Qualcomm now knows exactly how fast it is losing Apple. On July 29, CEO Cristiano Amon disclosed in a post-earnings Reuters interview that Apple-related chip revenue will fall roughly 50 percent sequentially between the September and December 2026 quarters — a sharper drop than the company had modeled, arriving sooner than its own guidance suggested. The catalyst is not a change of strategy but a change of pace: supply constraints across the semiconductor industry have accelerated Apple's rollout of its in-house C2 modem, compressing the timeline in ways that Qualcomm's models did not anticipate.

That acceleration has a concrete financial signature. Qualcomm had previously estimated it would supply modems for roughly 20 percent of the iPhone 18 lineup — itself already a dramatic reduction from its historical near-total dominance of Apple handsets. That figure is now off the table. CFO Akash Palkhiwala told analysts on the earnings call that the company's actual share in the upcoming iPhone launch will be "materially lower" than the 20 percent estimate. In plain terms, Qualcomm's last foothold in Apple's flagship line is shrinking faster than even the chipmaker expected.

"We kind of replaced Apple with the data center," Amon said in a Reuters interview — a line that captures both the scale of what is being lost and the confidence with which Qualcomm believes it can absorb it.

Qualcomm's Q3 Numbers: Strong Quarter, Steeper Cliff Ahead

The quarterly results themselves were solid. Qualcomm reported fiscal third-quarter revenue of $9.9 billion, landing at the high end of its own guidance and beating Wall Street's consensus estimate of $9.67 billion. Non-GAAP earnings per share came in at $2.21, broadly in line with analyst expectations. QCT — the chipmaking segment — contributed $8.5 billion, while the licensing arm QTL delivered $1.3 billion at a 69 percent earnings-before-tax margin.

But within those numbers, the handset division told a harder story. QCT handset revenue fell to $5.09 billion in the quarter, down 20 percent year over year, weighed down by unprecedented memory price inflation that pushed original equipment manufacturers toward lower-tier chips and older inventory. QCT gross margins fell below Qualcomm's historical 48-to-50-percent baseline range as a result.

For the fiscal fourth quarter — the period covering the iPhone 18 launch window — Qualcomm guided to total revenue of $9.7 billion to $10.5 billion, with QCT handset revenue of approximately $5.2 billion. Android growth is expected to partially offset the Apple shortfall. But the more consequential guidance was structural: CFO Palkhiwala confirmed that the December quarter will no longer be Qualcomm's annual revenue peak. Revenue is now expected to grow from December into the March quarter — a direct inversion of the seasonal pattern that had governed Qualcomm's financial calendar for years, driven entirely by the Apple iPhone launch cadence that is now exiting the model.

A Four-Year March Toward Modem Independence

The roots of this moment stretch back to 2019, when Apple acquired Intel's smartphone modem division and began building its own baseband engineering teams and intellectual property portfolio. The first tangible result arrived in early 2025 with the iPhone 16e, which debuted the C1 — Apple's inaugural in-house 5G modem. Within six months, the upgraded C1X appeared in the iPhone Air, and by March 2026 the C1X had also landed in the iPhone 17e, cementing Apple's value-tier lineup on its own silicon.

Throughout the iPhone 17 generation, however, the flagship Pro models retained Qualcomm's Snapdragon modems — and for a technically specific reason. Apple's C-series chips have not yet achieved competitive mmWave 5G performance, the ultra-high-frequency band (24-to-100 GHz spectrum) that US carriers including Verizon and AT&T have invested billions to deploy in dense urban environments. mmWave delivers multi-gigabit throughput but has limited range and cannot penetrate buildings — it works in stadiums, transit hubs, and dense downtown corridors, and US buyers of premium iPhones use it there. Without matching Qualcomm's mmWave capability, Apple cannot universally deploy its own modem in the US Pro line without delivering a functionally inferior network experience to the buyers paying the highest price.

The next-generation C2 modem — internally codenamed Ganymede, manufactured on TSMC's 4-nanometer N4 process — is designed to change that calculus, but not yet for US buyers. Leaked supply chain intelligence, corroborated by multiple analysts, suggests that international iPhone 18 Pro models will deploy the C2 while US models will retain a Qualcomm modem to preserve mmWave performance. The dual-modem approach — Apple silicon for the world, Qualcomm for the American market — means Qualcomm retains some presence in the flagship line. The question the Q3 earnings call answered is how small that presence will be: materially below 20 percent, and falling.

