Nvidia and Intel sidestepping questions about manufacturing in their new deal only reinforces that Intel's best path forward is to spin off its foundry business
The case for splitting one of the tech sector's most storied firms in two — If Intel is to rekindle its lost glory …
Context & Ripple Effects
Intel’s foundry separation debate predates this agreement: the company had already explored splitting product and factory operations amid broader restructuring options. The case sharpened after the foundry reported a $13B-plus loss in 2024, underscoring the gap between its manufacturing ambitions and current economics.
The new Nvidia arrangement matters less as proof of foundry demand than as a conspicuous absence of a disclosed manufacturing commitment. That follows reporting that Nvidia had tested Intel’s 18A process but did not proceed, leaving the foundry’s customer-validation challenge unresolved.
First-order effects
- Intel gains a commercial relationship with Nvidia but no publicly stated manufacturing endorsement, limiting the deal’s immediate usefulness as evidence that its foundry can win major external chip orders.
- Nvidia preserves sourcing and product flexibility by avoiding a disclosed Intel manufacturing commitment while pursuing its own AI-hardware roadmap.
Second-order effects
- The lack of a manufacturing component strengthens pressure on Intel to separate the capital-intensive foundry from its product business, a path previously proposed in calls for a foundry spin-off backed by purchase guarantees.
- Potential foundry customers and policymakers receive less signal that a flagship AI-chip buyer will anchor Intel capacity, making the economics of new factory investment harder to demonstrate.
Third-order effects
- If major chip designers continue to treat Intel as a partner without becoming fabrication customers, the industry may move toward a clearer split between Intel’s chip-design operations and a separately accountable manufacturing business.
- The episode highlights the difficulty of building a domestically significant foundry through industrial-policy support alone: durable capacity depends on repeat external customer commitments, not strategic announcements.
The trend: AI-era semiconductor partnerships are increasingly separating technology collaboration from manufacturing commitments, raising the premium on foundries that can independently attract anchor customers.