Intel and Apple's Un-Breakup

The Factory, Not the Chip

Ordinarily, when you have a corporate breakup, you don't look back. In 2020, Apple very publicly dumped Intel—a move that foreshadowed Intel's eventual decline—by ditching its processors to build its own superior M-series chips. And yet, this week brought a development so strange it demands attention: the two are reportedly back at the negotiating table, discussing a potential Apple investment in Intel.

The reason for this apparent reunion is not about getting back together. It is a pragmatic solution to a different, and much larger, strategic problem. The key to understanding the talks is to realize that Apple doesn't want Intel's chips. It wants its US factories.

The basic situation is that Apple's entire hardware empire is now critically dependent on a single point of failure: TSMC's advanced foundries in Taiwan. This is, of course, a massive geopolitical risk. A viable, US-based, leading-edge chip manufacturer is the perfect hedge, and the only company on the planet with a plausible shot at becoming one is Intel.

For Apple, an investment in Intel's foundry business allows the company to help build the TSMC alternative it needs without ever having to put an Intel processor in a Mac again. It is a way to solve a massive supply chain problem while simultaneously satisfying political pressure to invest in America's semiconductor comeback. Apple CEO Tim Cook has previously signaled his support for the idea. Here is Bloomberg:

Apple has sought to show that it’s investing heavily in the US — even as much of its production remains overseas. At a White House event in August, the company announced plans to spend $600 billion on domestic initiatives over a four-year period, up from a previous pledge of $500 billion.

...Tim Cook told CNBC's Jim Cramer that the investments would encourage other companies to add US production, creating a "domino effect."

When asked about Intel, he said that competition would be good for the chip foundry industry. "We’d love to see Intel come back," Cook said.

This is a story about industrial policy and geopolitical risk. The US government is already a 10% owner of Intel. Nvidia, a TSMC mega-customer, just invested $5 billion in Intel. What's happening is the slow formation of a "Team America" consortium, backed by both public and private capital, to prop up the only domestic foundry that matters. Apple may have divorced Intel's products, but it is now being conscripted into a national service partnership with its factories.

Supplying the Enemy

Ordinarily, when you spend five years developing a secret weapon, the goal is to use it against your enemies, not sell it to them. This week, Mercedes-Benz took its secret weapon—an elite Silicon Valley team designing the next generation of energy-efficient chips for self-driving cars—and set it up as an independent arms dealer.

The new company, Athos Silicon, was spun out with a "significant" investment and a parting gift of valuable intellectual property from its former parent. Its new mission is not to give Mercedes an exclusive edge, but to sell its advanced chiplet technology to the entire automotive industry, including Mercedes' direct competitors.

The technology itself is aimed at solving a critical problem in the electric vehicle transition. In an EV, where every component is fighting for a slice of the battery, power efficiency is paramount. As the new Athos CEO Charnjiv Bangar put it, "For an electric future, electricity is a new currency." His team's designs reportedly use 10 to 20 times less power than conventional chips.

Which makes the decision to set it free all the more interesting. The move, perhaps, is a tacit admission of a hard-learned lesson from the tech world: you cannot be a world-class chip designer and a world-class car company at the same time. To truly compete at the highest level of silicon design, a company needs the scale and feedback that only selling to the entire market can provide. An in-house team, serving a single automaker, is destined to be a niche player. Athos needed to be independent to succeed. Here is its CEO on that very point, via Reuters:

"Independence is important for Athos, so that we can reach out to other (carmakers), competitors of Mercedes. We need to make sure we have a neutral approach," Bangar said.

This is a calculated sacrifice of exclusivity in exchange for excellence and upside. As a customer, Mercedes ensures it has access to what could become the industry's best-in-class technology, sharpened by the demands of the open market. And as a minority shareholder, it gets a financial return on every chip Athos sells, whether it ends up in a Mercedes or a BMW. It is a sign that the century-old, vertically integrated model of the auto industry is starting to unbundle.

The Scoreboard

• Semiconductor: Trump mulls tariffs on foreign electronics based on number of chips (Reuters)
• AI: Anthropic to triple international workforce in global AI push (CNBC)