Youkeshu Technology Co. Ltd.

It earns money two structurally different ways today: one-time bulk sales of physical goods sourced from domestic factories and resold through overseas distribution channels, and fee-for-service software development and maintenance work. It is layering a third, usage-metered computing-rental model on top, charging through short-term rental, dedicated long-term capacity, private deployment and consumption-based billing, though that business's large signed contracts have yet to convert into recognized revenue.

It scales along two different paths at once: the trading side grows by adding sourcing and distribution relationships across overseas channels, consuming working capital roughly in proportion to volume, while the newer computing side grows by committing large amounts of capital to hardware and facilities ahead of revenue and then converting that capacity into billed usage over long contracts, tying growth to financing and cash management rather than to marginal cost. Its recent financial history includes years of net losses rather than steady compounding returns, and CompanyGraph's own reading of its current balance sheet shows cash sitting high relative to near-term obligations and total assets, a pattern consistent with a company holding a liquidity buffer around its capital-heavy expansion.

Its trading business depends on domestic manufacturing factories and brand owners for the goods it resells, and on third-party marketplaces such as Amazon, AliExpress and Shopee, platforms it does not control, to reach consumer buyers. Its computing business depends on external GPU hardware, including a named chip maker, Muxi Integrated Circuit, for which it holds domestic distribution rights, and on specialized technical talent in liquid cooling and heterogeneous computing that the company itself describes as scarce; CompanyGraph separately maps a broad base of upstream industries feeding this company beyond the specific suppliers its filings name.

Buyers span two businesses: on the trade side, overseas online-platform agents, offline channel distributors and end consumers reached through third-party marketplaces; on the computing side, domestic large-model companies, AI technology firms, telecommunications operators and AI institutions, including a named counterparty, Zhejiang Shenhu Technology, in one large equipment-sales agreement, while several other major contract counterparties are disclosed only as anonymized codes. No single named customer accounts for a dominant share of revenue, and even its largest several customers together remain a minority of sales.

The company states its own differentiators as combining its trade and computing businesses, supporting heavy-asset investment with external capital, sourcing globally, and integrating hardware deployment with software and maintenance end to end. This is the company's own claim about what sets it apart; CompanyGraph has no independent evidence about which of these capabilities its rivals do or do not have, so whether they are actually hard to copy cannot be assessed here.

The company's own account describes a split in how it sells computing capacity: dedicated clusters and private deployments, where a customer's workloads are built into specific infrastructure, sit alongside short-term rental and consumption-metered billing, which is designed to be flexible. Read structurally, the first group is more likely to carry real switching costs, and the second is designed not to; the company's own materials do not describe any switching-cost or lock-in mechanism for the separate bulk trade business.

The company itself points to capital, not customer retention, as its main limit: computing infrastructure requires large upfront spending with a long payback period, and cash gets tied up between paying for equipment and collecting from customers, while scarce specialized talent in areas like liquid cooling and heterogeneous computing is named as a second constraint on how fast it can grow. CompanyGraph's industry classification would otherwise suggest customer retention and switching costs as the starting hypothesis for this sector, but the company's own account emphasizes capital intensity and talent scarcity instead.

The company itself names a timing mismatch as a specific vulnerability: it pays for core computing hardware on terms set upstream while collecting from downstream customers on a different, slower timeline, which it says can strain cash. It also names reliance on external financing to fund heavy equipment purchases and dependence on retaining scarce R&D and project talent to deliver large computing orders on time, and it discloses two unresolved legal disputes, one an arbitration brought by a logistics counterparty, Shenzhen Qianhai Yuntu Logistics, and one a contract dispute brought by its own subsidiary against other named parties.

The company names securities regulators and the exchange it lists on as direct overseers, alongside customs, import-export, intellectual-property, cross-border data and overseas tax rules it must meet in the markets it trades into. It separately names global trade-policy shifts, tariffs and other trade barriers, and currency movements between the renminbi and foreign currencies as pressures that can reduce orders or raise costs, and discloses unresolved legal disputes with outside counterparties.