Qualcomm +75%: The Chip Company That Became the Secret Infrastructure of Every Major AI Device

On May 22, 2026, Qualcomm shares jumped 12% in a single session — hitting a record high after news broke that Stellantis had expanded its Snapdragon partnership across its entire vehicle lineup. But the single-day pop was a footnote. The real story: Qualcomm stock had already surged 75% over the prior month as investors finally started connecting the dots on what the company actually is — the unchosen, invisible backbone of nearly every AI device anyone’s building.

What Triggered the Pop

The immediate catalyst was a multi-year deal with Stellantis, announced May 21. Under the expanded agreement:

• Snapdragon Digital Chassis SoCs will power cockpit, connectivity, and ADAS across Stellantis’ global vehicle portfolio
• Snapdragon Ride Pilot — Qualcomm’s Level 2+ hands-free automated driving platform — is included
• A non-binding letter of intent was signed for aiMotive, Stellantis’ automated driving and simulation subsidiary, to join Qualcomm Technologies

That acquisition, if completed, would bring one of Europe’s more sophisticated simulation stacks directly inside Qualcomm, strengthening its push into full-stack autonomous driving rather than just chipsets.

The deal pushed Qualcomm’s automotive segment past $5 billion in annualized revenue — a record — and reinforced a pattern investors hadn’t fully priced in: Qualcomm’s auto business isn’t a hedge. It’s becoming the point.

The Earnings Were Already Strong

Before the Stellantis news, Qualcomm had already posted a clean Q2 FY2026 beat on April 29. Revenue came in at $10.6 billion, EPS at $2.65, both above consensus. The segment numbers:

• Automotive: +38% year-over-year to a record $1.33 billion
• IoT: +9% to $1.73 billion
• Handsets (QCT): Stable, with edge AI features driving premium tier demand

Gross margin: 55.1%. The company flagged it expects meaningful acceleration in the second half of 2026 as smart glasses launches begin shipping — a category that has been building quietly and is now expected to reach inflection.

The OpenAI Angle

On April 27, Qualcomm shares had already jumped roughly 7% on a separate headline: OpenAI was partnering with Qualcomm and MediaTek to develop the chip for its AI-native smartphone. That device — reported to be designed with Jony Ive’s io company and manufactured by Luxshare — replaces the traditional app grid with an AI agent layer. Mass production is currently targeting 2028.

Qualcomm being selected as a silicon partner (alongside MediaTek as primary chip supplier) confirmed what its CEO had been hinting at for months.

The CEO’s Bigger Claim

In a series of Fortune interviews in early May, CEO Cristiano Amon was unusually candid about the scope of Qualcomm’s device pipeline. When asked which AI companies Qualcomm is working with on next-generation hardware, his answer:

“Pretty much all of them."

He acknowledged “secret form factors I cannot tell you about” — wearable devices that he described not as smartphones or smartwatches, but as things you wear: glasses, jewelry, pins, pendants. The design logic is what he calls the “ecosystem of you”: glasses with outward-facing cameras, earbuds that capture ambient audio, and an autonomous AI agent tying the inputs together.

Amon’s stated thesis: 2026 is the year of AI agents. By 2028, he expects meaningful workloads to shift from smartphones to these new AI-first form factors. He frames Qualcomm’s position as the hardware neutral ground — working with Meta, OpenAI, and others across this transition, without being tied to any single platform.

Edge AI vs. Cloud AI: The Structural Argument

Why does running AI on-device matter? The cloud-vs-edge tradeoff has three dimensions:

Latency: A cloud request introduces round-trip time that on-device inference eliminates. For glasses that need to identify what you’re looking at in real time, or earbuds that need to translate speech before you finish a sentence, network delay is unacceptable.

Privacy: Data processed locally never leaves the device. For personal health data, ambient audio, or financial queries, this is a genuine differentiator — not marketing. Several enterprise customers, particularly in healthcare and legal, will not use cloud-dependent AI for specific workflows.

Cost and connectivity independence: Every cloud AI call has a per-token cost. At scale — billions of smartphones, millions of vehicles, wearables worn 16 hours a day — on-device inference eliminates an operational expense that would otherwise compound continuously. Off-grid use cases (automotive, rural, aviation) also require local inference.

The Snapdragon X2 Plus (announced at CES 2026) brings 85 TOPS of on-device compute to mainstream AI PCs — not just the premium tier. The Snapdragon Wear Elite extends the architecture to watches, pins, and pendants. Each category runs models locally that would have required a data center call two years ago.

The Apple Modem Loss and the Bull Case Anyway

Qualcomm has faced one significant headwind that analysts cite regularly: Apple completed its move to in-house cellular modems with the iPhone 17 line, removing Qualcomm from what had been a substantial revenue stream.

The stock’s 75% surge in a month suggests the market has repriced around the idea that the automotive + AI device expansion more than compensates. At $214 on May 22 — market cap approximately $225 billion, trailing P/E around 22, dividend yield 1.8% — the valuation isn’t cheap, but it isn’t pricing in a dominance scenario either.

Analysts are divided on how high this goes. Tigress Financial Partners (Ivan Feinseth) has a $280 price target with a buy rating; Daiwa sits at $225; other houses remain at hold, citing the Apple transition’s lagging revenue effects and potential MediaTek competition in the mid-tier smartphone market.

What Qualcomm Is and Isn’t

Qualcomm is not an AI lab. It doesn’t train frontier models, doesn’t have a cloud platform, and isn’t trying to be. What it has built — over decades — is a deep stack of power-efficient silicon, radio hardware, and driver software that runs on a significant fraction of the world’s connected devices.

That stack now has a new job: running AI models locally, at scale, without a data center. The companies building the most-discussed AI devices of 2028 and beyond — OpenAI’s AI-native smartphone, Meta’s Ray-Ban successors, the unnamed secret devices Amon is building with partners he won’t identify — are converging on Qualcomm’s hardware because no one else has both the NPU performance and the power envelope required for always-on wearable AI.

Nvidia dominates the data center. Apple dominates its own ecosystem. Qualcomm is quietly becoming the infrastructure layer for everything in between — the edge AI devices that will be the surface area through which most people experience AI agents in daily life.

The market is starting to notice.

Qualcomm (NASDAQ: QCOM) closed at $214 on May 22, 2026. This article is based on publicly available earnings reports, press releases, and news coverage. ChatForest does not hold any position in securities discussed. Nothing here is investment advice.