Enterprise DNA

Omni by Enterprise DNA

Enterprise DNA Resources

Latest AI and industry news. Practical AI operating-system thinking for owners, operators, and teams doing real work.

220k+

Data professionals

Omni

AI agents and apps

Audit

Map the manual work

News Breaking Product

OpenAI Is Building a Phone Where AI Agents Replace Apps

Reports confirm OpenAI is developing a smartphone with Qualcomm and MediaTek that runs AI agents instead of apps, targeting 2028 launch.

Enterprise DNA | | via TechCrunch
OpenAI Is Building a Phone Where AI Agents Replace Apps

OpenAI is reportedly building a smartphone. Not as a side project and not just a ChatGPT-branded device — but a full hardware play where AI agents replace the app model entirely.

Reports from TechCrunch and Fast Company, citing analyst Ming-Chi Kuo and multiple supply chain sources, confirm that OpenAI has selected Qualcomm and MediaTek as its chip development partners, with Luxshare designated as the exclusive manufacturing partner. Specifications are expected to be finalized by late 2026 or Q1 2027, with mass production targeting 2028. The ambition: 300 to 400 million annual smartphone shipments.

Qualcomm’s stock jumped 13 percent on the news.

The Vision: Agents Over Apps

The core idea is a fundamental rethink of how people use a phone. Instead of opening Instagram to scroll, Uber to book a ride, or Slack to message a colleague, you describe what you want and an AI agent handles it in real time — pulling context from multiple sources, executing the task, and returning a result. No switching between apps. No app store. No Apple or Google sitting in the middle.

Kuo framed it clearly: “Only by fully controlling both the operating system and hardware can OpenAI deliver a comprehensive AI agent service.”

That quote matters. It explains why OpenAI needs its own hardware. Right now, Apple and Google control the gates — what apps can access, what sensors agents can use, what runs in the background. A third-party app on an iPhone or Android device is always constrained. OpenAI’s own hardware removes those constraints entirely.

Sam Altman has been pointing toward this for months. He has said publicly that the app model will give way to AI agents that handle tasks across systems, adapting to the user in real time rather than making the user adapt to software.

This Is About Enterprise Software, Not Just Phones

Most coverage will focus on consumer implications — will people buy an OpenAI phone? That is the wrong question to start with.

The real signal is the enterprise software model. If AI agents can replace apps on a consumer device, the same logic applies inside a business. Why maintain 30 different software subscriptions when an agent could connect to the underlying data and APIs directly?

This is already happening in pockets. Companies using AI agents to replace manual workflows in finance, operations, and customer service are not replacing apps, they are working around them. The OpenAI phone is that logic extended to hardware.

The consequences for the software industry are significant. Every platform vendor — Salesforce, SAP, ServiceNow, Atlassian — has spent years building apps that sit between people and their work. An agent-first model challenges that entire architecture.

What This Means for Business

This is still three to four years from mass production, which means the immediate disruption is slow. But the direction it signals should inform how business leaders think about software investment today.

A few practical implications:

The app layer is becoming optional. Businesses locked into sprawling SaaS portfolios should be asking whether those tools will survive an agent-first world or whether they are paying for middleware that AI will bypass.

Custom software becomes more valuable, not less. If agents need structured data, clean APIs, and clear processes to work against, companies with solid data foundations and purpose-built internal tools will deploy agents far more effectively than those running 20-year-old legacy systems with no API access.

Whoever controls the hardware controls the experience. Apple has proven this. OpenAI is betting the same model works in AI. For enterprise buyers, this will create new vendor lock-in questions — not just which cloud or which software, but which hardware and agent runtime.

Training and literacy still matter. An agent-first world does not make people irrelevant — it makes people who understand how agents work, what they can do, and where they fail more valuable. The gap between organizations that invest in AI capability and those that do not will widen, not narrow.


If this is the kind of problem agents can help with, the free Working With Claude field guide is the practical next step. Thirty-two pages, no fluff. Get the free guide.

The phone itself may or may not succeed. The direction it represents almost certainly will.

Working With Claude field guide cover

Free Resource

Going deeper with Claude?

Get the free 32-page implementation guide for ANZ teams.

No spam. Unsubscribe any time.