OpenAI CEO Sam Altman has made a striking strategic choice: the company is shutting down Sora, its video generation tool, and Atlas, its web browser, to funnel resources into compute infrastructure and AI research instead. In a recent interview, Altman explained that these were "good" products, but protecting OpenAI's research and computational capacity took priority. This decision reveals how the world's most prominent AI company is fundamentally reshaping its business model away from building consumer-facing products and toward becoming a platform that powers others' applications.

Why Is OpenAI Abandoning Consumer Products?

Altman's reasoning centers on a simple but radical premise: OpenAI is not a product company, it's a platform company. Rather than competing with customers by building standalone applications like video generators or web browsers, OpenAI wants to provide the underlying AI infrastructure that everyone else builds on top of. The company recently merged ChatGPT and Codex, its coding tool, into a single interface and is aiming to reach 100 million new businesses and 8 billion people building applications on its platform.

The compute and research investments are already staggering. Altman described the infrastructure project as "probably already the most expensive infrastructure project in history," which includes building OpenAI's own chips and data centers. By killing Sora and Atlas, the company freed up engineering talent and capital that would have been spent maintaining and improving those products. Altman acknowledged that Atlas was "the best web browser," but the strategic calculus was clear: general intelligence for knowledge work mattered more than excelling in any single product category.

Altman

What Does OpenAI's Platform Strategy Actually Mean?

OpenAI's pivot reflects a belief that the real value in AI isn't in any single application, but in providing a flexible foundation that developers and businesses can customize for their own needs. The company is selling "AI at every point on the cost-performance curve," meaning it offers different versions of its technology at different price points and capability levels, rather than forcing everyone into one product box.

This approach mirrors how cloud computing companies like Amazon Web Services operate: they don't build every application themselves, they provide the infrastructure that makes it possible for others to build applications. OpenAI is betting that this model will generate far more value and reach than trying to compete in every consumer software category.

How to Understand OpenAI's Long-Term Vision

• Platform Over Products: OpenAI is consolidating its consumer-facing tools into a single interface and API (Application Programming Interface), allowing external developers to build on top rather than competing with them in specific product categories.
• Compute as the Bottleneck: The company views computational infrastructure, including custom chips and data centers, as the limiting factor for AI progress, not the ability to build polished consumer applications.
• Scale Through Partnerships: By positioning itself as a platform, OpenAI aims to reach 100 million businesses and 8 billion people indirectly, through applications built by others, rather than through direct user acquisition.

Is the AI Economy Moving Slower Than Expected?

Altman also revealed a humbling realization: the broader economy is adopting AI more slowly than he anticipated. After GPT-4 launched in 2023, he expected rapid disruption of software businesses. Instead, he found that "people keep buying from the same companies and using tools the same way." He compared the current moment to the smartphone era before the iPhone: all the technological pieces exist, but the product breakthrough that makes AI feel essential to everyday life hasn't arrived yet.

This admission is particularly striking because Altman himself hasn't fully integrated AI into his own workflow. He acknowledged that "the most psychologically inconsistent thing" about himself is that he still uses his computer the old way, scrolling email and maintaining a to-do list, despite having access to Codex. He called this "mostly a product failure," suggesting that AI hasn't yet had its "iPhone moment" for how people interact with technology.

"The economy has more inertia than he expected. Altman said he was wrong on timelines after GPT-4 in 2023: disruption of software businesses is coming slower because people keep buying from the same companies and using tools the same way," noted Altman in the interview.

Sam Altman, CEO at OpenAI

What Are Altman's Biggest Concerns About AI's Future?

Despite his optimism about AI's potential, Altman identified two major worries: loss of control over AI systems, and power becoming too centralized in one company or person. With around a billion weekly users, OpenAI has significant influence over how AI is deployed globally. Altman argued that safety comes from shipping imperfect models into the world and learning from real-world deployment, citing the Federal Aviation Administration's accident-reporting culture as a model. This pragmatic approach contrasts with calls from some researchers for more caution before deploying advanced AI systems.

Altman also criticized the AI industry, including OpenAI, for failing to explain AI's benefits and address public fears effectively. He noted that too many builders in the field are simultaneously saying there's a large chance of catastrophe while racing ahead to build more powerful systems. This communication gap, he suggested, is undermining public trust in the technology.

What Does This Mean for AI Startups and Small Businesses?

Despite the slower-than-expected adoption curve, Altman predicted "the greatest boom ever in people starting small businesses, empowered by AI." His platform strategy is designed to enable exactly this outcome: by providing accessible AI infrastructure, OpenAI hopes to lower the barriers for entrepreneurs to build AI-powered applications without needing massive engineering teams or research budgets.

Altman

The decision to kill Sora and Atlas, while painful in the short term, reflects Altman's conviction that OpenAI's role is to be the infrastructure layer, not the application layer. If that bet pays off, the company's influence over the AI economy could be even larger than if it had succeeded with individual consumer products. But it also means OpenAI is betting that developers and businesses will choose to build on its platform rather than competitors' alternatives, a wager that will play out over the next several years.