Sina published a market note on September 24 that attributed an accelerated annualized revenue run rate at DeepSeek to recent pricing moves and product demand, and also described a planned second financing round and preparations for a Shanghai or STAR Market listing. The update rests on anonymous sources and investor-meeting disclosures, and independent open filings or a company statement were not cited in the items reviewed for this report. The Sina note is available at a Chinese financial news feed and provides the core timeline that underlies much of the coverage in September. Separately, Reuters reported that DeepSeek had tapped CITIC Securities to prepare for a potential Shanghai STAR Market IPO. Because these accounts rely on unnamed sources and meeting disclosures rather than regulatory filings, readers should treat reported figures as provisional.
The reports raise questions about how DeepSeek would fund expansion and how investors would assess a potential domestic listing, but the available accounts provide reported targets rather than audited financial disclosures or completed deal documentation.
The central claims rest on a September 24 Sina account and must be handled as reported figures. A separate Reuters report on Shanghai STAR Market preparations provides earlier context but does not confirm the current fundraising target. For adjacent archive context, readers can consult an earlier EastFrontier DeepSeek item and a second related archive item; neither is used as factual support for the current update.
Run-rate and pricing
Sina reported that DeepSeek’s annualized revenue run rate reached $1 billion, a rise from under $500 million earlier in the year, and attributed the acceleration partly to an increase in API model-call prices that the report described as 2.3 to 4.5 times prior levels. The article said the number came from two people familiar with the matter and that DeepSeek CEO Liang Wenfeng disclosed the figure at a recent investor meeting subject to confidentiality constraints. Media summaries repeated the $1 billion run-rate figure while noting the same investor-meeting origin for the disclosure. Because the published accounts rely on unnamed sources and a meeting recap rather than a regulatory filing or company disclosure, the run-rate should be treated as a reported metric rather than audited or recognized revenue.
The coverage paired the pricing changes with claims that customers did not abandon the platform after the hikes and that demand for larger models remained robust, according to the meeting account. Sina reported that nearly all revenue came from paid model API calls, while the free, ad-free chatbot app generated no revenue. Those revenue-mix claims are reported measures and are subject to the caveat that the underlying articles did not provide audit detail or vendor statements. For readers who want to consult the original Chinese market note, the Sina item can be accessed directly. For context on the listing preparations and adviser engagement, see the Reuters item that summarized the company tapping a domestic securities firm to prepare for an IPO.
Fundraising and IPO plans
Sina said DeepSeek was advancing a second financing round with a target of ¥50 billion and a stated valuation target of ¥500 billion, and it reported a goal of closing the round before the end of October. Dealroom and other contemporaneous summaries explicitly stated that the proposed raise had not closed and presented alternative dollar conversions that differ from Sina’s renminbi presentation. For context on earlier capital moves, Reuters reported in September that DeepSeek had tapped CITIC Securities to prepare for a Shanghai STAR Market IPO and that the company raised about $7.4 billion in June at a post-money valuation that was reported at that time. Those items are distinct from the September financing target and should not be conflated with a completed transaction.
The discrepancy in dollar conversions across outlets underscores exchange-rate and reporting differences; the underlying renminbi targets are the clear point of reference in the sourced coverage. The Sina article’s headline used wording that suggested a completed financing while its body described an advancing round and a target timetable. Dealroom’s statement that the round was not closed supports treating the September funding as a reported plan rather than a finalized deal. EastFrontier covered DeepSeek’s reported August funding round separately and identified that earlier financing as one phase in a broader capital campaign, which investors will view as context for any new round. Reporters and market participants should therefore treat the September targets as fundraising objectives that may change before any close.
Product, margins, and compute
Sina reported that DeepSeek V4.1-Flash and V4-Flash attracted global customers seeking lower-cost AI-agent systems, and described the models as potentially able to benchmark against larger models; these are source-attributed claims, not independently verified benchmarks or customer evidence. The product-performance and competitive-comparison language in the articles came from investor-meeting statements and anonymous sources and should not be treated as independently benchmarked results. The reporting also included an internal metric that placed API-call gross margin for the first seven months of the year at a high level, and it said in-house figures showed a larger share of compute dedicated to training than to inference over that period. A separate media summary published in July placed annualized revenue for an earlier slice of the year in a lower range, and contemporaneous summaries cited different renminbi or dollar measures for overlapping intervals. Those numbers appear to reflect different measures or time slices and were presented in the open reporting as distinct items rather than a reconciled accounting.
Sina reported an 82.9% API-call gross margin for the first seven months and reported that Liang said more than 70% of compute was allocated to training; the implications for capital needs and scale remain analytical rather than established facts. As of the supplied reporting, DeepSeek was described as pursuing a proposed second financing round and reported Shanghai/STAR Market IPO preparations. Market reaction and investor calculus will likely hinge on verification of the reported figures and on clarity about how DeepSeek will spend any new proceeds. The September reporting describes a firm pursuing ambitious capital targets and domestic listing steps while citing internal metrics that, if confirmed, could support a large valuation. However, the published items stop short of documentary confirmation.
The September reports place DeepSeek at a pivotal point: reported revenue momentum, reported pricing changes, and reported fundraising targets have reopened questions about how the company will finance continued model development and go-to-market expansion and about how regulators and investors will assess a potential Shanghai or STAR Market listing. Because the available accounts rest on anonymous sources and investor-meeting disclosures rather than on formal filings, the market will need clearer documentation or company statements before treating the reported run rate and the proposed round as definitive.