ON Semiconductor Corporation

Revenue comes almost entirely from one-time sales of semiconductor products, with a much smaller share from product-development agreements. It is spread unevenly across power, analog-and-mixed-signal and sensing product lines, power the largest of these and sensing the smallest, and reaches customers through both direct accounts and independent distributors in broadly similar proportions. The company has stayed profitable in each of the last several years covered here.

As a producer whose output is capped by physical plant capacity, CompanyGraph reads its scaling as moving in discrete steps rather than smoothly: it has added capacity and technology through acquisitions of a silicon-carbide business and rights to a power-technology line, while separately running a realignment program that cut manufacturing capacity and staff elsewhere in its network. Its cash-flow pattern also shows it converting a larger share of revenue into operating cash than most peers in its industry group, while putting a smaller share of that cash into capital spending than most peers do, a combination CompanyGraph reads as relatively capital-light cash conversion for a plant-based producer. It sits within a large, well-populated category of companies that scale the same way, by expanding or shrinking physical plant capacity.

The company's own filings name dependence on suppliers of silicon and silicon-carbide wafers, substrates, metals, lead frames, mold compound, ceramic packaging, chemicals and gases, which feed both its own plants and outside contract manufacturers used mainly for wafer fabrication and finished-goods assembly and test. Those filings also flag the manufacturing network itself as a source of risk: interdependent owned and third-party facilities include single-source sites, such as an internal location that alone produces a silicon-carbide crystal material used elsewhere in the network. It further depends on independent distributors to reach much of its customer base, and names substantial operations and foreign-currency cost exposure across Asia-Pacific and Europe. Beyond named inputs, CompanyGraph's mapping of the industries that feed into this one shows it draws on a wide base of supplying sectors relative to the narrower set it sells into.

Its own account names direct customers as large multinational companies and selected regional OEMs, alongside independent distributors who resell onward to OEMs, contract manufacturers and other end customers in automotive, industrial and other markets, including AI data centers, computing, consumer, and networking and communications. Of its end markets, automotive is the largest source of revenue and industrial the second largest, by the company's own account. Beyond named customers, CompanyGraph's mapping of the industries it supplies shows a narrower base of buying industries downstream, compared with a wider base of industries it draws inputs from upstream.

CompanyGraph places this company in a very large, common category: many other producers run production systems shaped the same way, converting material inputs into outputs inside fixed physical plant, so the basic shape of the business is not unusual by itself. In its own filings, the company points to breadth of intellectual property, manufacturing capability across fabrication and packaging, and long-standing customer and technical-support relationships as what it considers its strengths, particularly in its power and sensing product lines. CompanyGraph has no independent basis in the evidence gathered here to say whether these are actually difficult for competitors to replicate, since that would require visibility into rival companies that this evidence does not include.

By its own account, the limit named most directly is supply and manufacturing capacity: suppliers may extend lead times, limit supply or raise prices when they face capacity constraints of their own, and revenue expected from long-term supply agreements can be affected by manufacturing or supply-chain constraints, changes to customer agreements, and regulatory change. This is consistent with a broader expectation for producers that convert inputs to outputs inside fixed plant, that their scale is ultimately bound by how much that plant can physically process, though that broader expectation is an industry-level one rather than something measured for this company beyond what its own filing states.

By its own account, the risks it names first include reliance on interdependent owned and third-party manufacturing facilities and suppliers, some of them single-source, so a disruption at one node can affect output elsewhere in the network. One of its assembly-and-test sites is also organized as a majority-owned joint venture rather than a wholly owned facility. Its revenue is concentrated in two end markets, automotive most of all and industrial next, so pressure on demand or pricing in either one would weigh heavily on results given how much of revenue the two represent together. It further names substantial operations and foreign-currency cost exposure across Asia-Pacific and Europe, and is defending litigation alleging misleading statements about its silicon-carbide business specifically.

The company's own account discloses active securities litigation, Hubacek v. On Semiconductor Corp, alleging misleading statements about its silicon-carbide business, along with related shareholder derivative claims that are paused pending that case. A motion to dismiss the amended complaint is fully briefed. It also names substantial operations and foreign-currency cost exposure in Asia-Pacific and Europe as a pressure on its results. Separately, and as a general expectation for the industry rather than something measured about this company specifically, producers whose output is bound by physical plant capacity typically face pressure from the gap between what a plant can physically process and what demand or input supply allow it to run at, and from the margin between conversion cost and selling price.