OpenAI to launch AI-centric smartphone in 2028, in talks with Qualcomm, MediaTek for mobile chipset

OpenAI is reportedly laying groundwork to enter smartphones, with talks underway with Qualcomm and MediaTek for chipsets and Luxshare expected to design and produce the device. The company is also still working on an AI wearable and a smart speaker with camera, with the smartphone timeline pointing to specification finalization by late 2026 or Q1 2027, mass production in 2028, and launch potentially later that year. The report is strategic but early-stage, with no confirmed product or revenue impact yet.

Analysis

The near-term market impact is not the consumer-device story; it is the probability of a multi-year design-in cycle for a custom AI stack that forces a revaluation of the silicon supply chain. If OpenAI truly pursues a proprietary, voice-first OS with meaningful on-device inference, the first beneficiaries are not handset brands but component vendors that can tolerate low-volume, high-complexity ramps: baseband, AP, memory, packaging, thermal, and camera subsystems. Qualcomm’s upside is asymmetric because even a partial win on modem/connectivity or edge AI silicon would extend its relevance beyond the mature smartphone replacement cycle into a new category with higher ASPs and stickier software pull-through.

The more important second-order effect is competitive discipline on Apple. OpenAI does not need to ship a mass-market phone on day one to pressure the premium ecosystem; it only needs to redefine the interface expectation around ambient, voice-native computing. That threatens the hardware stack where Apple still monetizes frictionless integration, while also pulling attention from its services narrative if consumers begin to treat the phone as an AI endpoint rather than the center of the experience. The real risk for Apple is not unit displacement in 2028, but a slow compression of upgrade urgency as AI-native interaction migrates to new form factors and weakens the premium justification for incremental handset refreshes.

The biggest contrarian point is timing: this is a credible strategic direction, but the market is likely overestimating how quickly a vertically integrated AI phone can be brought to scale. The dependency chain is long, and any product that must balance latency, battery, privacy, and cloud fallback will be exposed to brutal iteration risk. For now, the right trade is to express the theme through enablers rather than the end-product, because the probability-weighted value accrues to suppliers well before consumer adoption is proven.

Catalyst-wise, watch for supplier announcements and any evidence of silicon tape-out or reference-design selection over the next 12-24 months; those are the signals that move this from narrative to executable pipeline. If those milestones slip, the market will likely fade the story quickly because the option value sits far enough in the future that execution slippage matters more than headline ambition.

Key Decisions for Investors

• Buy QCOM on weakness over the next 1-3 months; use a 6-12 month horizon. Risk/reward favors upside if it wins even one design-in on connectivity/edge compute, with downside limited by existing handset and auto exposure.
• Sell AAPL relative to QCOM via a pair trade (short AAPL / long QCOM) into any AI-hardware enthusiasm over the next 3-6 months. Thesis: the market underprices slow erosion of upgrade urgency more than near-term unit risk.
• Add a small-basket long of AI hardware enablers with QCOM as the core and complementary exposure to high-end component suppliers. Use 12-18 month horizon; expect multiple expansion if design wins become visible.
• Avoid shorting handset OEMs outright. The payoff is too far out and the first-order impact is likely narrative-driven, not revenue-driven, for at least 2-3 years.
• Optionality trade: buy long-dated QCOM calls or call spreads expiring after the first meaningful supplier-confirmation window. Best setup if the market starts pricing a 2028 launch path before execution proof is visible.