BlackRock Summit Bombshell: Sam Altman Predicts AI Joins Electricity, Water, and Internet as a Paid Utility

OpenAI chief executive Sam Altman has predicted that artificial intelligence will join electricity, water and the internet as a paid utility, with users buying intelligence by the unit rather than through subscriptions. Delivering the remarks at BlackRock's U.S. Infrastructure Summit in Washington, D.C., in March, Altman described a coming era of metered AI access that could transform how the technology is consumed.

The comments have reignited discussions about the economics of artificial intelligence as the sector races to build the necessary infrastructure.

Altman's Vision for AI Billing

Rev reported that the OpenAI boss told delegates that the core business of model providers would revolve around selling tokens, the units used to measure computational effort. 'We see a future where intelligence is a utility like electricity or water and people buy it from us on a meter and use it for whatever they want to use it for,' he said.

This approach would let consumers and businesses tap into AI on demand, paying strictly for what they use – whether running complex reasoning tasks or background assistance. Altman stressed the aim was to flood the world with intelligence so that future generations expect instant access to advanced capabilities in any field.

An Instagram reel noted that 'Sam Altman says AI is increasingly looking like a utility, where people pay for intelligence the same way they pay for electricity, water, or cloud computing.' The prediction aligns with OpenAI's broader push towards abundance rather than scarcity in AI capabilities.

The Infrastructure Challenge Ahead

Realising this utility model will require unprecedented investment in compute power, data centres and energy generation. Altman highlighted the unique difficulties of the sector, noting that infrastructure must be committed years in advance and at enormous cost.

OpenAI as per CNBC has set a target of $600 billion (£446.6 billion) in total compute spend by 2030 to meet demand and avoid capacity constraints that could push prices higher or limit access to the wealthy. Broader industry spending as per Moneywise is even larger, with tech giants planning $700 billion (£521.0 billion) in capital expenditure this year alone to scale AI infrastructure.

Power supply remains a key bottleneck, though Altman expressed optimism about advances in model efficiency – citing a 1,000-fold cost reduction in solving hard problems between early and recent models – and new energy sources including nuclear.

Industry Bets on the Utility Future

The remarks come as artificial intelligence shifts from experimental tool to everyday essential, much like the internet two decades ago. Analysts say the metered model could democratise access if infrastructure scales successfully, but risks concentrating power among a handful of providers with the deepest pockets.

BlackRock's summit underscored the role of infrastructure investors in enabling the transition, with Altman announcing a new partnership with the North American Building Trades Unions to train workers for data centre construction.

As of June 2026, the debate continues over whether AI can truly become too cheap to meter or if usage-based pricing will simply add another line to household bills alongside electricity and water. The vision of AI as a paid utility is gaining traction as compute demand shows no signs of slowing.