OpenAI has enlisted chipmakers Qualcomm and MediaTek to develop custom processors for its own AI-powered smartphone, with mass production targeted as early as 2027, according to multiple industry reports that surfaced in recent weeks. If the project reaches consumers, it would represent OpenAI’s most ambitious push beyond software and into the hardware business that Apple and Samsung have dominated for over a decade.
The effort comes roughly a year after OpenAI acquired Jony Ive’s hardware design firm io in a deal valued at approximately $6.5 billion, a transaction that signaled the company’s serious interest in building physical products. Now, with two of the world’s largest mobile chip suppliers on board, that interest appears to be taking a concrete shape: a phone built from the ground up around on-device AI.
What the chip partnerships reveal
The dual-chipmaker arrangement is the most firmly established detail so far. Qualcomm, which already powers a large share of the world’s premium Android phones, brings mobile processor expertise and critical modem technology. MediaTek, the world’s largest smartphone chip supplier by volume, offers cost-efficient manufacturing at massive scale. By working with both, OpenAI appears to be keeping its options open between a premium device and a more affordable one, or potentially building different models for different price tiers.
Chip development partnerships at this level typically involve signed engineering agreements and dedicated teams, which makes the collaboration itself difficult to misreport. What the partnership produces, however, is still an open question. No reporting has surfaced details about the processor architecture, and it remains unclear how much of OpenAI’s large language model capability the chip would be able to run locally versus offloading to the cloud.
That local processing question is central to the project’s pitch. Today’s AI features on iPhones and Pixel devices still rely heavily on remote servers to handle the most demanding tasks. Apple Intelligence processes some requests on-device through its Apple silicon, and Google has pushed its Gemini Nano model onto Pixel phones for select features, but both approaches have significant limitations. An OpenAI phone designed to run more powerful AI models directly on the handset could offer faster response times, better privacy, and the ability to function without a constant internet connection.
The agent-first vision
Several accounts of the project describe a phone that would replace traditional app-based interactions with AI agents capable of handling multi-step tasks directly. In this model, a user might ask the phone to plan a weekend trip, and the AI assistant would search for flights, book a hotel, coordinate calendar invites, and manage confirmations without the user ever opening a separate app.
That vision, if realized, would challenge the app-store economy that generates tens of billions in annual revenue for Apple and Google. It would also require deep integration between the AI layer and core phone functions like messaging, payments, and location services. Building that kind of seamless experience is something even Apple and Google have struggled to achieve on their own platforms, despite controlling both the hardware and the operating system.
No reporting has confirmed whether the OpenAI phone would run a modified version of Android, a fully custom operating system, or something else. That decision carries enormous consequences. A custom OS would give OpenAI full control over the user experience but would launch without the millions of existing Android and iOS apps that consumers expect. Running Android would provide app compatibility but could limit how deeply OpenAI can reshape the interface around its AI models.
Timeline and scale questions
The biggest point of disagreement across reports is when the phone would actually ship. One account, citing industry sources, places mass production in early 2027 with an initial run of 30 million units. A separate report suggests the timeline may stretch closer to 2028. OpenAI has not publicly commented on either projection.
The 30 million unit figure, if accurate, would be a remarkably aggressive opening. For comparison, Google shipped an estimated 10 million Pixel phones in 2023 after years of iterating on its hardware line. Hitting 30 million in a first production run would place OpenAI’s phone roughly on par with a mid-tier Android manufacturer’s entire annual output and would require a supply chain, distribution network, and retail presence that the company does not currently possess.
That gap between ambition and infrastructure is worth taking seriously. OpenAI has no carrier relationships, no retail stores, and no track record managing hardware logistics. The graveyard of failed smartphone challengers is well-populated: Amazon’s Fire Phone, Essential Phone, and multiple efforts from Microsoft all collapsed despite backing from companies with far more hardware experience than OpenAI has today. Even well-funded startups like Nothing have found it extraordinarily difficult to capture meaningful market share against Apple and Samsung.
Where Jony Ive fits in
OpenAI’s acquisition of io brought Jony Ive, the designer behind the iPhone, iMac, and Apple Watch, into the company’s orbit. Early reporting on that deal described a device concept focused on ambient computing, something closer to a wearable or screenless companion than a traditional smartphone. Whether the phone project reported in May 2026 is the same effort, a parallel track, or an evolution of the original concept remains unclear.
Ive’s involvement does lend the project a degree of design credibility that most hardware newcomers lack. His track record at Apple demonstrated an ability to create products that consumers form emotional attachments to, which is arguably as important as raw technical specifications in the smartphone market. But design alone has never been enough to sustain a phone business. Execution on software, services, carrier deals, and after-sale support matters just as much.
What this means for the AI-phone race
OpenAI is not entering a vacuum. Apple has been steadily expanding Apple Intelligence across its devices, integrating on-device processing with cloud-based models through what it calls Private Cloud Compute. Google has embedded Gemini deeply into Android and its Pixel hardware. Samsung has partnered with Google to bring Galaxy AI features to its flagship phones. All three companies have the advantage of controlling existing ecosystems with hundreds of millions of active users.
What OpenAI brings to the table is arguably the most recognized consumer AI brand in the world. ChatGPT crossed 400 million weekly active users by early 2025, a user base that dwarfs what most hardware startups can claim before launching a single product. The question is whether that software loyalty translates into hardware purchases, especially at a price point and form factor that has yet to be revealed.
For now, the strongest signal is strategic intent. OpenAI’s decision to invest in custom mobile silicon with two established chip partners, combined with its acquisition of a world-class hardware design team, points to a company that is serious about owning the device layer, not just the software running on someone else’s phone. How quickly that intent becomes a product consumers can buy, and whether it arrives in 2027 or later, will depend on engineering milestones, manufacturing partnerships, and business decisions that are still being made behind closed doors.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