What the C2 brings that Qualcomm's Snapdragon does not offer is NR-NTN support — New Radio Non-Terrestrial Networks, a 3GPP Release 17 standard that enables a smartphone to connect directly to low-earth orbit satellites for broadband internet access, treating them like a very distant cell tower. This goes well beyond the emergency satellite SOS functions on current iPhones. International buyers of the iPhone 18 Pro will receive a modem that can access satellite broadband from LEO constellations — a capability no Qualcomm modem in current production matches.

The Long Game: Why This Transition Is Irreversible

The standard frame for the Apple-Qualcomm story is financial: Apple saves the Qualcomm royalty per unit and improves its margins; Qualcomm loses a key customer. That frame is accurate but incomplete. The deeper implication of Apple's modem vertical integration is architectural.

Qualcomm's Snapdragon modems are discrete chips that communicate with Apple's A-series application processor over a bus interface — a design that imposes a hard ceiling on how efficiently power can be managed across the system, because two separately designed chips must negotiate how they share thermal headroom and battery draw. Once Apple integrates the C-series baseband with the A-series SoC on a unified die or package — as it has already done with Wi-Fi, Bluetooth, and the Ultra Wideband chip — it will be able to optimize power draw and thermal output across the full compute stack in ways that a third-party modem supplier relationship structurally cannot replicate. The C1 and C2 are not yet integrated with the A-series SoC. They are the first steps toward that integration. When it arrives, no price concession or technical specification will make a return to an external modem commercially rational for Apple.

The royalty licensing relationship provides a financial bridge in the near term. Qualcomm's patent licensing agreement with Apple — the successor to the 2019 settlement that ended a years-long legal dispute — runs through at least March 2027, continuing to generate licensing revenue from Apple even as chip revenue declines. That bridge softens the cliff but does not flatten it.

Qualcomm Raises Prices as Margins Come Under Pressure

Alongside the Apple disclosure, Amon announced that Qualcomm will raise prices across its entire chip portfolio beginning September 1. He framed the move not as a margin expansion but as a cost pass-through. "Cost went up, prices are going to go up," Amon said. Palkhiwala estimated that QCT Android revenue is running roughly 20 percent below prior-year levels, with the memory bill-of-materials inflation imposing an EPS headwind of more than $1.50.

The memory crisis is structural, not cyclical. The same AI infrastructure build-out driving demand for high-bandwidth memory in data centers is permanently reallocating DRAM production away from mobile devices, raising costs for every smartphone maker simultaneously. Qualcomm's price increases will layer in over several quarters as customer contracts roll off — terms are negotiated individually, Amon said.

Palkhiwala acknowledged that QCT gross margins will remain below the historical 48-to-50-percent range for the near term, with an additional 1.5-to-2-percentage-point drag expected from the initial wave of custom data center silicon, which carries margins significantly below the corporate baseline.

Automotive: The Near-Term Replacement That Actually Works

While data center revenue dominates the strategic narrative, it is automotive that is doing the actual near-term work of replacing Apple's contribution. QCT automotive revenue reached a record $1.59 billion in fiscal Q3 2026, up 61 percent year over year, extending a multi-year streak of consecutive quarters of double-digit annual growth.

On the strength of that performance, Qualcomm raised its automotive annualized revenue run-rate target exiting fiscal 2026 to approximately $7 billion — up from a prior guidance of $6 billion. A new multi-generation chip supply agreement with BMW for digital cockpit systems, and an expanded collaboration with Stellantis extending into the 2030s, anchored the revision.

The fifth-generation Snapdragon Digital Chassis, ramping in September 2026, is expected to bring an order-of-magnitude increase in processing power and silicon content per vehicle — driving per-vehicle revenue above prior expectations as automakers migrate from socket-by-socket design awards to multi-generation strategic engagements. CEO Amon attributed the momentum to exactly that shift: "Customers are shifting from socket-by-socket design awards to multi-generation strategic engagements."

Data Center: Revenue Begins in December

Qualcomm's data center ambitions are real and funded — the company spent the first half of 2026 confirming the $3.92 billion all-stock acquisition of AI software startup Modular, whose Mojo programming language and MAX inference engine allow developers to run AI models across chips from multiple vendors without rewriting their code. Meta Platforms has signed a multi-generation agreement to deploy Qualcomm's Dragonfly C1000 server CPUs.

But the revenue timeline is important to hold precisely. Two unnamed hyperscale custom silicon customers already have purchase orders in hand and wafer production underway — revenue from those engagements is expected to begin in the December 2026 quarter. That is near, but it is not now. Qualcomm's stated data center revenue target is $5 billion by fiscal 2027, rising to $15 billion by fiscal 2029.

