The possible agreement that would test chip manufacturing in the US

Donald Trump has claimed that Apple will collaborate with Intel to design and manufacture chips in the United States. The announcement, made through Truth Social and reported by various financial outlets, immediately shook the market because it touches on three sensitive industry points: Apple’s reliance on TSMC, Intel’s need to attract external clients for its foundry business, and Washington’s strategy to rebuild an advanced semiconductor supply chain domestically.

For now, it’s important to view this news with caution. Neither Apple nor Intel has publicly detailed what types of chips would be involved, which manufacturing processes would be used, when wafers would start, or if the agreement affects secondary components or main processors. That nuance is key. Manufacturing controllers, connectivity chips, or auxiliary circuits is very different from producing future Apple Silicon for Macs, iPads, or iPhones.

Still, the move aligns with a clear trend. Apple aims to expand its industrial footprint in the US. Intel needs to demonstrate that Intel Foundry can compete for strategic clients. And the US government wants major tech companies to buy more capacity domestically, even if that doesn’t immediately eliminate dependence on Taiwan, South Korea, or Japan.

The agreement Intel needs to validate its foundry

For Intel, a contract with Apple would be worth much more than direct revenue. Apple is one of the most demanding clients worldwide for semiconductors. Its chips prioritize performance per watt, vertical integration, cost control, volume, reliability, and strict schedules. If Intel succeeds in manufacturing any critical component for Apple, it can use that as a credibility boost with other clients.

That is precisely the weak point Intel has been trying to fix for years. The company doesn’t just want to design its own processors; it aims to manufacture chips for third parties and compete with TSMC and Samsung Foundry. To achieve this, a compelling roadmap isn’t enough. Intel must demonstrate manufacturing execution, strong production performance, support for external customers, and the ability to produce at scale.

Intel approaches this possible partnership at a more favorable moment than a few years ago. Its Intel 18A-P process just entered risk production. The company claims this evolution offers 9% more performance at the same power or 18% less power at the same performance compared to Intel 18A, along with thermal improvements and design rule compatibility. This is significant because 18A-P is designed to facilitate reuse of IP and design flows already prepared for 18A.

However, risk production does not mean mass manufacturing. It’s a validation phase where the process is tested, early designs are produced, and efforts are focused on stabilizing yields. Apple wouldn’t entrust its most critical chips to a foundry unless there’s evidence it can meet volume, quality, and schedule commitments.

Apple wants to diversify, but TSMC remains central

Apple still depends heavily on TSMC for its most advanced chips. That relationship has been a key to Apple Silicon’s success: iPhone, iPad, and Mac have benefited from leading-node manufacturing at TSMC, with efficiencies hard for rivals to match. Changing that base is neither simple nor likely in the short term.

But diversification makes sense. Global demand for advanced chips is driven by AI. NVIDIA, AMD, Apple, Qualcomm, and others are competing for cutting-edge capacity. TSMC continues expanding, including its facility in Arizona, but its supply chain remains concentrated in Asia for many critical steps.

In 2025, Apple announced a commitment to invest $600 billion in the US over four years, alongside an American Manufacturing Program with partners like Corning, GlobalWafers, Applied Materials, Texas Instruments, Samsung, GlobalFoundries, Amkor, Broadcom, and TSMC. Including Intel in this network would be consistent with that strategy—if costs and quality align.

At least initially, the most likely approach would be to begin with less critical chips or specific designs. Apple typically adopts supply changes gradually. Before moving its core devices’ processors, it requires years of validation, performance testing, packaging, security, availability, and per-wafer costs.

Washington re-enters semiconductor supply chain politics

The potential Apple-Intel collaboration cannot be separated from US industrial policy. In 2025, Trump’s administration allocated part of the funds linked to the CHIPS Act and strategic programs into an $8.9 billion investment in Intel common stock, holding nearly 10%. Intel described this agreement as a move to accelerate US leadership in advanced manufacturing.

Soon after, NVIDIA announced a $5 billion investment in Intel and a product alliance to develop multiple generations of data center and PC products. This didn’t make Intel a NVIDIA GPU manufacturer but reinforced the idea that US tech giants have financial, industrial, and political incentives to support Intel.

The political message is clear: the US wants to ensure that the design, manufacturing, and packaging of strategic chips are not reliant on a single region. Rebuilding such a supply chain takes years. Announcing investments isn’t enough; it also requires engineers, suppliers, tools, materials, factories, water, energy, advanced packaging, and willing customers to assume the risks of moving production.

