AXT, Inc.

AXT earns most of its revenue from finished substrate wafers, with a smaller, secondary stream from selling raw materials on their own rather than converting them further into substrates. It sells under short-term customer purchase orders rather than long-term contracts, priced deal by deal through negotiation, formula or list pricing, with revenue recognized only once control of the goods actually transfers to the buyer. Very little of its future revenue is therefore locked in by contract at any given point.

AXT scales by physically building more crystal-growing and wafer-finishing capacity inside China, which takes construction time and customer requalification before the added output can actually be sold, rather than by a lighter mechanism like software or licensing. Its recently reported earnings have run ahead of the cash the business itself generates, and it has funded new capacity in part by raising money externally rather than relying only on operating cash flow. CompanyGraph groups AXT with a large set of companies that scale the same physical way, expanding output only as fast as capacity can be built, qualified and brought on line.

AXT's own account names a set of raw-material supply companies in China, several of which are partly owned affiliates rather than independent third parties, that produce gallium and related inputs it depends on. Separately, it names quartz tubing, arsenic, phosphorus and polishing solutions as inputs it can source from only one or a limited number of suppliers. All of its manufacturing sits inside China, so it also depends on Chinese government permitting and export-control decisions simply to move its own finished products out of the country. CompanyGraph separately maps its supply chain as reaching into a small number of upstream industries beyond these named suppliers.

AXT's direct customers are mainly companies that apply epitaxial layers to its substrates before selling the modified wafers on to wafer fabs, chip designers and LED makers, along with universities and research organizations that buy substrates directly. Its own account describes that customer base as spread across a number of individual buyers rather than concentrated in one or two, with its largest handful together making up a meaningful but not dominant share of revenue. CompanyGraph separately maps AXT as feeding several downstream industries.

AXT's basic way of running a plant, converting feedstock into finished product within a capacity ceiling, is a common shape; CompanyGraph places a large number of companies in that same broad category. AXT itself names Sumitomo Electric Industries, Japan Energy (JX), Freiberger Compound Materials, Umicore, China Crystal Technology and Vital Materials as its primary competitors, and claims a narrower distinction against them: that only a few suppliers worldwide can qualify as indium-phosphide substrate makers, and that it is the only compound-semiconductor substrate maker that also holds a stake reaching into its own raw-material supply. Whether rivals could replicate that particular combination, or simply have not, is not something this evidence can settle.

AXT's own account describes new customers going through a lengthy internal evaluation before they will place an order, so switching to an entirely new, unqualified supplier is slow by its own description. At the same time, it says customers typically keep more than one supplier already qualified against the same quality, delivery and support requirements, so once a customer has qualified more than one source, moving volume between those already-approved suppliers is not slowed by that same evaluation process.

Companies that convert fixed inputs into output at a capped physical rate are generally limited by that physical ceiling; CompanyGraph tests whether this holds for each specific company rather than assuming it. AXT's own account points to several specific limits rather than plant capacity alone: export permits, the length of customer qualification, the availability of raw materials, its ability to add and use manufacturing capacity, manufacturing yields, and access to qualified technical staff. It gives two concrete instances of a limit binding: demand for one substrate type rising while the revenue it earned from that product fell because export permitting stood between the two, and its own decision to sell less of another wafer type after the cost of the raw material behind it rose substantially.

AXT's own account lists what it names first among its risks: broad economic and political conditions including tariffs and import-export restrictions, intervention and fast-changing rules from the Chinese government, and the possibility of complications tied to its shares being listed in the United States while its operations and audit sit in China. Because every one of its manufacturing sites is inside China, and because a small number of suppliers provide inputs it says it cannot easily source elsewhere, its own account points to disruption where Chinese regulation and its own supply chain meet. It also notes that its customers are not obligated to buy any set quantity or give it binding forecasts, so the demand it currently serves is not contractually secured.

AXT's own filings describe pressure from two governments at once: Chinese export-permit requirements now cover all of its wafer families, following China's decision to add indium-phosphide substrates to its export control list, while it separately names volatile United States tariffs and import restrictions, and the possibility of further Chinese export controls, as live exposures. Because all of its manufacturing sits in China, it also names PRC government intervention, fast-changing rules, and currency-transfer approval requirements among the conditions it operates under, alongside a pending shareholder lawsuit and a related appeal.