Enflame Gets More Than 80% of Revenue From Tencent. Is That a Risk?

 

A company derived 83.79% of its 2025 revenue from a single customer, and that customer is also its largest shareholder. The combination alone is enough to draw regulatory scrutiny. It is not, however, the most alarming figure in this prospectus.

 

Per Beijing News reporting, Enflame Technology's top five customers accounted for 96.50%, 92.60% and 96.89% of revenue across 2023 through 2025, with sales to Tencent Technology (Shenzhen), including both direct sales and sales under the AVAP model, representing 33.34%, 37.77% and 83.79% respectively. Tencent Technology (Shenzhen) is a related party, and both channels constitute related-party transactions. The same report notes the Shanghai Stock Exchange has formally asked whether the company's existing AI chips are highly adapted to Tencent and whether it has the capability to develop and adapt to other customers.

 

What deserves unpacking sits one layer down. In 2024 total revenue was RMB 722 million with RMB 273 million from Tencent; in 2025 total revenue was RMB 990 million with RMB 830 million from Tencent. Subtracting from the disclosed figures, non-Tencent revenue fell from roughly RMB 450 million to roughly RMB 161 million, a decline of about 64%. This is simple arithmetic from prospectus data and should be verified against the official filing, but it points to a more fundamental problem than concentration itself.

 

 

 

Tencent Technology (Shenzhen) accounted for 83.79% of revenue in 2025, against 37.77% in 2024, 33.34% in 2023 and 8.53% in 2022, a one-directional climb.

 

The corresponding amounts were RMB 100 million, RMB 273 million and RMB 830 million, against total 2025 revenue of RMB 990 million.

 

Top five customers accounted for 96.50%, 92.60% and 96.89% of revenue across the three years, never falling below 90%.

 

Calculated from public data, non-Tencent revenue fell from roughly RMB 450 million in 2024 to roughly RMB 161 million in 2025, a decline of about 64%, while total revenue grew roughly 37%.

 

Tencent Technology is the largest shareholder, holding roughly 20.26% together with its concert party Suzhou Paiyi, making the related sales connected transactions.

 

Revenue grew from RMB 301 million in 2023 to RMB 990 million in 2025, a three-year compound rate of 81.32%, while net losses attributable to shareholders were RMB 1.665 billion, RMB 1.510 billion and RMB 1.164 billion.

 

First quarter 2026 revenue was RMB 287 million, up 1,474.85% year over year, with a net loss of RMB 444 million, and the company guides first-half revenue of RMB 1.06 billion to RMB 1.15 billion.

 

The STAR Market IPO application cleared the listing committee on June 15, 2026, was submitted for registration that month and received registration approval in July, seeking RMB 6.0 billion with CITIC Securities as sponsor.

 

 

 

Looking only at 83.79% invites the conclusion that the company depends heavily on one customer, but that framing is too broad. The trajectory carries more information.

 

Per Ifeng's summary, the revenue share from Tencent Technology (Shenzhen) surged from 33.34% in 2023 to 83.79% in 2025, meaning that for every RMB 100 of 2025 revenue, nearly RMB 84 came from Tencent. Beijing News supplies the earlier starting point: from 2022 through 2024 and the first three quarters of 2025, the ratios were 8.53%, 33.34%, 37.77% and 71.84%.

 

Four years from 8.53% to 83.79% with no year of retreat. For a company approaching public markets, the healthy path is concentration declining as scale grows. What appears here runs the other way.

 

 

Beyond the single customer, the broader structure warrants attention. Beijing News shows the top five accounting for 96.50%, 92.60% and 96.89% over the three years.

 

One analysis pushed the observation further. Per a piece carried on Sina Finance, top five concentration exceeded 92% for four consecutive years while the names on the list changed almost entirely each year with only Tencent holding the top spot every year, implying no genuinely stable repeat customer base outside Tencent. This is third-party interpretation, and the actual composition of the customer list should be verified against the prospectus.

 

The company's own explanation partly corroborates it. The prospectus attributes high concentration to large individual contract values in the intelligent computing systems and cluster business and to large-scale deployment of its products in internet customers' AI application scenarios. The first reason explains the annual turnover mechanically: project-based large contracts inherently lack year-to-year continuity.

 

 

 

This is the calculation readers should verify for themselves. Per Sina Finance, combined direct and AVAP-model sales to Tencent were RMB 100 million, RMB 273 million and RMB 830 million across 2023 to 2025, while total revenue grew from RMB 301 million to RMB 990 million.