Which node could Intel use for Apple?

A key technical question is which process Intel would use. If the deal is limited to less advanced chips, mature nodes or processes suitable for connectivity, power management, or controllers could suffice. For main processors, the focus shifts to Intel 18A, 18A-P, or future generations.

Intel 18A and 18A-P are significant because they incorporate RibbonFET and PowerVia. RibbonFET is Intel’s implementation of gate-all-around transistors, while PowerVia delivers power from the back of the chip. In theory, this combination boosts efficiency, scalability, and power distribution. In practice, everything depends on manufacturing performance and whether design tools are mature enough for external clients.

Apple doesn’t choose a node based on marketing. It chooses based on real performance, power consumption, costs, yield, volume, packaging, and risk. Intel can offer an attractive narrative, but must deliver in execution. That’s the part that has historically set TSMC apart: not just having technology but delivering it with extreme industrial discipline.

TSMC remains in the equation

An agreement with Intel doesn’t mean Apple will abandon TSMC. That would be an overreach. TSMC remains Apple’s primary partner for advanced chips, and its Arizona fab is already part of the company’s US strategy. Moreover, the most cutting-edge TSMC nodes don’t reach the US at the same pace as Taiwan.

The most reasonable scenario is a more diversified supply chain, not a replacement. Apple could buy chips from TSMC in Arizona, expand deals with GlobalFoundries for specific components, use Amkor for packaging, and explore Intel Foundry for future designs. This diversification reduces geopolitical risk and improves Apple’s bargaining position.

AI demand also favors diversification. While Apple doesn’t compete like NVIDIA for training GPUs, it requires advanced capacity for its own SoCs, private AI servers, device chips, and future products. Having more industrial options is always beneficial for a company selling hundreds of millions of devices annually.

The risk: turning industrial policy into commercial promise

The biggest risk here is mistaking political desire for actual production capability. Trump’s announcement of collaboration doesn’t guarantee Intel will produce Apple’s main chips, nor that the deal will have an immediate impact. Semiconductor movements don’t follow social media or press releases; advanced chips can take years from design to volume production.

There’s also the risk of overhyping the headline. A deal to produce “some chips” might be very relevant for Intel Foundry but doesn’t necessarily change the architecture of iPhones or Macs. Until Apple or Intel provide detailed scope, analysis should remain cautious.

Markets may react quickly, but factories do not. Intel needs clients, but also needs to prove that its advanced nodes can reliably produce external designs. Apple needs diversification but won’t sacrifice performance and efficiency for political symbolism.

A potential turning point for Intel, if confirmed by actions

If Apple were to start using Intel Foundry for relevant chips, the impact would be huge. It would validate Intel’s strategy as a third-party manufacturer, reinforce US industrial policy, and offer a serious alternative to TSMC’s dominance. It would also send a message to other designers: Intel is back as a viable option.

But the path there remains long. The agreement must translate into designs, nodes, wafers, packaging, testing, and final products. Success will be measured, not by political declarations but by chips functioning within Apple devices and meeting the company’s quality standards.

The most cautious view is that Apple and Intel may be taking initial steps towards a more serious foundry relationship. For Intel, this could be an exceptional opportunity. For Apple, an industrial insurance policy. For the US, a political victory in the race to regain advanced manufacturing.

Now, the hard part remains: knowing what will be manufactured, when, and with what results.

Frequently Asked Questions

What did Donald Trump say about Apple and Intel?

Trump claimed that Apple agreed to work with Intel to design and produce chips in the US. Details about the specific chips, nodes, or volumes have not been publicly disclosed.

Has Apple confirmed the agreement?

At the time of initial reports, neither Apple nor Intel provided immediate public details about the scope of the collaboration.

What chips might Intel manufacture for Apple?

It could range from secondary components, controllers, or connectivity chips to more significant processors. Without official confirmation, it’s uncertain whether Intel will produce advanced Apple Silicon.

Why is Intel 18A-P important?

Intel 18A-P is an evolution of Intel 18A that has entered risk production. Intel states it improves performance, power consumption, and thermal characteristics, while maintaining design rule compatibility with 18A.

Will Intel replace TSMC as Apple’s supplier?

Short-term, that seems unlikely. The most reasonable approach is a diversified supply chain, with Intel providing some specific chips, while TSMC continues to lead in advanced nodes in Taiwan.