OpenAI is reportedly building an iPhone rival

For the past two years, AI startups have tried everything to convince us that smartphones are dead. They pinned badges to our shirts and sold us bright orange walkie-talkies. Turns out, nobody actually wants to throw their phone away.

Now, OpenAI seems to have gotten the memo. Despite years of insisting it had no interest in building a smartphone, the AI giant is reportedly working on an iPhone competitor of its own.

According to renowned supply chain analyst Ming-Chi Kuo, OpenAI is actively developing a smartphone targeting mass production in 2028.

Custom chips and a familiar factory

Kuo revealed that OpenAI is working with both Qualcomm and MediaTek to develop custom smartphone processors. Luxshare has reportedly signed on as the exclusive co-design and manufacturing partner.

Component suppliers and core hardware specifications are expected to be locked in by late 2026 or the first quarter of 2027, paving the way for a 2028 rollout.

This is a massive pivot from previous rumors. We already know OpenAI is teaming up with former Apple design legend Jony Ive on a family of hardware, but earlier leaks pointed to smart glasses, a lamp, custom earbuds, and a ChatGPT-powered HomePod-style speaker with a built-in camera. OpenAI's Chief Global Affairs Officer Chris Lehane even noted earlier this year that the company's first hardware device would arrive in the second half of 2026.

A full-blown smartphone, however, takes OpenAI straight into the deep end of personal computing.

Killing apps with autonomous agents

OpenAI doesn’t just want to build another glass slab running Android. Kuo reports that the device is being designed around AI agents that completely replace traditional apps.

Right now, Apple and Google tightly control system permissions, effectively boxing AI into isolated apps. By controlling the entire hardware and OS stack, OpenAI could give its AI deep, unrestricted access to manage tasks across the entire operating system.

The phone is expected to blend on-device small models with giant cloud models, continuously soaking in user context. With ChatGPT approaching 1 billion weekly active users, OpenAI has the scale to push a post-app interface where you simply state your goal and agents execute the background chores.

Not everyone agrees that hardware needs reinventing, though. Just last week, Perplexity CEO Aravind Srinivas argued that AI won't disrupt the iPhone at all, suggesting Apple's handset only becomes more indispensable as models get smarter.

Why it matters

We think this is the most clear-headed hardware decision OpenAI has made so far. Gadgets like the Humane Ai Pin proved that ditching the screen is a dead end. Instead of trying to bypass the smartphone, OpenAI is finally admitting that the phone remains the undisputed center of human digital life.

For consumers, 2028 is a long way out, so there is nothing to rush into today. But the concept of an agent-first device is tantalizing. If OpenAI pulls this off, you will no longer switch between dozens of siloed apps to book flights, message friends, or manage bills. The agent will just do it.

For the tech industry, this is a declaration of war against Apple and Google. Right now, OpenAI relies on Apple to distribute ChatGPT inside Siri. But playing in someone else's walled garden means living under their rules, their revenue cuts, and their privacy limits. Owning the hardware gives OpenAI direct access to raw, contextual user data that no standalone app can ever see.

That said, building hardware is notoriously brutal. Supply chain logistics, cellular certifications, and thermal constraints have crushed companies far larger than OpenAI. By 2028, Apple and Google will have several generations to bake their own agents into iOS and Android. OpenAI might be building a killer device, but it is walking straight into the lion's den.