Fort Worth 24

collapse
Home / Daily News Analysis / DeepSeek Restarts Second Funding Round

DeepSeek Restarts Second Funding Round

Nov 02, 2026  Twila Rosenbaum  47 views
DeepSeek Restarts Second Funding Round

Hangzhou-based artificial intelligence start-up DeepSeek has restarted its second funding round, seeking nearly $8 billion (£6 billion), after pausing the process in July when remarks made to investors were leaked, according to local media reports.

The company is seeking a valuation of nearly 500 billion yuan, equivalent to about $74 billion or £55 billion. The reported target would mark a sharp step up from its first round, which closed earlier in the summer at a valuation of about 350 billion yuan. DeepSeek has not published detailed terms, and the figures cited in reports rely on unnamed sources.

The renewed fundraising effort comes as investors weigh the company’s low-cost model strategy against the capital demands of frontier AI development. DeepSeek gained international attention for training high-performance models at comparatively low expense and offering access at prices well below those charged by many competitors. That approach challenged assumptions held by some Western companies that have committed hundreds of billions of dollars to infrastructure for training and deploying models.

Key facts at a glance

  • DeepSeek has resumed its second funding round after pausing it in July.
  • The start-up is seeking nearly $8 billion (£6 billion).
  • It is targeting a valuation of nearly 500 billion yuan, or about $74 billion (£55 billion).
  • Monolith Management is reported to be in talks to contribute.
  • DeepSeek plans a large data centre in Inner Mongolia as part of efforts to secure 1 gigawatt of computing power.
  • The company has told users to expect a substantial price increase, without giving details.
  • Its first funding round closed earlier in the summer at a valuation of about 350 billion yuan.

Why the round was paused and restarted

The second round was suspended in July after the company’s remarks to investors were leaked. The episode underlined the sensitivity surrounding fundraising talks in the fast-moving AI sector, where valuations, compute plans and pricing strategies can shift quickly and where confidentiality is closely guarded. The restart suggests that DeepSeek has resolved the immediate disruption and is again seeking outside capital, though it remains unclear whether the terms have changed from the original process.

For a company that has built its reputation on efficiency, the decision to raise a large sum is significant. Training and operating advanced AI models requires substantial computing capacity, and access to chips and data-centre space has become a central constraint for AI developers in China and elsewhere. DeepSeek’s plan to secure 1 gigawatt of computing power illustrates the scale of infrastructure needed to support current models and future development.

Investor interest and the reported backer

Monolith Management, an investor in Beijing-based Moonshot AI, the developer of the Kimi line of AI models, is in talks to contribute to the round, according to reports. The potential involvement of an existing AI investor would signal continued appetite for DeepSeek’s prospects despite the company’s unconventional cost structure and the broader caution in parts of the venture market.

Still, the round is not confirmed publicly, and investor conversations can change. The reported $74 billion valuation would place DeepSeek among the most valuable private AI companies, but it would also raise questions about how quickly the company can convert technical attention into durable revenue. DeepSeek’s low prices have made its models accessible to developers and businesses, yet they also limit revenue per token. A higher price regime could improve unit economics, but it could also test customer loyalty in a market with more expensive alternatives.

The Inner Mongolia data centre plan

DeepSeek reportedly plans to construct a massive data centre in Inner Mongolia, part of a broader initiative to gain access to 1 gigawatt of computing power. The plan may include leasing capacity from third parties as well as building its own facilities. Some of the new funding would be used to finance the data centre, according to reports.

Inner Mongolia has attracted data-centre investment because of its energy resources and cooler climate, both of which can help reduce the cost of running power-intensive computing infrastructure. For AI companies, however, energy is only one part of the equation. Advanced chips, networking equipment, cooling systems and reliable operations all contribute to the total cost of ownership. Securing 1 gigawatt of capacity is a large undertaking, and the timeline, sourcing and financing of such a project remain unclear.

The move also reflects a wider race among Chinese AI developers to lock in compute. Access to advanced semiconductors has been constrained by export controls, prompting companies to design more efficient models, seek domestic suppliers and explore alternative arrangements. DeepSeek’s emphasis on efficiency has been a hallmark of its approach, but even efficient models require large-scale infrastructure to train and serve users at commercial scale.

Pricing strategy under pressure

DeepSeek told users last Thursday that it intended a substantial rise in prices, without specifying its upcoming pricing plans. The notice marks a potential shift for a company whose low prices have been central to its identity. DeepSeek currently charges $0.14 per million input tokens and $0.28 per million output tokens. By comparison, Moonshot’s latest Kimi K3 is priced at $3 per million input tokens and $15 per million output tokens, while Anthropic’s latest Fable 5 is priced at $10 per million input tokens and $50 per million output tokens.

The gap is striking. DeepSeek’s pricing has been far below that of several rivals, helping it attract developers who are sensitive to inference costs. A substantial increase would narrow that gap, though the company has not said how far prices will rise or when the change will take effect. Customers building applications on top of DeepSeek’s models will need to reassess their cost assumptions, particularly those running high-volume workloads.

Pricing is also a signal of strategic intent. Low prices can drive adoption and data collection, but they can compress margins and make it harder to fund heavy infrastructure. As DeepSeek contemplates a large data centre and a higher-cost compute base, a more sustainable price structure may become necessary. The challenge will be to raise prices without eroding the cost advantage that helped the company stand out.

Valuation and the funding environment

The reported target valuation of nearly 500 billion yuan would represent a substantial increase from the first round’s approximately 350 billion yuan. That jump, if confirmed, would reflect continued investor enthusiasm for AI despite a more selective funding environment. Private markets have grown more disciplined after a period of rapid valuations, and investors are increasingly focused on revenue, margins and capital efficiency.

DeepSeek’s case is unusual. It has achieved outsized visibility with a relatively lean public profile and a reputation for cost-efficient training. But the capital intensity of AI infrastructure means that even efficient companies may need large sums to compete at the frontier. The second round will test whether investors are willing to pay a premium for DeepSeek’s technical approach while accepting uncertainty around pricing, compute access and long-term profitability.

The company’s first round closed earlier in the summer at a valuation of about 350 billion yuan. The second round, if completed at the reported terms, would add significant new capital and a higher valuation marker. For now, the plans remain subject to negotiation and market conditions, and the company has yet to disclose final details.


Source: Silicon UK News


Share:

Your experience on this site will be improved by allowing cookies Cookie Policy