Intel’s Foundry Unit Posts 31% Growth, but Losses Still Run Deep

Intel’s push to become a major supplier of chip manufacturing capacity is showing signs of momentum, but the economics remain unforgiving. In a report published Monday, the company’s foundry business grew 31% in the most recent quarter while still losing about $2.1 billion in the period, highlighting the gap between ramp growth and durable profitability.

The foundry operation is Intel Foundry Services (IFS), Intel’s effort to manufacture chips for customers as well as for its own products. Unlike a typical standalone manufacturer, IFS initially depends heavily on internal orders. According to the report, almost every dollar of the foundry unit’s revenue came from Intel’s own business, meaning external customer demand has not yet become the dominant driver.

The article also pointed to a still-worsening, but improving, loss rate. It said the factory arm’s loss was roughly half as large per revenue dollar as it was a year earlier. That kind of metric is often used to gauge whether operating costs, depreciation, and start-up inefficiencies are easing as factories get fuller and processes mature, but it also implies Intel has not yet reached a scale where operating losses begin to disappear quickly.

Intel’s foundry growth, therefore, looks more like a ramp than a turnaround. Revenue increases can occur as wafer starts, capacity utilization, and product demand inside Intel rise, even if the unit remains loss-making. In that scenario, investors tend to focus on whether the growth is increasingly supported by third-party customers, rather than primarily by Intel’s own chip needs.

The shift Intel is aiming for is structural. For a foundry model to become economically self-sustaining, a greater share of output typically needs to come from external clients who commit to production volumes. The report’s emphasis that most foundry revenue still originates within Intel suggests that the external customer base is not yet big enough to meaningfully change the unit economics.

Sector context matters because the semiconductor manufacturing business is capital-intensive and time-consuming. Scaling capacity requires large upfront spending, long technology qualification cycles, and careful yield improvement. Even when customers begin placing orders, the ramp can take multiple quarters for those commitments to translate into higher utilization and better margins across the manufacturing footprint.

Still, the quarter-by-quarter picture can remain mixed. A foundry unit may show revenue growth while losses persist, and even improving loss per revenue dollar can coexist with a large absolute loss if the unit is still at an early stage relative to its cost base. The report’s framing suggests Intel is making progress on efficiency, but the absolute gap has not narrowed enough to produce profits.

What Intel does not disclose in the report itself is as important as what it does. The publication does not provide additional breakdowns in the excerpted framing, such as how much revenue came from specific process nodes, whether margin trends differ between customer and internal production, or how much of the $2.1 billion loss is tied to depreciation versus operating costs. Those details, typically found in company financial footnotes, would be needed to assess whether the improvement is sustainable or largely tied to short-term ramp dynamics.

Looking ahead, investors and analysts will likely watch whether Intel Foundry can increase the share of revenues coming from outside customers and whether improvements in loss per revenue dollar continue. If Intel’s external wins translate into higher utilization and better yield on a broader scale, the operating picture could shift more decisively. If not, growth may continue to come with substantial cash burn even as efficiency improves.