AI: The New Utility Bill?
OpenAI's Sam Altman predicts a future where AI is sold like water and electricity. But will this create a tech divide?
Imagine a world where AI is just another line item on your utility bill. That's the vision laid out by Sam Altman, CEO of OpenAI, at the BlackRock Infrastructure Summit in Washington, DC. According to Altman, artificial intelligence could soon be as ubiquitous, and as metered, as electricity and water, making intelligence a utility to be purchased on demand.
A New Kind of Utility
Altman envisions a future where tokens, which AI systems use to process and price data, become the currency of choice. 'We see a future where intelligence is a utility like electricity or water,' he said. In this setup, compute capacity, the backbone of AI models, will determine who gets access to this new form of power.
With demand for AI skyrocketing, the ability to offer compute capacity will become a competitive edge. But here's the catch: Not everyone gets to plug in. If OpenAI and other tech giants can't build enough capacity, AI could become a luxury item, accessible only to those who can pay top dollar or subject to government distribution.
The Infrastructure Race
Major tech companies are already in a race, pouring hundreds of billions into infrastructure to meet AI's growing appetite. AMD's CEO Lisa Su highlighted the need for a mind-boggling '10 yottaflops' of compute power, 10,000 times more than what existed globally in 2022, by 2031. Yet, even with this massive build-out, the road to more AI capacity is filled with hurdles.
AI data centers are power-hungry beasts, consuming as much energy as small cities. The strain on the power grid, coupled with slow permitting and transformer shortages, could bottleneck this expansion. Even Elon Musk chimed in on a podcast, suggesting China's rapid energy build-out might give it an edge in total AI compute over the US.
The Consequences of Scarcity
Within tech firms, compute resources are already scarce, leading to fierce competition among engineers who now negotiate AI compute budgets alongside salaries. OpenAI has committed a whopping $1.4 trillion on data center projects over the next eight years to try and stay ahead. But as OpenAI President Greg Brockman admitted last December, keeping up with demand is a monumental challenge, no matter how ambitious their dreams are.
So, what's all this mean for everyday folks? Will AI become just another utility bill, or will it widen the tech divide? In Buenos Aires, stablecoins aren't speculation. They're survival. The question isn't if AI will be sold like utilities, but rather, who gets to buy in. Who will explain AI to the street vendor in Medellín?
Get AI news in your inbox
Daily digest of what matters in AI.
Key Terms Explained
The science of creating machines that can perform tasks requiring human-like intelligence — reasoning, learning, perception, language understanding, and decision-making.
The processing power needed to train and run AI models.
The AI company behind ChatGPT, GPT-4, DALL-E, and Whisper.
The neural network architecture behind virtually all modern AI language models.