal":null,"litigationEvents":null,"insurancePostureDisclosed":null,"unitEconomicsDisclosed":null,"tickerSymbol":null,"stockExchange":null,"secCik":null,"revenueUsd":null,"revenueBasis":null,"revenueAsOf":null,"employeeCount":160,"employeeCountBasis":"reported headcount, Wikipedia synthesis of Chinese press","employeeCountAsOf":"2025-01-01T00:00:00.000Z","marketCapUsd":null,"marketCapAsOf":null,"marketCapSource":null,"createdAt":"2026-08-28T01:48:32.056Z","updatedAt":"2026-10-10T20:54:34.367Z","jsonLd":{"@context":"https://schema.org","@type":"Organization","@id":"https://registry.deploy.report/companies/deepseek","url":"https://registry.deploy.report/companies/deepseek","name":"DeepSeek","description":"DeepSeek is a Hangzhou-based AI model developer spun out of the Chinese quantitative hedge fund High-Flyer in July 2023 by founder Liang Wenfeng. Unlike the US neocloud pack (CoreWeave, Lambda, Crusoe) that rents GPUs by the hour to third parties, DeepSeek runs its own fleet as a captive research resource: the Fire-Flyer 2 cluster built out from 2021 onward, roughly 10,000 Nvidia A100 GPUs plus a later top-up of H800 accelerators procured before US export controls tightened. The lab does not sell compute; it sells inference on its own frontier models through a low-priced API.\\n\\nThe company vaulted into global view in January 2025 when DeepSeek-R1 briefly displaced ChatGPT as the top free iOS download and Chinese researchers claimed a training run for the underlying V3 base model costing roughly $5.6 million, an order of magnitude below Western frontier labs. That efficiency claim, whether taken at face value or read as a narrow marginal-cost figure, triggered a one-day $600B drop in Nvidia's market cap and reframed the \"how much compute do you need for a frontier model\" debate. Successive releases (V3.1, V3.2, the V4-Flash and V4-Pro tiers in 2026) have kept the cadence.\\n\\nDeepSeek is privately held and until 2026 was funded entirely off High-Flyer's balance sheet. In May 2026 it closed a $7B Series A at a $52B post-money valuation and began IPO preparation targeting a 2027 listing. Its distinctive posture in the neocloud/AI-infra map is that it is an operator of its own inference stack rather than a rentable cloud, and it is doing frontier model work under Chinese export-restricted silicon, which has made it both a proof point for compute-efficient training and a target of US and Australian government use bans.","identifier":"1e9bc5ec-550c-4b27-9033-4ea00f8e33aa","foundingDate":"2023","address":"Hangzhou, China","publisher":{"@id":"https://deploy.report/#organization"}},"framework_metadata":{"framework_schema_version":"0.1.0","verification_status":"verified","maturity_stage":null,"lifecycle_state":null,"architectural_position":{"cohort":null,"sub_cohorts":[]},"within_cohort_verified_vs_claimed_pair":null,"cap_flags":[],"verification_depth":{"sources_count":7,"primary_source_types":["knowledge-base","preprint"]}}}