Altman says AI lacks iPhone moment, OpenAI revenue hits $6.7B

OpenAI Chief Executive Sam Altman said AI has yet to experience its "iPhone moment" — a product inflection point that changes human-computer interaction — and acknowledged adoption has lagged since GPT-4 launched in 2023.

"I thought when we got to GPT-4, which was back in 2023, that very quickly after that there was going to be much more disruption, software businesses up for grabs right away, than it turned out to be," Altman said on a podcast hosted by David Senra, aired August 24. "We've all been too ambitious on timelines."

Altman said the economy and society carry "so much inertia" that even powerful technology takes time to reshape behavior. People continue buying from the same companies, using familiar tools, and working in established ways. He compared the current phase to the smartphone era before the iPhone, when devices like Palm already contained relevant technology but lacked a unified interaction model that made users abandon old habits.

The remarks carry weight for an industry that has absorbed billions in investment since ChatGPT's launch in late 2022. OpenAI's revenue rose 18 percent quarter over quarter to $6.7 billion in the second quarter, while rival Anthropic's sales doubled in the same period, according to The Wall Street Journal — a gap that has put OpenAI on the back foot and forced a business and leadership overhaul.

What's Missing Isn't Models — It's a New Way of Interacting

Altman said even AI insiders remain constrained by old habits. He acknowledged that despite having Codex, OpenAI's coding assistant, his own approach to using computers hasn't fundamentally changed in 20 years — he still copies and pastes information, processes email, and maintains traditional to-do lists.

The problem stems partly from human behavioral inertia and partly from insufficient product capability, he said. Users today find themselves caught between "two worlds": continuing legacy computer operations while experimenting with AI tools, without clarity on when to delegate tasks to AI. The true "iPhone moment" isn't merely the first appearance of a technology — it's when a product makes the public feel there's no going back to the old way.

Altman believes AI already has most of the necessary technological components but has yet to produce a unified interface that reorders the relationship between people and computers. He hopes for an AI agent that continuously understands the user, holds vast personal context, and proactively offers assistance — capable of processing information the user lacks time to read and providing advice on major decisions.

OpenAI Bets on Models, Compute, and a Platform Approach

On strategy, Altman said OpenAI wants to operate more like a platform company than a product company covering every category. The company has already integrated ChatGPT with Codex, and the future focus is on providing a unified AI interface along with APIs that allow other developers to build applications on top. OpenAI does not intend to manufacture every product and will not try to "swallow the entire economy."

He described compute infrastructure buildout as potentially one of the most expensive infrastructure projects in human history, involving chip R&D, wafer fab supply chains, server racks, data centers, power systems, financing, and policy coordination. Altman revealed that OpenAI had discontinued Sora and the Atlas browser, redirecting those resources to Codex and upstream areas such as models, chips, and infrastructure software.

On safety, Altman summarized his greatest concerns into two categories: loss of control once model capabilities become too powerful, and concentration of power in a single company, model, or individual. He criticized industry voices that use safety as a pretext to restrict ordinary people's access to the most powerful models, calling that an "anti-human" proposition. OpenAI's approach is "iterative deployment" — gradually introducing models into real-world environments and identifying alignment flaws through user feedback.

For investors, Altman's remarks suggest AI's commercialization timeline may not advance in lockstep with technological iteration. Rapid improvements in model performance do not necessarily mean enterprise processes, consumer habits, and industry structures will change immediately. OpenAI's revenue of $6.7 billion in Q2, while growing 18 percent quarter over quarter, still trails Anthropic's growth trajectory — and the competitive gap is widening as Claude Code gains traction among developers. The decisive product may well be an interaction system that lowers switching costs and integrates naturally into existing workflows, rather than any single model release.

This article is for informational purposes only and does not constitute investment advice.