The market’s broad classification of artificial intelligence winners and losers has left some good businesses overlooked. Maurits Ovaa from our offshore partner, Orbis, explains how the Orbis Optimal SA Fund is finding opportunities.
The stock market has been sorting companies into artificial intelligence (AI) winners and losers with a coarse sieve, and some good businesses have fallen through it. This quarter, we highlight two significant positions built in the Orbis Optimal SA Fund this year.
United Integrated Services (UIS) is a Taiwanese industrial supplier that builds cleanrooms inside semiconductor factories, or fabs. A leading-edge fab is less a building than a giant piece of laboratory equipment. Particle count limits in its advanced cleanrooms are thousands of times tighter than in a hospital operating theatre, and vibration is carefully controlled so that lithography tools costing US$200 million each can hold sub-nanometre accuracy. UIS is a key manufacturing partner for Taiwan Semiconductor Manufacturing Company (TSMC), involved in every advanced fab that TSMC has constructed in Taiwan, Arizona and Japan.
Demand for AI computing is driving TSMC’s capital spending higher, while each new generation of chip technology requires more cleanroom space built to higher specifications. On our estimates, UIS’s revenue should grow by around 60% this year and 30% next year, with earnings per share up by more than half and then by more than a third. The pipeline of fabs already announced supports double-digit growth for several years beyond that. The company has no borrowings and practically no debt, a substantial net cash position and pays out around 85% of its earnings as dividends.
Yet the shares trade at only 13 times next year’s earnings. Were UIS trading at a higher multiple, we would worry more about what happens when growth slows, but at current valuations, we are paying very little for the company’s growth runway. If earnings compound as we expect and investors come to value UIS as the reliable specialist it is, the shares could more than double over the next four years.
UIS’s revenue follows the timing of fab construction, so the clear near-term risk to our projections would be that a pause between fabs opens an air pocket, as it did in 2021 and 2024. Margins have roughly doubled since 2023, and management is candid that part of that gain could be given back in a weaker market. Geopolitics is another risk, though its franchise lies more in its people and expertise than its physical plants. We nonetheless think the risk of permanent capital loss is low. If TSMC’s spending simply reverted to its 10-year trend rather than accelerating, we estimate UIS would still be worth almost twice the current valuation. If margins also fell back to the mid-teens while the shares derated to 10 times earnings, we would still expect to break even on our investment.
UIS plays an essential role in semiconductor expansion but sits outside the usual technology universe. That is exactly the overlooked opportunity we like.
Finding potential among the so-called AI losers
Experian sits in another part of what we view as the AI misclassification spectrum. The company holds credit records on some 1.5 billion consumers and 175 million businesses. About three-quarters of its revenue comes from businesses, with the rest from consumers through a free credit-score app with more than 200 million members and a marketplace for loans and cards. The same chief executive and finance director have run it since 2014 and, over the past decade, revenue has doubled at margins above 25%.
Why, then, does the stock appear cheap? Two letters: AI. Between July 2025 and March 2026, the shares fell by around 40% from 30 times forward earnings to just 17 times. Investors worried that AI would enable banks to underwrite on their own data, that generative models would commoditise Experian’s analytics software, or that new tools would bypass the credit bureaus altogether. Experian sold off in step with other data and information companies, seemingly without the market making a distinction between their competitive positions.
We think these concerns underestimate the durability of Experian’s core asset. Its credit bureau is a contributory database: Virtually every major lender supplies data because it needs access to everyone else’s. The system is tightly regulated, and it has taken decades to build. AI can make better sense of that data, but it cannot generate borrowers’ actual repayment histories. AI could also increase the demand for this data by allowing banks to refresh their credit models more frequently and feed in more variables each time. Additionally, Experian charges by usage rather than by seat.
The numbers agree. Revenue grew organically by 8% last year and earnings per share by 15%. When many of Experian’s largest North American financial services clients came up for renewal last year, everyone renewed. On average, the new contracts were longer and higher in value than before. The company itself uses AI internally, with margins continuously expanding. These are not the numbers of a company being disrupted.
Our due diligence included extensive primary research with industry practitioners, which reinforced our view that bureau data remains deeply embedded in lenders’ operations, with little appetite to reduce reliance on it.
Today, the shares remain at 17 times forward earnings, against an average of 25 to 30 times over the preceding five years. A free cash flow yield of more than 5%, growing at a double-digit rate, could compound our capital in the mid-teens annually, even absent a recovery in the multiple. If disruption fears fade as we expect, a rerating would come on top, potentially doubling the shares over the next four years. We cannot predict when the clouds lift. But in our experience, they usually do, and while we wait, we are being paid to be patient by a business compounding its earnings rapidly.
As long as the market continues sorting companies into AI winners and losers with a coarse sieve, good businesses will keep dropping through, and you can count on us to seek opportunities to catch them.