 

Subtracting Tencent revenue from the total: non-Tencent revenue was roughly RMB 450 million in 2024 and roughly RMB 161 million in 2025, a decline of about 64%. Total revenue over the same period rose from RMB 722 million to RMB 990 million, growth of about 37%.

 

In other words, all of 2025's revenue growth came from Tencent alone, while outside Tencent the business did not stagnate but contracted materially. This is a simple arithmetic illustration from disclosed data whose precise definitions and reconciliation should be verified against the prospectus, but it converts the static description of high concentration into a dynamic question: is market-facing commercial capability deteriorating?

 

The Shanghai Stock Exchange's inquiry is directly relevant. Beijing News shows the exchange asking whether existing AI chips are highly adapted to Tencent and whether the company has the capability to develop and adapt to other customers, and separately questioning the reasonableness of the sharp increase in Tencent-related transactions. The framing itself indicates regulators are focused on product generality rather than the percentage.

 

 

Another dataset reinforces this. Per a separate Ifeng report, revenue across 2023 to 2025 was RMB 300 million, RMB 722 million and RMB 990 million for an average compound growth rate of 81.5%, but 2025 growth slowed to 37%.

 

Total growth decelerating from triple digits to 37% while single-customer share climbed from 37.77% to 83.79% means the base of growth is narrowing rather than broadening.

 

 

Tencent's dual role pushes this beyond ordinary customer concentration. Per Sina Finance, Tencent is both Enflame's largest shareholder and its largest customer. Beijing News gives specifics: prior to this offering, Tencent Technology is the largest shareholder holding 20.26% combined with concert party Suzhou Paiyi. Other summaries cite Tencent Technology (Shanghai) at 16.9569%, so sources differ on the holding entity and percentage, and readers should rely on the prospectus.

 

The structure cuts both ways. On the positive side, Huxiu's summary notes the company began working with Tencent in 2019 and progressed from small-scale single-scenario validation through multi-scenario large-scale validation to deep strategic cooperation, forming a stable relationship through multiple product generations and sustained software optimisation. For an early-stage domestic chip designer, an anchor customer willing to supply genuine high-concurrency workloads is difficult to replicate.

 

The side requiring caution is pricing. Beijing News notes that both direct and AVAP-model sales to Tencent Technology (Shenzhen) constitute related-party transactions, and raises the question of how the company avoids improper benefit transfer through connected transactions, noting the company had not responded as of publication. Third-party analysis has further questioned whether a gross margin holding near 36% while a single customer exceeds 83% of revenue implies non-market pricing arrangements in related-party transactions. That is analytical scepticism rather than an established conclusion, and the regulatory responses and prospectus disclosures are the authoritative sources.

 

Additionally, Ifeng's summary notes the company sells not only directly to Tencent but also through the AVAP model to server makers designated by Tencent, with Huxiu listing Customer A, Customer D and Customer J. Actual exposure to the Tencent ecosystem may therefore be broader than the direct sales figure indicates.

 

For investors seeking background on the company's business model and how it trades, this explainer on Enflame Technology and how to trade it provides a starting point for understanding the risk structure.

 

 

 

Loss scale is another necessary dimension. Ifeng's summary shows net losses of RMB 1.665 billion, RMB 1.51 billion and RMB 1.164 billion across 2023 to 2025, cumulative losses above RMB 4.3 billion, with R&D expenses of RMB 1.23 billion, RMB 1.41 billion and RMB 1.35 billion, keeping R&D above 130% of revenue throughout. Huxiu adds that unrecovered losses on a consolidated basis reached RMB 4.441 billion at the end of 2025.

 

Sina Finance puts cumulative R&D at RMB 3.676 billion, or 1.83 times cumulative revenue over the same period. A company with roughly RMB 2 billion of three-year revenue spending close to RMB 3.7 billion on R&D indicates that current pricing power does not yet cover the technology investment.

 

The loss trend is narrowing. Ifeng's summary notes the company stating in its inquiry response that, based on an order backlog exceeding 100,000 chips, mass production of the third-generation S60 inference accelerator card and contributions from kilo-card intelligent computing centre projects, it expects to reach breakeven as early as 2026. Sina Finance reports the company projecting consolidated profitability in 2026 or 2027.

