//php echo do_shortcode(‘[responsivevoice_button voice=”US English Male” buttontext=”Listen to Post”]’) ?>

ChatGPT creator OpenAI is working closely with Qualcomm and MediaTek to develop processors for an AI-first smartphone, which is reportedly co-designed and manufactured by China’s Luxshare with a launch planned in 2028. The news comes from Ming-Chi Kuo, an analyst at China-based TF International Securities, who is known for sniffing out plans of large tech firms through supply-chain sources.

Kuo, based in Taiwan, has made accurate projections about Apple’s products in the past. He claims that while OpenAI plans to mass-produce this AI smartphone by 2028, its hardware specifications will be finalized by early 2027.

Kuo also explained in his X post why AI agents will shape the OpenAI smartphone, making it work and feel very different from an iPhone. More importantly, he reckons that AI agents will replace smartphone apps, and that it’s going to require both on-device edge intelligence and cloud AI integration.

Kuo also sketched out a potential revenue model in which OpenAI might tie hardware sales to subscription plans and cultivate a developer ecosystem around AI agents. He added that OpenAI’s high-end smartphone is likely to entail annual shipment volumes of 300 million to 400 million units.

Is Copper Sintering the Key to Advancing Wide Band Gap Semiconductors?

By Rejoy Surendran, Market Strategy Manager & Xinpei Cao, Sr. Principal, Application Engineering, Henkel  04.27.2026

SK hynix Receives 2026 IEEE Corporate Innovation Award

By SK hynix  04.26.2026

Intelligent Cost Optimization is a Competitive Advantage  

By TechInsights  04.24.2026

But there is a twist in OpenAI’s smartphone story.

OpenAI’s smartphone about-face

Last year, when OpenAI acquired Jony Ive’s startup io Products for $6.4 billion, CEO Sam Altman said the io collaboration would yield a new category of AI devices that complement phones and laptops rather than replace either. These devices, optimized for on-device AI tasks, would enable faster local inference for smaller models, improve power efficiency, and offer agent-like capabilities.

For instance, there have been industry reports on consumer-oriented AI devices such as AI-powered earbuds. The device, while heavily relying on cloud-based AI processing, would be powered by Samsung’s 2-nm Exynos chip for on-device processing. Then, there has been talk of a screenless device taking the form factor of a pen or pendant.

Ive, former chief designer at Apple, would lead the creation of what Altman called a “third core device” alongside phones and laptops. In short, OpenAI, while keen to break into hardware, wasn’t focused on smartphones. But a day before Kuo’s post, Altman wrote on X, “While OpenAI clearly believed at one point that it could disrupt the iPhone with its first AI devices, perhaps it now sees the smartphone as central to AI uses for many years to come.”

So, what led to this about-face?

It’s likely that, while OpenAI believed it could disrupt the iPhone with its new category of AI devices, it now sees the smartphone as central to AI use for many years to come. Kuo reinforced this premise, saying that smartphones will remain the largest-scale device category for the foreseeable future. So, it makes sense for OpenAI to develop a device entirely run by AI, since the company has accumulated user data over the years.

“The smartphone is the only device that captures the user’s full real-time state, which is the most important input for real-time AI agent inference,” he wrote in his post. “Only by fully controlling both the operating system and hardware can OpenAI deliver a comprehensive AI agent service.”

OpenAI’s smartphone could aim to capitalize on the device as the most valuable source of data for real-time AI inference.

Kuo believes that AI agents will shape the OpenAI smartphone, making it work and feel very different from an iPhone. However, he acknowledges that OpenAI will need to emulate Apple’s ‘whole widget’ approach by controlling both hardware and software.

Smartphone design wars

Last year, Qualcomm announced that OpenAI’s first reasoning model, gpt-oss-20b, is running on devices powered by its Snapdragon processors. This 20B parameter model—enabling chain-of-thought reasoning entirely on-device—provided a glimpse into the future of AI where assistant-style reasoning could become local.

The gpt-oss-20b model is designed for lower latency, local, or specialized use cases without a cloud dependency.The gpt-oss-20b model is designed for lower latency, local, or specialized use cases without a cloud dependency. (Source: OpenAI)

Qualcomm, with early access to the gpt-oss-20b model, performed integration testing with Qualcomm AI Engine and the Qualcomm AI Stack. Next, OpenAI’s gpt-oss-20b model running on Snapdragon processors enabled devices to leverage on-device inference while complementing cloud solutions via AI agents.

However, while Qualcomm has incorporated neural processing units (NPUs) into its Snapdragon processors, the OpenAI smartphone affair isn’t just another Android processor tweak. It marks an attempt to build silicon for ChatGPT-style experiences that’s device-native, rather than cloud-dependent. Nevertheless, it broadens Qualcomm’s smartphone narrative and poses a potential threat to iPhone maker Apple.

It’s important to note that Apple has already partnered with OpenAI to integrate ChatGPT into Siri; Apple is also opening Siri to third-party AI assistants such as Google’s Gemini. In fact, Apple was forced to rely on Google’s Gemini AI models to power its upcoming Siri voice assistant. So, while OpenAI’s smartphone is expected to arrive by 2028, it gives Apple breathing room to get its Siri act together and build an “AI-native” iPhone from the ground up with deeper agentic capabilities.

OpenAI seems to be challenging Apple’s iPhone turf with a smartphone that will serve as a real-time data repository to enable continuous AI agent inference. That clearly poses a competitive threat to both Apple and Samsung, which together account for roughly 40% of global smartphone sales.

The next round of smartphone design wars is in sight now.

See also:

Open-Weight Models and the Future of AI

How to Plan Agentic AI Deployment for Chip Design

Arm Unveils Lumex Platform, Boosting AI Processing at Edge