Washington recently made news about Apple’s possible collaboration with Intel on the manufacture of chips, causing the technology world to go haywire with speculation. The concept of necessity and ambition is simple on the surface, but the reality of this huge paradigm shift in the semiconductor industry is far more intricate and complex than many industry observers had imagined. The deal, if it ever becomes official, is an intriguing mix of corporate planning, national industrial policy and the drive to keep developing new technologies in a fiercely competitive world market.
Essentially, this potential deal would represent a combination of Intel’s resolute attempt to regain the trust of its customers as a contract chipmaker and Apple’s urgent need for greater manufacturing capacity. The world’s main supplier of Apple components, Taiwan Semiconductor Manufacturing Company, has been grappling with forces it cannot control as it seeks to ramp up production of AI chips to satisfy growing demand from Nvidia and other leading players in the AI arena. The chip shortage problems at TSMC have already had a real impact on the company, and in April, Apple’s CEO Tim Cook said that the chip shortage has been impacting its iPhone sales. This admission highlighted an exposure that can be very dangerous when relying on a single manufacturing partner, no matter how strong that partner is.
But the road to production is littered with big hurdles that cannot be ignored or downplayed. The worst case scenario, said chip research firm Future Horizons’ CEO Malcolm Penn, would be two to three years before the first chips rolled off the production line. The design effort for such a system on chip can take about 2 years just for the design phase, and a further 4 months of production cycle time to reach volume ramp up. This timeline assumes Intel’s technology is fully developed and running, and the design tools are sufficiently reliable to be relied upon by Apple. In fact, Penn described the possible deal as a shotgun wedding – Intel has had problems with chip yields and production schedules in the past, and, without history, the commercial and financial risks are quite a leap of faith.
It’s especially interesting to note the timing of this possible partnership, as Intel has recently made significant strides in the semiconductor manufacturing industry. Intel is in the midst of a “latecomer’s success” in the early days of the AI boom. In April, the company became a customer of Tesla, a major endorsement of its foundry goals. It has helped Intel on the road to what may be an even more pivotal relationship with Apple, but the future is still up in the air.
It remains unclear what Intel will use for its chips, and experts have differing opinions on which manufacturing process Apple will select. Some watchers believe Apple is following in the footsteps of Tesla and moving to Intel’s next-generation 14A process, which is still years away from production, but is based on the most sophisticated chip manufacturing equipment in the world. Analyst Bob O’Donnell at TECHnalysis Research noted Apple likely would want to use Intel’s 14A process technology, which may be available in 2028 or 2029. If this is indeed the case, O’Donnell noted, it’s an extremely significant development for the foundry business, and for U.S.-based semiconductor manufacturing in general.
Other analysts, however, are more conservative, predicting that Apple will compromise on the cutting edge for reliability. If this is the case, Apple could choose to go with 18A-P, a version of Intel’s most advanced process which is now in initial production this month, or an older and more established node such as Intel 3. The options of bleeding edge technology versus proven reliability is just part of the equation that Apple will have to juggle when deciding on this deal.
But Daniel Newman, CEO of tech research firm Futurum Group, painted a more nuanced picture, citing a late-2027 or early-2028 start to volume production of Apple-designed chips. Newman surmised that the first will be less critical elements, possibly those found on the MacBook Air or certain versions of the iPad Pro. This way, Apple can experiment with Intel’s technology on less important products before employing its most important chips in the Intel process.
This approach of dissing Intel with low-end products and then committing to the full range of products is indicative of the cautious way analysts expect Apple to play the game. Intel has had big problems in the past with delivering chips on time and of the quality that Apple requires – and it will have to find a way to do that with the company’s chips in the future, especially when it comes to yield, the number of chips which were not already defective that it must deliver to Apple, the world’s largest consumer electronics company. The percentage of the silicon wafer that is ready to use when manufacturing is finished is an important measure known as yield, and it is directly related to profitability and availability of product.
The logic of such an eventual agreement is more than just a corporate one: it also reflects national policy concerns in terms of industry. Intel has become a major component of the United States’ chipmaking capacity plan, which is based on tariffs and incentives. The ten per cent stake in this operation, plus a massive five billion dollar fund from Nvidia, at the urging of President Donald Trump, highlights the political and economic value of putting semiconductor manufacturing back on the U.S. soil.
But the challenges to this potential partnership are not to be taken lightly. The semiconductor industry has a very long development cycle, high capital investment, and constant technological progress that cannot be complacent with. Intel’s manufacturing woes have been well-documented in recent years and have experienced major setbacks which have hurt the company’s credibility in the eyes of prospective customers. It’s going to take a lot of work and proof to rebuild that trust and establish our ability to meet Apple’s high expectations.
