Sam Altman rarely describes the future of artificial intelligence (AI) in narrow terms. When asked where the technology will ultimately go, he portrayed AI as a force that can rapidly move beyond software, reshape physical infrastructure, accelerate scientific discovery, and drive broad-based economic growth.
“You can imagine billions of humanoid robots building more data centers, mining more materials, building more power plants,” OpenAI’s chief executive said in a conversation with Cisco CPO Jeetu Patel at Cisco’s AI Summit on Tuesday, February 3. “With all kinds of amazing new services and scientific discoveries, you can imagine the economy growing at an unprecedented rate.”
But the image anchored a broader theme that repeatedly surfaced throughout his remarks. The idea is that AI’s trajectory is no longer limited to narrow use cases and productivity improvements. It is moving toward systemic change, even though most companies are not structurally ready to absorb it.
Upper limit: Fully AI-powered enterprise
When the conversation turned to where today’s AI systems ultimately go, Altman focused on capabilities rather than short-term implementation. “
I think the upper limit is a full AI company,” he said, describing organizations where AI systems are active participants in how work gets done, rather than tools layered on top of workflows.
He added that a key turning point is the move from models that produce output to agents that can interact directly with computers.
Advertisement: SCROLL TO CONTINUE
“Code is really powerful,” Altman says. “But code and general computer use are even more powerful.”
With agents that can navigate browsers, applications, and authenticated environments, AI can complete tasks end-to-end rather than stopping at recommendations or drafts. Altman suggested that once you experience that interaction model, it becomes difficult to think of AI as a passive system waiting for human prompts.
He extended that logic beyond individual workflows to coordination and collaboration. Altman described the potential for entirely new interaction models in which agents communicate with each other on behalf of humans. He positioned this as a natural outcome of increasing capabilities, an interaction system designed primarily for machines to exchange information and coordinate tasks, rather than humans manually managing the exchange.
Security and data access remain the most challenging issues
Despite rapid advances in AI capabilities, the most binding constraints are no longer technical, Altman said.
“How do you balance the security and the kind of data access and the practicality of all these models?” he asked. Existing security and authorization systems were designed for human users making individual, intentional requests. They are not well suited for continuous monitoring and always-on agents operating system-wide.
“I feel like this requires us to invent a new kind of security or data access paradigm,” Altman said. Until that happens, he said, organizations will continue to limit their adoption of AI, even as capabilities advance.
Altman repeatedly returned to what he described as a widening gap between what AI systems can do and what companies are ready to deploy.
In his view, this gap is not caused by the technology itself, but by unresolved issues around governance, security, and data access. As a result, adoption of tools, even those that already exist, is delayed.
“I feel like it’s really important to think about how to set up a company so that it can absorb these new tools quickly, without spending years of internal friction and debate,” Altman said.
He warned that delays could affect competition. If companies fail to adapt their structures quickly enough, they may fall behind not because the technology is not available, but because they are not prepared to work with it.
“I don’t want to make too dramatic a prediction, but companies that aren’t equipped to quickly deploy what you might call AI colleagues will be at a huge disadvantage,” Altman said.