The more technically specific development from the Q3 call was the completion of the tape-out for HBC Gen 1 — Qualcomm's High Bandwidth Compute chip, which addresses a core AI inference bottleneck by integrating compute directly with high-density memory on a single package, eliminating the external DRAM accesses that currently dominate inference power consumption and cost. Silicon demonstrations are expected in coming quarters, with first solutions arriving in mid-2027. The HBC Gen 1 is the product that Amon says potential new customers need to see physically before committing to custom silicon engagements — it is the near-term pipeline's gating milestone.

CFO: Non-Handset Revenue Growth Tops 60% in Fiscal 2027

The strategic case that Qualcomm can absorb the Apple departure rests on a single mathematical claim: combined non-handset revenue growth — spanning automotive, IoT, data center, and adjacent segments — is forecast to exceed 60 percent in fiscal 2027, fully replacing Apple product revenue lost in fiscal 2026. By fiscal 2029, Qualcomm now targets $40 billion in total non-handset revenue, up from a prior $22 billion target, with non-handset chips projected to account for roughly two-thirds of all QCT chip revenue.

The China exposure remains a structural risk. Chinese customers accounted for 46 percent of Qualcomm's total revenues in its most recent fiscal year, at a time of sustained US-China trade tension. Palkhiwala said China handset revenues hit their cyclical bottom in the June quarter, with double-digit sequential growth forecast for September as channel inventory normalized — but that recovery is happening inside a market Qualcomm's CEO acknowledged is down low-teens in fiscal 2026 versus 2025.

Amon's framing at the close of the earnings call was deliberately optimistic. "We are incredibly excited about the next chapter of Qualcomm," he told analysts. "Our relevance in the next phase of AI and distributed intelligence from edge to cloud remains firmly focused on execution." The iPhone 18 lineup is expected to be unveiled at Apple's traditional September event. It will be the first generation in which Apple's own modem silicon reaches the flagship Pro tier — and the one in which Qualcomm's presence in that lineup falls to its lowest point in the company's history as an Apple supplier.

Frequently Asked Questions

Will the US version of the iPhone 18 Pro use a Qualcomm or Apple modem?

Based on supply chain intelligence corroborated by multiple analysts, US models of the iPhone 18 Pro are expected to retain a Qualcomm Snapdragon modem to maintain mmWave 5G performance — the high-frequency band US carriers have deployed in dense urban environments. International models are expected to use Apple's C2 modem. Apple's C2 does not yet match Qualcomm's mmWave performance, and Apple's licensing agreement with Qualcomm runs through at least March 2027. Apple has not officially confirmed modem configurations for the iPhone 18 Pro.

Why is Apple able to make its own modems now, when it couldn't for so many years?

Apple acquired Intel's smartphone modem division in 2019 and spent six years building internal baseband expertise and intellectual property before shipping its first in-house 5G modem (the C1) in the iPhone 16e in early 2025. The delay was primarily a performance problem: Apple's early prototypes reportedly could not match Qualcomm's power efficiency or mmWave throughput. The C2, built on TSMC's 4-nanometer process, closes enough of that gap to deploy in most global markets — but the US mmWave gap persists for this generation.

How is Qualcomm planning to replace the revenue Apple's modem business represented?

Qualcomm is running a three-track replacement strategy. In the near term, automotive revenue — at a record $1.59 billion in fiscal Q3 2026 — is already absorbing much of the impact, with a multi-year BMW digital cockpit agreement and a Stellantis partnership extending into the 2030s. In the medium term, two unnamed hyperscale data center customers are expected to begin generating revenue for Qualcomm's custom silicon in the December 2026 quarter. Long term, Qualcomm's targets call for $15 billion in annual data center revenue by fiscal 2029 and $40 billion in total non-handset revenue — nearly double its prior $22 billion target.

What does Apple's modem independence actually mean for the iPhone's technical capabilities?

In the near term, it means the iPhone 18 Pro gains NR-NTN support in international markets — the ability to connect directly to low-earth orbit satellites for broadband internet, going well beyond the emergency satellite SOS features on current iPhones. Longer term, the strategic significance is architectural: once Apple integrates its C-series modem with the A-series application processor on a single chip package, it will be able to optimize power draw and thermal management across the full computing stack in ways that a separate third-party modem cannot match. That integration has not yet happened, but the C1 and C2 are the steps that make it possible.