According to ExtremeTech, analysts like Patrick Moorhead of Moor Insights & Strategy are heaping praise on Intel’s future 14A process node, calling it “the real deal.” This optimism stems from groundwork laid by the current 18A node, which Moorhead believes can stack up against any leading-edge production node globally. However, a major issue remains: the 18A node is largely unproven and lacks major external customers lined up to buy it. Intel CEO Pat Gelsinger has reportedly shifted the company’s focus to 14A, putting most of its eggs in that basket. He’s even suggested that if 14A isn’t a commercial success, it might be Intel’s last new process node. The company has secured backing from Nvidia and the U.S. government and convinced Apple to use its fabs for future chips.
The Confidence Gap
Here’s the thing: the analyst praise for 18A sounds great. “Industry leadership,” “tens of millions of leading-edge chips.” That’s the language Intel desperately wants to hear. But there’s a glaring disconnect between the technical confidence and the commercial reality. If 18A is so good, where are the customers? It’s one thing for an Intel-focused analyst to talk to Intel’s existing partners; it’s another to pry a major design win away from TSMC. That’s the real test. And right now, that test is still pending. It’s like having a critically acclaimed restaurant that’s always empty. The reviews might be stellar, but you start to wonder.
The 14A All-Or-Nothing Bet
So, Intel’s leadership is making a huge, public pivot. They’re basically saying, “18A is the foundation, but 14A is the whole house.” Gelsinger’s statement that failure on 14A could mean the end of Intel’s process development is staggering. It turns the next node into a existential company bet. That kind of pressure does two things: it focuses the entire organization like a laser, and it makes potential customers nervous. Who wants to commit to a platform the CEO has hinted could be a dead end? The logic from analysts is that success with 18A will naturally lead customers to 14A. But in the cutthroat foundry business, logic doesn’t always win. Execution and proven reliability do. For companies designing multi-billion dollar chip products, stability is often more attractive than a leap to the next big thing, especially from a supplier rebuilding its reputation.
The Long Road Back
Look, the backing from Nvidia and Apple is undeniably huge. It’s validation. But let’s be clear: Apple using Intel’s fabs is not the same as Apple using Intel’s process node. They’re likely bringing their own chip designs to Intel’s manufacturing floor—a foundry customer. That’s the model Intel needs to master. The government money helps build the floor, but it doesn’t fill it with other people’s work. Moorhead’s optimism is based on what he’s seen and who he’s talked to, and that carries weight. But the semiconductor industry is littered with “next big things” that stumbled on the ramp to volume production. Intel’s promise is finally matching its ambition. Now it has to deliver, consistently, for someone other than itself. The next 18-24 months will tell us if the praise was prescient or premature.
