Chip revenues have been down across the largest semiconductor companies in recent quarters, but the bigger concern for policymakers and visionary investors is what happens to chip production in the coming two years and beyond.
Intel expressed the worst of it during first quarter earnings Thursday, with its largest quarterly loss in company history (a 133% annual drop in earnings per share) and revenue down 36% year over year to $11.7 billion. Yet, based on earlier expectations and a truly disastrous fourth quarter, the first quarter was “solid,” the company demurred, and shares jumped nearly 5% on Friday afternoon to $31.24.
RELATED: Chip revenues crash for Intel and Samsung in Q1
Longer term, CEO Pat Gelsinger is touting the company’s mission to become a productive fab for both its internal designs and designs of other companies to be a competitor to Taiwan’s TSMC and South Korea’s Samsung with cutting edge nodes down to 2nm and less and with novel chip packaging designs and the latest production technologies.
But the fruits of Intel’s labors will not really be known for years. Much of Intel’s future is tied up in making the US and Western Europe less dependent on Asian companies for cutting-edge chips, which was a major reason Gelsinger met with elected leaders to promote the CHIPS Act and other initiatives. Intel wants the government largesse, of course, to help finance expensive fabs.
Patriotism, at least for the US, is at work in what Intel is up to, and that patriotic sentiment makes some sense if China invades Taiwan someday and throws off the entire global economy for next-generation chips. US officials see domestic production of advanced chips critical and clearly have fingers crossed in hopes a possible (probable?) China-Taiwan military debacle is pushed beyond 2025, or hopefully, never takes place.
It is hard to think of a time in history when industry, diplomacy, geography and military strength have come in such close connection to each other. The global economy lives on the back of semiconductors and the latest generative AI apps run on the slickest chips, so what happens to Intel matters greatly.
“The US needs a strong Intel,” said Moor Insights and Strategy analyst Patrick Moorhead on a Friday Six Five Podcast. In an email to Fierce Electronics, he amplified: “The United States and Western Europe both need a strong Intel to produce leading edge semiconductors. TSMC in Taiwan and Samsung in Korea are both very good foundries but they are both at risk of invasion from China and North Korea.”
TSMC is in the process of building multibillion-dollar fabs in the US, but it has gone on record saying it will not produce any of its leading edge chips in the US for fear of IP theft, Moorhead added. “Once 2nm becomes mainstream, I do believe we will see it produced in the US by TSMC,” he added.
Intel has laid out plans to make chips as advanced as TSMC’s by 2026 so it can compete for custom jobs like the Apple A-series chips used in iPhones. On its first quarter call, Intel claimed that it is on track to hit its goals. “We still have more work to do as we reestablish process, product and cost leadership, but we continue to provide proof points each quarter,” Gelsinger added on the call with analysts.
However, Lucas Keh, semiconductors analyst at global research firm Third Bridge said Intel’s IFS faces multiple challenge in the next few years. The most pertinent challenge is building long-term trust with fabless chip customers (companies that could include Apple or Nvidia) who worry Intel will take their designs and turn around to produce them among Intel’s own product lines for a lower price, Keh said.
“In order for Intel to hope to have any competition with TSMC, they must provide this anti-competitive environment to ease this customer sentiment and win new business,” Keh added.
Intel already gets some credit for work in advanced packaging methods and chiplet architectures such as Foveros, Ponte Vecchio or its Hybrid Architecture used in Raptor and Meteor Lake to penetrate new markets. “But to try to compete with TSMC, this penetration will largely rely on execution and delivery, an area where Intel has seen massive failures,” Keh said.
One key failing was the delay in shipping Sapphire Rapids, which was resolved only recently, Keh said. “Customization to customer needs and specific applications will be key for Intel,” he added.
Given recent quarters and challenges facing IFS, could Intel flop before 2006?
Thinking broadly, Keh said investors and customers should expect, if all goes well, to see Intel’s new fabs up and running in four to five years. “Obviously, given macroeconomic turbulence, inflation and geopolitical factors, it is possible they flop,” he said.
“I don’t see Intel failing between now and 2005,” said Moorhead. “Intel is gaining share in PCs in the current state of the market. While AMD and Ampere are both taking share, I don’t see wild, double-digit swings, but rather, single digit swings. Finally, TSMC doesn’t have enough wafer or packaging capacity to replace Intel in 2023 to 2024.”
So, as both Key and Moorhead indicated, it is hard to envision progress for Intel or even the entire chip market in two to five years. Too many factors are at play to have a clear crystal ball. However, analysts at the Semiconductor Industry Association have tracked trends for years and see the overall industry growing at a healthy pace to reach $1 trillion in sales globally by 2030.
But to look at this one chipmaker, Intel, in two or 10 years is more difficult, partly because it has undergone so many changes and has multiple tough competitors. No wonder investors depend so heavily on quarterly earnings calls to make predictions where they only need reading glasses instead of telescopes for vision.