Flex builds things for other companies. It does not sell products under its own brand. Instead, it runs factories in about 30 countries and charges customers each time it designs, assembles, or ships a product on their behalf. Those customers include data center operators, car makers, medical device companies, and consumer electronics brands. Flex earns money on every order it fulfills, which means its revenue rises and falls with what its customers decide to build. The diagram below traces where the money goes.

Five years of financial data tell a story with two distinct chapters. From 2022 to 2025, revenue moved in a narrow band between $24.6 billion and $28.5 billion, rising sharply in 2023 when component shortages pushed up order volumes, then slipping back as demand softened in consumer electronics and automotive. Free cash flow was thin throughout that stretch, hitting a low of $0.3 billion in 2023. Gross margins hovered around 7 percent, which is typical for a contract manufacturer competing on price and scale.

The second chapter started quietly and accelerated fast. Flex built a new segment called Cloud and Power Infrastructure, which makes power systems, cooling equipment, and integrated compute racks for data centers and AI workloads. That segment generated $3.2 billion in revenue in 2024, then $4.8 billion in 2025, then $6.6 billion in 2026, a 38 percent jump in a single year. As that higher-margin business grew, it pulled the whole company's gross margin up from 7.1 percent in 2024 to 9.2 percent in 2026. Free cash flow recovered too, reaching $1.1 billion in both 2025 and 2026.

The operating cash story improved alongside margins. Flex generated $1.0 billion in operating cash in 2022, dipped to $0.9 billion in 2023, then climbed steadily to $1.7 billion by 2026. That improvement matters because Flex needs cash to invest in specialized manufacturing equipment, fund working capital for large new contracts, and service its debt.

Now Flex plans to split itself into two separate publicly traded companies. The Cloud and Power Infrastructure segment would become its own independent business. The remaining company would keep the traditional manufacturing operations serving automotive, healthcare, industrial, and consumer electronics customers. The split is targeted for the first quarter of 2027, but it still needs approval from Flex's board, shareholders, and a Singapore court, as well as clearance from the United States Securities and Exchange Commission.

The planned split creates real risks alongside the potential upside. Running a separation of this scale requires management attention that could otherwise go toward winning new contracts and running factories efficiently. If the split is delayed or cancelled, the uncertainty itself creates problems. If it goes ahead, both resulting companies will be smaller, which could make each one more vulnerable if their specific markets soften. The traditional manufacturing business, stripped of its fast-growing data center segment, would have lower average margins than the combined company does today.

Three other risks are documented and specific. First, Flex's ten largest customers account for 45 percent of its sales, and those customers include very large cloud companies with enormous bargaining power. Those customers can demand lower prices, change their order volumes, or decide to build things themselves. Second, Flex sources some components from single suppliers or from suppliers that its customers choose, which means it has limited ability to find alternatives if something goes wrong. A missile strike on its Ukraine facility in August 2025 caused $51 million in charges and showed how physical disruptions can hit the business without warning. Third, the data center business depends on customers continuing to spend heavily on AI infrastructure, and that spending could slow if electricity becomes scarce, water for cooling runs short, or if the pace of AI adoption turns out to be slower than today's order books suggest.

The traditional manufacturing segments face a separate structural pressure. In the Integrated Technology Solutions segment, which covers consumer electronics and communications gear, revenue fell from $12.6 billion in 2024 to $11.1 billion in 2026 as consumer demand weakened. The Automotive business inside the Regulated Manufacturing Solutions segment also declined 2 percent in 2026. These are cyclical markets, meaning their revenues move up and down with the broader economy and with product cycles.