DeepSeek raises $16 billion in funding at $74 billion valuation

A massive capital injection

DeepSeek has closed a significant funding round totaling approximately 105.5 billion yuan, which converts to nearly $16 billion. This final tally was reported by The Information and other outlets between October 22 and October 30, 2026. The amount represents a major increase from the 50 billion yuan the company initially sought for the round.

The valuation of the company remains a point of some discrepancy among reporting outlets. Some sources cite a pre-money valuation of roughly 500 billion yuan, or $74 billion, while others peg it higher at approximately 615.5 billion yuan. These figures suggest a significant markup from the company's first funding round in June, which was valued at around 350 billion yuan.

This capital injection marks the second time DeepSeek has accepted outside investment. By restricting the round to state-backed and corporate funds, the lab is prioritizing strategic partners over financial investors. This move ensures that the company's cap table is populated by entities that can provide essential infrastructure like chips, memory, and distribution channels.

The industrial supply chain

The investor list reveals a clear strategy focused on the physical supply chain of AI hardware and consumer distribution. Tencent and the battery maker CATL provided large checks, with Tencent contributing 10 billion yuan in each of the two funding rounds. Other notable backers include SMIC, which manufactures chips, and ChangXin Memory Technologies, which produces memory components.

Beyond hardware, the round includes heavyweights in the consumer electronics and services space. Companies like vivo, Meituan, Geely Automobile, and Midea Group are participating as investors. These entities can integrate DeepSeek models into a wide array of products, including smartphones, apps, vehicles, and household appliances, creating a built-in customer base.

The inclusion of telecom carriers like China Mobile and China Telecom further reinforces this distribution play. These carriers operate massive subscriber bases and cloud businesses, providing a direct channel for a domestic model. This mix of investors suggests that DeepSeek is being positioned as a foundational layer for the broader Chinese tech economy.

State backing and chip costs

One of the most notable shifts in this round is the jump in state involvement. State-owned institutions and government-backed funds supplied 27% of the capital, a sharp increase from just 2% in the previous financing round. This includes contributions from five major institutions, including China Life and CICC, as well as the national AI industry fund.

The funding was reportedly upsized because of the extreme costs associated with acquiring high-end hardware. Reports indicate that black-market prices for Nvidia Blackwell hardware in China have surged significantly compared to US prices. DeepSeek needed a larger war chest to navigate these high costs and provide a financial buffer against the premium prices of gray-market chips.

The involvement of SMIC and ChangXin Memory also suggests a push toward domestic silicon. By investing in DeepSeek, these suppliers gain a flagship customer to help improve yields and performance for homegrown chips. This creates a feedback loop where the lab's success directly supports the development of a self-reliant domestic hardware ecosystem.

Efficiency vs. Scale

DeepSeek originally gained industry attention in early 2025 by demonstrating that strong AI results could be achieved at a lower cost than Western labs. This reputation for frugality is now being tested by a funding round that rivals the scale of well-known Western competitors. The shift suggests that while architecture matters, frontier competition eventually demands massive capital for talent and compute.

The lab recently released V4.1 Flash, a model designed to balance high capabilities with faster inference speeds. This release is significant because cheap, quick inference is what allows partners like Meituan or China Mobile to deploy models at scale. Investors are paying for this momentum, looking for a balance between the lab's research-first culture and corporate demands for product integration.

For the broader market, this round shows that capital is still flowing to model builders outside the US hyperscaler orbit. It creates significant exposure for Chinese listed companies like Tencent and CATL to a private AI asset. Analysts will now be watching how these companies book their stakes and how the lab manages its transition from a lean lab to a capital-heavy powerhouse.