Intel’s Big Bet on Becoming Someone Else’s Factory
Intel built its reputation over five decades on a single, powerful premise: design the chip, own the factory, control the outcome. That vertical integration was the source of its pricing power, its engineering culture, and its competitive moat. Now, facing pressure from TSMC’s manufacturing dominance and Nvidia’s fabless success, Intel is preparing to spin off its foundry division into a separate legal entity – a move that would make Intel Foundry Services an independent business competing for external customers while still serving Intel’s own chip design teams.
The structural separation sounds clean on paper. In practice, it places Intel in an uncomfortable position: asking the semiconductor industry to trust a company that has repeatedly missed process node targets as a neutral manufacturing partner, while simultaneously asking its own investors to believe that splitting the business will unlock value rather than expose how deeply the two sides have been subsidizing each other.
Why the Spinoff Makes Sense – and Why It Doesn’t
The logic behind separating Intel’s foundry from its chip design business follows a well-established playbook. TSMC’s dominance was built precisely on neutrality – it has no competing chip products, which makes Apple, Qualcomm, and AMD comfortable handing over their most sensitive designs. Intel, as a combined entity, will always carry the suspicion that its foundry arm prioritizes its own silicon over external clients. A legally separate entity with its own board and financials theoretically addresses that concern.
The problem is that structural separation doesn’t automatically generate customer trust or manufacturing capability. Intel’s foundry has struggled with yield issues and process delays that pushed major clients toward TSMC and Samsung. Simply reorganizing the corporate chart doesn’t fix tool calibration, lithography throughput, or the engineering talent gaps that have accumulated over years of underinvestment in process development. Customers evaluating a foundry partner look at on-time delivery records and defect rates, not organizational diagrams.
There is also the question of who Intel Foundry’s anchor customers actually are. Without internal Intel chip orders, the division would be producing at far below economic capacity. With them, it remains tied to Intel’s design fortunes – which are themselves under pressure from AMD in data center CPUs and from Arm-based alternatives gaining ground across client computing. The foundry business and the chip design business are financially tangled in ways that a spinoff announcement cannot immediately undo.
The Manufacturing Capability Gap
Intel’s 18A process node – its most advanced – has been positioned as the technology that will close the gap with TSMC’s N2. Early test results have been watched closely by the industry, and while Intel has reported progress, no major external customer has publicly committed production volume to 18A at scale. That silence matters. Foundry economics only work with high utilization rates, and high utilization requires external customers willing to commit before a process is fully mature – a circular confidence problem Intel hasn’t yet broken.
TSMC, by contrast, begins taking customer orders for a new node while that node is still in risk production, because years of consistent delivery have earned it the benefit of the doubt. Intel is trying to rebuild that credibility from a lower starting point while simultaneously restructuring its corporate organization and managing a balance sheet that has been strained by capital expenditure commitments made during a more optimistic period for semiconductor demand.
What the Spinoff Actually Changes
Separating Intel Foundry into an independent entity does accomplish a few concrete things. It allows the division to raise outside capital independently, potentially attracting sovereign wealth funds or government investment tied to domestic chip production goals – particularly relevant given U.S. CHIPS Act funding tied to advanced manufacturing on American soil. It also creates clearer financial visibility into how much the foundry operation actually earns or loses, without internal transfer pricing obscuring the numbers.
That transparency could be a double-edged instrument. If Intel Foundry’s standalone financials reveal sustained operating losses even after accounting for internal chip orders, it becomes much harder to argue the business is viable without a sustained period of below-cost pricing to attract external customers. The foundry business in its current state is likely not self-funding. Making that explicit through a spinoff structure doesn’t solve the problem – it just makes the problem harder to ignore at the board level.
There is a scenario where the spinoff works. If Intel 18A delivers competitive yields, if the U.S. government continues channeling CHIPS Act subsidies toward Intel’s fabs, and if geopolitical pressure on TSMC’s Taiwan concentration pushes even one or two major American chip companies to dual-source with Intel, the foundry could reach the utilization levels needed for economic sustainability. That scenario requires multiple favorable variables aligning simultaneously over a 3-to-5 year window – a long time for a company burning capital at Intel’s current rate.
The deeper tension in this story is about identity. Intel’s engineering culture was built around designing and building its own chips – the two activities informed each other, with process engineers and chip architects working in close proximity. Separating those teams into different legal entities, with different incentives and different P&L accountability, changes how those engineers interact and what they optimize for. A foundry optimizes for yield, throughput, and customer flexibility. A fabless chip designer optimizes for performance per watt and time to market. Those priorities don’t always point in the same direction, and Intel is betting it can manage both simultaneously in organizations that no longer share a balance sheet.
Whether Intel’s foundry ambitions can survive the structural split may ultimately depend less on corporate governance than on whether 18A’s performance data, when fully disclosed, gives the industry a genuine reason to diversify away from TSMC – or whether it confirms that Intel is still chasing a lead that has already moved two nodes further ahead.
Frequently Asked Questions
Why is Intel spinning off its foundry division?
Intel is separating its foundry into an independent entity to attract external customers and capital, and to remove the perception that it favors its own chip designs over outside clients.
Can Intel Foundry compete with TSMC?
Intel’s 18A process node is its most advanced attempt to close the gap, but no major external customer has publicly committed volume production to it yet, leaving TSMC’s lead intact for now.
