Artificial intelligence may be creating one of the largest investment booms in history. But billionaire venture capitalist Bill Tai thinks the more important story is what comes after all the data centers get built.
The longtime tech investor, who backed companies including Zoom and Canva and served for years as the founding chairman of Bitcoin miner-turned-AI infrastructure company Hut 8, believes AI is approaching a point where it begins changing not just technology, but the basic economics of how work gets done.
"We are in a moment that is going to re-define all of technology and the way we live every day," Tai told Coinage. "I know that sounds a little bit gargantuan, but it's true."
His thesis stretches from Nvidia's booming AI infrastructure business to the future of white-collar employment, autonomous vehicles and even stablecoins.
At the center of it is a resource Tai thinks investors may soon come to view very differently.
"There was a saying that I used to use around 2011: Data is the new oil," Tai said. "For now, it's basically compute is the new oil." That shift is already beginning to show up in the physical economy.
Nvidia increasingly describes the massive data centers being built around its chips as "AI factories," infrastructure that effectively converts electricity and computing power into tokens generated by AI models and agents. For Tai, that helps explain why the AI boom is fundamentally different from previous waves of software investment.
Hyperliquid Is Up 220% And Just Hit An All-time High. Why It Could Keep Climbing
Why A New Bitcoin Bull Market Is Finally Here
Google dramatically lowered the cost of finding information. AI can increasingly perform the intellectual work that came after finding it. Tai describes that as a move from simply giving humans access to information to giving them access to the knowledge that can be extracted from it.
"All of that knowledge can then be sucked into one node of compute power to do what you would have done searching through 50,000 documents to come up with a solution," Tai said.
Tai argues that the crypto mining industry unknowingly developed many of the techniques the AI industry now needs.
Bitcoin miners progressed from CPUs to GPUs and eventually specialized ASIC machines. Ethereum also relied heavily on GPUs before abandoning proof-of-work. In both cases, miners had to figure out how to secure massive quantities of electricity, cool dense computing equipment and operate industrial-scale data centers.
"A lot of the techniques that are in practice today for running gigantic — and now these are a lot bigger than the Bitcoin mining sites — giant computation centers that Jensen would refer to as AI factories," Tai said, "were all pioneered in the crypto industry."
That history has suddenly become extremely valuable.
Hut 8 now reports 949 megawatts of contracted AI data center capacity representing approximately $26.6 billion in expected aggregate base-term contract value. Its Beacon Point project in Texas is being designed around Nvidia's DSX reference architecture.
In other words, a company built to mine Bitcoin is increasingly monetizing one of the resources the AI boom needs most: enormous amounts of power connected to enormous amounts of compute.
But Tai thinks what happens inside those data centers will ultimately matter much more than the buildings themselves.
The uncomfortable implication of AI becoming better at transforming information into useful work is that a lot of the global economy currently pays humans to perform exactly that function.
Tai compares the shift to the mechanization of agriculture.
Before tractors, harvesters and other machines dramatically improved agricultural productivity, farming required huge amounts of human labor. Mechanization made individual workers dramatically more productive, allowing farms to grow larger while requiring fewer people to produce the same amount of food.
The result was enormously beneficial for economic productivity. It was also deeply disruptive to the workers living through the transition.
"Now we have the same kind of displacement and acceleration of productivity in some of the nodes for information workers," Tai said.
That could make AI adoption less of a choice for corporations than it sometimes appears.
Tai argues that once one competitor discovers how to use AI to accomplish the same amount of work with fewer resources, every other company in the industry has an incentive to follow.
"Companies by their very nature are responsible to their shareholders to drive the highest ROI they can in a reasonable way," he said. "And there's always a fear that if you don't do it and all of your competitors do, you're out of business."
That may be one of the more important ways investors should think about AI.
The question is no longer simply whether companies want to spend money experimenting with artificial intelligence. Increasingly, it may be whether companies can afford not to if their competitors begin realizing material productivity gains.
Tai said corporate AI budgets are already being consumed much faster than many companies expected as workers experiment with models from companies including Anthropic and Google.
That dynamic could eventually divide public companies into very different groups: those selling the infrastructure necessary to power AI, those successfully using AI to increase margins and growth, and those whose existing business models are being automated away.
Zoom Communications, one of Tai's early investments, offers an example of how an incumbent software company might land in the second category.
The original value of Zoom was relatively straightforward: making virtual meetings easier.
But once AI can understand everything said during those meetings, Tai argues the economic value of the platform potentially expands.
A transcript can become a summary. The summary can identify decisions. Those decisions can create assignments. An AI system can then track whether those assignments were ever completed.
"Zoom is becoming a workplace AI automation instrument," Tai said. The more difficult future may belong to software products whose primary function can be replicated entirely by an AI agent.
Tai pointed to accounting and tax preparation as examples of highly rules-based work that may prove relatively straightforward to automate. He stressed that he was not making a specific bearish call on Intuit, the company behind TurboTax and QuickBooks. But the example reveals the larger economic threat and white-collar jobs may only be the beginning.
Tai pointed to Uber, Lyft and DoorDash as companies whose gig-economy platforms have helped provide income to people navigating changes elsewhere in the labor market. Then he pointed to Waymo, Alphabet's autonomous driving business is now providing fully autonomous trips in 14 cities and continues expanding into new markets.
"What's going to happen as some of these other, not just white-collar workers in front of screens, but also some of these kinds of functions get automated with AI and robotics?" he asked. "It's something to think about."
It is a particularly difficult economic problem because technology can eliminate an existing job faster than society can imagine the profession that ultimately replaces it. But Tai is still an optimist.
His argument is that nearly every major jump in productivity has looked threatening when viewed through the jobs it displaced rather than the opportunities it eventually created.
Agricultural mechanization reduced the number of people required to work farms. Industrialization created different jobs. Computers eliminated certain types of clerical work while creating industries that could scarcely have been imagined before them.
Tai thinks AI could eventually be remembered the same way.
"If we clock forward 30 years from now and we look back at a world where you had to sit in front of a computer screen with a keyboard for ten hours a day or more, that's not very exciting," he said.
The jobs created on the other side are impossible to predict, which is precisely what makes the transition frightening.
But Tai believes removing more repetitive work could ultimately give humans the opportunity to spend their time differently.
"As productivity shifts forward, I think people, as newer jobs get created — and no one knows exactly what those will be — it's like you're getting unlocked," he said. "It's freedom."
Subscribe to the free Coinage newsletter to stay up-to-date on all things crypto and finance.