Cristiano Amon, CEO of Qualcomm, walked into the company's Investor Day on June 24, 2026, and laid out a number built for headlines: over $15 billion in annual data center AI chip sales by fiscal 2029, climbing from $5 billion in 2027. A broader pitch followed - a $40 billion non-handset revenue target and adjusted earnings above $18 per share that same year. Qualcomm shares jumped as much as 15% after hours, then surrendered nearly all of it. The stock closed June 25 at $204.90, down 7.66% from the prior week. The reason was simple: the chips needed to reach that $15 billion do not ship yet, and investors are being asked to fund a roadmap, not a product line.
Qualcomm's data center CPUs are scheduled for production in 2028, with a custom AI inference processor - the Dragonfly AI300 - teased for later. The company also announced a $3.92 billion stock purchase of AI software firm Modular and a multiyear CPU supply agreement with Meta Platforms. Yet the Q2 fiscal 2026 filing described hyperscaler engagement only as "on track for initial shipments later this calendar year." That phrasing has not been upgraded to "shipping."
The math behind the promise
Qualcomm's current business offers a sharp contrast. Total Q2 FY26 revenue was $10.599 billion, down 3.46% from the prior year. Handsets still produced $6.024 billion of that but fell 13%, pressured by memory supply constraints and soft demand from Chinese OEMs. Data center revenue from the recently closed Alphawave Semi acquisition is not yet separately reported. So Amon is telling shareholders that a segment generating essentially nothing in fiscal 2026 will produce a sum equal to about 34% of the company's entire fiscal 2025 revenue base three years later. For sales professionals, the gap between ambition and auditable revenue is a familiar tension - this is a large deal still in the proposal stage.
Amon's projections require more than silicon. The CEO said Modular's software stack and the Dragonfly architecture will deliver high efficiency on a "dollar per token" and "dollar per kilowatt" basis. Those efficiency arguments only win when power costs, not chip performance, become the binding constraint in data centers. That point has not yet arrived for most buyers.
Nvidia's moat is the variable Qualcomm cannot ignore
AI for Executives & Strategy is exactly what Amon was selling - a future-state narrative built on market share gains. Yet Bloomberg Tech's Ed Ludlow framed the competitive reality bluntly. "NVIDIA absolutely dominates the market and whatever's left, AMD kind of comes up and gets the rest," he said, characterizing NVIDIA's position as a technical monopoly.
NVIDIA reported $75.246 billion in data center revenue in a single quarter during Q1 FY27, growing 92% year over year. Qualcomm's entire 2029 data center target would equal one quarter at Jensen Huang's current run rate. And while AMD is cementing the second-place slot with its Meta-anchored MI450 deployment, both NVIDIA and AMD ship new architectures annually. Qualcomm's parts remain years away. Ludlow acknowledged that Qualcomm's designs could be "highly performing, very efficient chips on a dollar per token basis," but in sales terms, that is a champion-challenger narrative - not a disruption story.
Micron's quarter is the warning embedded in the same news cycle
On the afternoon Amon pitched the future, Micron Technology reported its fiscal Q3 results. Revenue reached $41.46 billion, beating the $35.25 billion consensus by 17.6%. Non-GAAP EPS came in at $25.11 versus $20.28 expected, and GAAP gross margin hit 84.6%, up from 37.7% a year earlier. CEO Sanjay Mehrotra called the result "the strategic value of memory in the AI era" and highlighted newly signed multiyear strategic customer agreements that lock in HBM4 demand. Micron didn't sell a vision - it sold product that ships now, with contracts that show committed buyers.
The contrast matters for anyone carrying a quota. Qualcomm's future depends on a combination of design wins, software stack maturity, and customer adoption that hasn't started. The next four quarters will test whether the unnamed hyperscaler customer gets named, whether Meta's CPU order converts into a deposit, and whether Modular's benchmarks materialize.
Why this matters for Sales
Qualcomm's pitch is an exercise in selling a roadmap, not a product. The numbers are large enough to move a stock, but the revenue isn't real yet. Sales professionals who face long deal cycles know that a verbal commitment doesn't equal a forecast, and a design win doesn't equal a shipped part. When your own pipeline faces similar scrutiny - buy-in without a deposit, a customer referenced only as a "hyperscaler" - this is the same gap between promise and proof. AI for Sales training increasingly means learning to distinguish between supplier excitement and verified demand. The Micron contrast drives that home: inventory that ships, gross margins that print, and contracts that name real names always outweigh a $15 billion number three years out.
Your membership also unlocks:
