Apple gains tighter control over power, performance, and 5G integration
In the long arc of technological self-determination, Apple is preparing to sever one of its last significant dependencies on an outside supplier — the modem, the invisible engine that connects every iPhone to the world's wireless networks. Beginning in 2023, the company plans to manufacture its own 5G modem chips through TSMC's most advanced fabrication processes, a move that traces back to its quiet acquisition of Intel's modem division in 2019 and signals a deeper ambition: to own not just the experience of its devices, but the fundamental physics of how they communicate. For Qualcomm, long the unchallenged sovereign of smartphone modem supply, the reckoning is already being counted in projected market share.
- Apple is on the verge of eliminating its last major chip dependency, designing a custom 5G modem that would replace Qualcomm components in iPhones as early as 2023.
- Qualcomm has openly acknowledged the threat, forecasting its iPhone modem share will collapse from near-total dominance to just 20 percent within two years — a stunning concession from a company that once seemed irreplaceable.
- The transition is being engineered in stages, with Apple testing designs at 5nm before committing to TSMC's cutting-edge 4nm process, a node so new no commercial product has yet shipped using it.
- TSMC's deep operational entanglement with Apple — including hundreds of engineers embedded in Cupertino — makes it the only credible manufacturing partner for a component this complex and performance-critical.
- The shift lands Apple in a position of rare vertical integration, giving it direct control over power efficiency, 5G performance, and the negotiating leverage that comes from no longer needing anyone else's modem.
Apple is preparing to build its own 5G modems for iPhones starting in 2023, manufactured by TSMC using its 4-nanometer process — a fabrication technology so advanced that no commercial product has yet shipped with it. Before committing to that node, Apple is designing and testing the chips at the 5-nanometer scale, a process already familiar to its chip teams.
The move has been anticipated for years. Qualcomm, which currently supplies modems for every iPhone 13 model, recently projected that its share of Apple's modem orders will fall to just 20 percent within two years — a dramatic acknowledgment of what's coming from a company that has long dominated this corner of the smartphone industry.
Apple's path here began in 2019 with the acquisition of Intel's 5G modem division, a purchase that made its intentions unmistakable. That same year, a bruising patent war with Qualcomm ended in settlement, freeing Apple to develop its own technology without legal constraint. The peace was useful — but always temporary.
TSMC is the natural partner for this effort. The foundry already manufactures Apple's A-series and M1 chips, and maintains hundreds of engineers stationed directly in Cupertino. That embedded relationship makes TSMC the only realistic choice for something as performance-critical as a modem.
The deeper significance of this transition lies beyond supply chain arithmetic. By controlling its own modem, Apple gains tighter command over power consumption, 5G integration, and its relationships with carriers and regulators. For Qualcomm, losing Apple's modem business — even partially — marks a structural shift in how the smartphone industry is organized, and a reminder that no supplier relationship, however entrenched, is permanent.
Apple is preparing to build its own 5G modems for iPhones starting in 2023, according to reporting from Nikkei, marking another significant step in the company's push to control the silicon at the heart of its devices. The chips will be manufactured by Taiwan Semiconductor Manufacturing Company using its cutting-edge 4-nanometer process—a technology so new that no commercial product has yet shipped with it. Before that leap, Apple is designing and testing the modems at the slightly larger 5-nanometer scale, a familiar territory for the company's chip teams.
This move has been widely anticipated in the industry. Qualcomm, which currently supplies modem components for every iPhone 13 model sold, recently acknowledged the writing on the wall by projecting that its share of iPhone modem orders will plummet to just 20 percent within two years. For a company that has dominated smartphone modem manufacturing, the shift represents a dramatic loss of business from Apple alone.
Apple's path to this moment began in 2019, when the company acquired Intel's 5G modem division—a strategic purchase that signaled clear intent to eventually abandon its reliance on external suppliers for this critical component. That same year, Apple and Qualcomm ended a bitter patent dispute that had consumed years of litigation and cost Qualcomm more than $4 billion in settlement payments. The peace deal allowed both companies to move forward, but it also gave Apple the breathing room to develop its own technology without the legal entanglement that had previously constrained its options.
TSMC is the natural manufacturing partner for this effort. The Taiwanese foundry already produces all of Apple's A-series processors—the chips that power iPhones—as well as the M1 systems-on-chip used in Mac computers. The relationship runs deep: hundreds of TSMC engineers are stationed in Cupertino, working directly alongside Apple's chip development teams. This embedded presence makes TSMC the only realistic choice for manufacturing a component as complex and performance-critical as a smartphone modem.
The timing aligns with Apple's broader chip roadmap. The 2022 iPhone models are expected to use TSMC's 4-nanometer process for their main processors, while certain iPad models will reportedly jump to 3-nanometer chips in 2023. Apple is said to be planning an even faster transition for iPhones, potentially moving to 3-nanometer production as soon as next year—a pace that underscores how aggressively the company is pushing the boundaries of semiconductor manufacturing.
What makes this transition significant extends beyond the supply chain. By designing and manufacturing its own modems, Apple gains tighter control over power consumption, performance, and the integration of 5G capabilities with the rest of the phone's silicon. It also reduces dependency on a single external supplier and strengthens Apple's negotiating position with carriers and regulators. For Qualcomm, the loss of Apple's modem business—even if it retains some market share—represents a fundamental shift in the smartphone industry's structure.
Notable Quotes
Qualcomm expects to account for just 20 percent of iPhone modem orders in two years' time— Qualcomm (via Nikkei reporting)
The Hearth Conversation Another angle on the story
Why does it matter that Apple is making its own modem instead of buying from Qualcomm?
Control. A modem is the radio that connects your phone to the network. If Apple designs it, they can optimize it specifically for their hardware, their software, their power management. They're not constrained by what Qualcomm thinks is good enough.
But Qualcomm has been doing this for years. Why is Apple suddenly capable of doing it better?
They're not necessarily better yet. But they have the resources, the engineering talent, and now the foundational technology from Intel's acquisition. They're also willing to invest years and billions to get it right. For Qualcomm, it's one product line among many. For Apple, it's existential.
The article mentions TSMC has hundreds of engineers in Cupertino. That seems like a lot.
It is. That's not typical. It means Apple and TSMC have built something closer to a joint venture than a vendor relationship. TSMC isn't just manufacturing what Apple designs—they're embedded in the design process itself.
What happens to Qualcomm?
They shrink. They go from supplying every iPhone to supplying maybe one in five. They'll survive—there are other phone makers—but Apple was their crown jewel. Losing it is a wound that doesn't heal quickly.
Is 2023 a hard deadline or could it slip?
In Apple's world, deadlines slip all the time. But the momentum is real. They've been planning this for years. The only question is whether the first generation works well enough to ship, not whether they'll try.