 

Operating quality shows improvement too. Ifeng's summary notes accounts receivable as a proportion of revenue falling from 82.15% at the end of 2023 to 29.34% at the end of 2025. One caveat: separate third-party analysis cites cash collection ratios of 0.45, 0.48 and 0.48 across 2023 to 2025, framing the picture differently. This article cannot fully reconcile the two measures, and prospectus data should be treated as authoritative.

 

The 2026 data suggests another possibility. Per reporting carried by ZAKER, first quarter 2026 revenue was RMB 287 million, up 1,474.85% year over year, with a net loss of RMB 444 million, and the company expects first-half revenue of RMB 1.06 billion to RMB 1.15 billion, growth of 258.68% to 289.13%. First-half guidance already approaches or exceeds full-year 2025.

 

 

Judging whether concentration can fall requires examining specific non-Tencent orders. Ifeng's summary discloses two verifiable projects: City A's phase two project, one of eight national computing hub nodes, repurchased 313 sets of third-generation intelligent computing systems with a tax-inclusive contract value of RMB 280.6045 million, delivered and accepted in December 2024; and in June 2025 the company signed an investment agreement with City B's government and partners for an intelligent computing centre project with total investment of RMB 500 million and computing scale of no less than 4,000P, expected to generate roughly RMB 180 million to RMB 200 million of revenue.

 

The nature of these projects matters. Both are government-led intelligent computing centre builds, operating on different logic from commercial procurement by internet customers, with sustainability depending on fiscal spending cadence rather than competitive outcomes. Note also that the RMB 280.6 million contract was delivered and accepted in December 2024 and therefore falls in a prior period, while the RMB 180 million to RMB 200 million remains unrealised.

 

The product line is also evolving. Eastmoney's summary shows AI accelerator card and module revenue rising from RMB 186 million in 2023 to RMB 856 million in 2025, growth exceeding 360%, and notes products currently serve inference scenarios with the fourth generation, its first combined training and inference product, about to enter mass production alongside a fifth generation in development. A combined training and inference product theoretically widens the addressable customer base, but market validation has not begun.

 

 

The first risk is demand substitutability. Tencent's procurement follows its own capital spending cycle, and any adjustment transmits directly. On the 2025 structure, if Tencent purchases went to zero, that year's revenue would fall below RMB 160 million. This is an extreme hypothetical derived from public data and reflects no expectation.

 

The second is product generality. The Shanghai Stock Exchange's inquiry has already put this on the table, asking whether existing AI chips are highly adapted to Tencent and whether the company can develop and adapt to other customers. The year-over-year decline in non-Tencent revenue is an unfavourable footnote to that question.

 

The third is related-party pricing. Tencent is simultaneously shareholder and customer, the related sales are connected transactions, and how market-based the pricing mechanism is remains a regulatory focus that directly affects the comparability of gross margin data.

 

The fourth is the profitability timeline. The company expects breakeven as early as 2026 and consolidated profitability in 2026 or 2027, but unrecovered losses reached RMB 4.441 billion at the end of 2025, so even after turning profitable, clearing historical losses would take considerable time.

 

On scenarios, the base case is 2026 revenue growing sharply on continued Tencent volume with concentration staying elevated, shifting attention to the absolute size of non-Tencent revenue. A second is the combined training and inference product plus intelligent computing centre projects delivering genuine customer expansion, pulling concentration back from 83.79% and shifting the valuation logic from single-customer dependence toward a scaled domestic substitution narrative. A third is Tencent's procurement cadence slowing without non-Tencent business filling the gap, where the current cost structure and R&D intensity could interrupt the narrowing loss trend.

 

Status distinctions: the financial data, customer percentages and regulatory inquiry content come from officially filed documents. The June 15, 2026 committee approval, the registration submission that month and the July registration approval are all on public record. As of writing, no formal announcement of pricing or listing date has been identified, and any statement about listing timing or valuation should rely on exchange and company announcements. Third-party valuation estimates and interpretations are those institutions' judgements rather than statements of fact.

 

 

The most important number in this prospectus is not 83.79%. It is a difference you have to calculate yourself. Subtract Tencent revenue of RMB 273 million from 2024 total revenue of RMB 722 million and non-Tencent revenue is roughly RMB 450 million. Subtract RMB 830 million from 2025's RMB 990 million and it is roughly RMB 161 million. A decline of about 64%. In the same year total revenue grew 37%, the business outside Tencent shrank by nearly two-thirds. Rising concentration has two possible causes: the large customer growing faster, or other customers being lost. Those carry entirely different implications, and the data here points to the second.

