Qualcomm guides $15B data-center AI revenue by 2029, up from zero three years ago

Qualcomm set a $15 billion annual data-center AI chip revenue target for fiscal 2029 during its June 24 Investor Day, a figure the company disclosed alongside confirmation that meaningful volume shipments have not yet begun. The guidance represents a 200% increase from the $5 billion run-rate the company expects by fiscal 2027, itself an estimate built on products still in early-stage deployment. Three years ago, Qualcomm recorded no data-center revenue. The company now positions itself as a credible alternative to Nvidia in the inference layer, where margins compress but deployment scale multiplies.

The timing matters more than the number. Qualcomm announced the target six months after completing its acquisition of Modular, a compiler startup whose technology stack is designed to translate AI workloads across heterogeneous chip architectures without vendor lock-in. The acquisition cost was undisclosed, but the strategic intent is clear: Qualcomm is building a software moat to compensate for arriving late to a market Nvidia has owned since 2016. The $15 billion figure was constructed to be repeatable in earnings calls and allocator presentations, a benchmark that forces Wall Street to model Qualcomm as something other than a smartphone chipmaker in structural decline. The company's mobile business, which still accounts for roughly 70% of revenue, grew 3% year-over-year in the most recent quarter, a pace insufficient to justify the multiple expansion management seeks.

The market Qualcomm is entering is not the one Nvidia dominates. Training workloads, where gross margins exceed 70%, remain concentrated in H100 and next-generation Blackwell deployments. Inference, where Qualcomm is targeting volume, operates at margins closer to 40% and requires a different commercial model: lower unit economics, tighter integration with hyperscale infrastructure, and price sensitivity that rewards second-source suppliers. Amazon, Microsoft, and Google have each signaled interest in diversifying inference silicon away from single-vendor dependence, a tailwind Qualcomm is designed to capture. The company's existing relationships with Android OEMs provide no structural advantage here; data-center procurement operates on different timelines, requires different validation cycles, and responds to different risk models. Qualcomm is building this business from scratch, and the $15 billion target assumes execution on product roadmaps that have not yet been tested at hyperscale.

Operators and allocators should monitor three events over the next eighteen months. First, volume shipment announcements from at least one Tier 1 hyperscaler, expected in Q4 2026 or Q1 2027, will validate whether Qualcomm's inference silicon meets deployment thresholds. Second, gross margin disclosure on data-center revenue, likely segmented out by fiscal Q2 2027, will clarify whether the business model sustains returns above Qualcomm's corporate average or dilutes them. Third, competitive response from Nvidia, which has already begun pricing Hopper-generation chips more aggressively for inference workloads, will determine whether the market Qualcomm is targeting remains structurally profitable or compresses into a low-margin commodity game. Each of these milestones carries binary implications for whether the $15 billion figure represents credible guidance or aspirational marketing.

The $15 billion target is not a forecast. It is a financing instrument, a number designed to anchor capital allocation decisions before the products that generate the revenue exist at commercial scale. Qualcomm needs this narrative to work, because the alternative is a company whose core business is shrinking and whose valuation multiple depends on a story it has not yet earned the right to tell.