 

Three misreadings look likely. The first is treating high concentration itself as the core risk. For an early-stage domestic chip designer, an anchor customer supplying genuine high-concurrency workloads is difficult to replicate, and the company's relationship with Tencent since 2019 progressed through a full cycle from small-scale validation to deep cooperation, which is itself evidence of product maturity. The real risk is not having one large customer but having no stable second growth curve beyond it. The second is underweighting the valuation impact of shareholder and customer overlap. Tencent and its concert party hold roughly 20.26% and the related sales are connected transactions, meaning gross margin, payment terms and pricing all require additional verification, and those are among the most sensitive inputs in any valuation model. The third is treating first quarter 2026 growth of 1,474.85% as proof of successful expansion. Against an extremely low prior-year base, that rate mainly reflects delivery cadence rather than customer structure, and judging expansion requires the absolute level of non-Tencent revenue rather than the growth rate.

 

Three verifiable metrics deserve tracking. First, the direction of the Tencent share in the 2026 interim report and subsequent disclosures. The company has already stated in its prospectus that it expects the high Tencent share to persist for some period, so a fall below 70% would indicate expansion is working while holding above 80% or rising would indicate the structural issue remains. Second, the absolute level of non-Tencent revenue. From a base of roughly RMB 161 million in 2025, the pace of recovery on that line says more about market-facing capability than total revenue growth does. Third, the customer composition of the fourth-generation combined training and inference product entering mass production, where whether the first customers come from outside the Tencent ecosystem is the most direct evidence of product generality.

 

For cross-asset investors, the lesson is that revenue concentration must be assessed alongside the nature of that revenue. Eighty percent of revenue from a single counterparty means something entirely different when that counterparty won a competitive tender than when it is also the largest shareholder, because the latter's procurement decisions incorporate non-commercial factors that may not persist at the same intensity after listing. The logic maps directly onto crypto, where many protocols' locked value or trading volume depends heavily on a few market makers or affiliated entities, creating a gap between headline metrics and market-driven demand that public data cannot easily separate. When evaluating names like this, establish who the counterparty is and what its relationship to the company is before judging whether that business survives a change in the relationship. That sequence gets closer to the substance of the risk than tabulating concentration percentages. From a risk management standpoint, when shareholder and customer overlap, run the valuation once more with the related-party contribution stripped out; the difference between the two results is the risk premium attached to that structure.

 

This analysis rests on published prospectus disclosures, regulatory inquiry records and credible reporting available now. Subsequent financial data, changes in customer composition and listing progress could each change the conclusion, and no single scenario should be treated as a fixed expectation.

 

 

 

Sales to Tencent Technology (Shenzhen), including direct sales and sales under the AVAP model, accounted for 83.79% of revenue in 2025, roughly RMB 830 million against total revenue of RMB 990 million. That share was 8.53% in 2022, 33.34% in 2023, 37.77% in 2024 and 71.84% in the first three quarters of 2025, a one-directional climb. Top five customers accounted for 96.50%, 92.60% and 96.89% over the three years.

 

 

Partly. The company explains in its prospectus that high concentration stems from large individual contract values in the intelligent computing systems and cluster business and from large-scale deployment in internet customers' AI applications, describing the situation as similar to comparable companies. The internet sector genuinely is the primary source of domestic AI compute demand. The distinction, though, is that top-five concentration is common while a single customer exceeding 80% is not.

 

 

Calculated from public disclosures, non-Tencent revenue was roughly RMB 450 million in 2024 and roughly RMB 161 million in 2025, a decline of about 64%, while total revenue grew roughly 37%. This is simple subtraction from prospectus figures and the precise definitions should be verified against the official filing. The direction implies all of 2025's growth came from Tencent while other customers contracted, which is the context for the exchange's inquiry into product generality.

 

 

Principally related-party transactions and pricing. Tencent Technology is the largest shareholder holding roughly 20.26% with its concert party, and the related sales are connected transactions. Reporting has raised the question of how the company avoids improper benefit transfer, noting no response as of publication. Third-party analysis has also questioned whether a gross margin near 36% is plausible with a single customer above 83% of revenue. These are analytical questions, and regulatory responses and prospectus disclosures are authoritative.

 

 

The company's stated timeline is breakeven as early as 2026 and consolidated profitability in 2026 or 2027, based on an order backlog exceeding 100,000 chips, mass production of the third-generation S60 inference accelerator card and contributions from kilo-card intelligent computing centre projects. For context, net losses across 2023 to 2025 were RMB 1.665 billion, RMB 1.510 billion and RMB 1.164 billion, with unrecovered consolidated losses reaching RMB 4.441 billion at the end of 2025.

 

 

Heavy R&D investment. R&D expenses across 2023 to 2025 were RMB 1.23 billion, RMB 1.41 billion and RMB 1.35 billion, keeping R&D above 130% of revenue throughout and totalling roughly RMB 3.676 billion, or 1.83 times cumulative revenue over the same period. The high fixed-cost nature of the chip industry contributes as well. That intensity indicates current pricing power does not yet cover the technology investment and improvement requires greater scale.

 

 

The STAR Market application was accepted on January 22, 2026, the first A-share IPO accepted that year, cleared the listing committee on June 15, was submitted for registration that month and received registration approval in July. The company seeks RMB 6.0 billion with CITIC Securities as sponsor, offering new shares equal to no less than 10% and no more than 15% of post-issue share capital. As of writing, no formal announcement of pricing or listing date has been identified, so exchange and company announcements are the authoritative source.

 

 

Three verifiable directions. First, the direction of the Tencent share in subsequent disclosures, where a fall below 70% would indicate expansion is working. Second, the absolute level of non-Tencent revenue, since from a base of roughly RMB 161 million in 2025 its recovery pace says more about market-facing capability than total growth does. Third, the customer composition of the fourth-generation combined training and inference product, where first customers from outside the Tencent ecosystem would be the most direct evidence of product generality.

 

 

This article is provided for informational and research purposes only and does not constitute investment advice, financial advice, legal advice, tax advice or any recommendation to transact. The financial data, customer composition, shareholding percentages and regulatory inquiry content referenced here come from publicly disclosed prospectus filings, inquiry responses and credible reporting; the company's official filings are the authoritative source; sources differ on the holding entity, percentages and certain financial measures, which the text notes where applicable; every subtraction and ratio calculation from public data in this article is a simple arithmetic illustration whose precise definitions and reconciliation should be verified against the prospectus and which does not constitute a determination about the company's operations; third-party analytical questions, valuation estimates and views are those institutions' judgements rather than statements of fact; and as of writing no formal announcement of pricing or listing date has been identified. Prices of crypto assets, equities and other related financial instruments can move sharply over short periods, particularly for unprofitable companies with highly concentrated customers, and newly listed securities are especially prone to severe volatility in early trading, with the possibility of total loss of principal. Futures contracts referencing a stock carry leverage risk that magnifies losses proportionally, holding such contracts is not equity ownership and confers no shareholder rights, and the risk profile differs fundamentally from holding shares directly. Historical performance, technical indicators and on-chain data cannot guarantee future outcomes. Readers should conduct their own independent research, verify official information directly, and evaluate any decision against their own financial circumstances, investment objectives, experience and risk tolerance, consulting a qualified professional adviser where appropriate. The MEXC Crypto Pulse team accepts no liability for any direct or indirect loss arising from use of or reliance on the information in this article.

 

 

James Mitchell specializes in technical analysis, market trends, and trading strategies for both Bitcoin and altcoins. Based in London, he has over 10 years of experience in financial markets. Before joining MEXC Learn, James worked as a senior analyst at a leading European investment firm, where he developed expertise in risk management and quantitative trading. His transition to cryptocurrency markets began in 2017, and he has since become recognized for his data-driven approach. He holds a Master's degree in Financial Economics from the London School of Economics. His analytical approach combines traditional technical analysis with on-chain metrics to provide readers with actionable insights.

 

Areas of Expertise:

 

 

Want the fastest access to MEXC's latest updates? Join our official Telegram group now!

MEXC Platform: Download App | Zero Fee Trading | Buy USDT Now

Join MEXC Community: X (Twitter) | Telegram | Discord

Get Insights: MEXC Learn | Terms of Service | MEXC Blog | MEXC News

Account Verification: Understand KYC | How to Complete KYC

External Content Platforms: Substack | Medium | Paragraph | LinkedIn | X(News)