Billionaire investor Mark Cuban has sounded the alarm on what he sees as a brewing storm in the artificial intelligence industry, warning that leading companies may be overspending in a race reminiscent of the 1990s search engine boom. His remarks come as Japan-based investment giant SoftBank, another major player in the sector, has made headlines by selling its entire stake in Nvidia—currently the world’s most valuable public company—freeing up billions to funnel into other AI ventures, especially those linked to OpenAI.
Speaking on the Pioneers of AI podcast, Mark Cuban directly addressed industry heavyweights including Perplexity, OpenAI, Anthropic, Google, Microsoft, and Meta, expressing concern that the frantic push to develop the world’s most powerful foundational AI model could end in a costly bubble. According to Business Insider, Cuban drew a parallel to the dot-com era, recalling how a crowded field of search engines ultimately yielded to Google’s dominance. “You’ve got five, six, whatever it is, companies that are trying to create the ultimate foundational model that we all depend on,” Cuban said. “It’s almost like in the ’90s when all the search engines were competing pre-Google… Now, we know with search engines it’s Google, and then there’s Bing, as, like, 1 or 2% and DuckDuckGo has got a half a percent. So it’s effectively a winner-take-all.”
Cuban’s warning isn’t just about market share. He’s deeply concerned about the sheer scale of investment pouring into AI, especially the vast sums being spent on infrastructure like massive data centers. “They may be overspending,” he cautioned. “And if they overspend or get too caught up, the bubble is in the competition between all those models because that could pop just like that with any new technology.” His skepticism centers on the idea that technology will inevitably improve, making today’s costly investments potentially obsolete. “I just can’t imagine over a 10-year period that we aren’t going to improve the technology enough that if you overspend on today’s technology, it just doesn’t feel right to me,” Cuban added.
He further suggested that the real disruption in AI is unlikely to come from incremental improvements by the current leaders. Instead, he predicted a game-changing breakthrough from an unexpected source. “Somebody’s going to come up with some incredible shit, right? If I knew what it was, I’d do it,” Cuban quipped, hinting that the present arms race could be blindsided by innovation from outside the usual suspects.
Meanwhile, SoftBank, one of the world’s largest investment holding companies, has been making its own seismic moves in the AI sector. As reported by The Motley Fool, SoftBank sold its entire Nvidia stake for about $5.8 billion on November 26, 2025, marking the second time the company has fully exited its position in the chipmaker. SoftBank, which had previously invested $4 billion in Nvidia in 2017 and cashed out in 2019, began buying shares again in 2020, amassing 32.1 million shares before this latest sale.
The rationale behind SoftBank’s exit is strategic rather than a loss of faith in AI or Nvidia specifically. Instead, the company is reallocating capital to deepen its partnership with OpenAI, the creator of ChatGPT. SoftBank has committed around $30 billion to support OpenAI’s expansion and model development, signaling a shift from investing in AI infrastructure—such as Nvidia’s GPUs, networking equipment, and AI chips—to betting on the software side of the AI pipeline, including applications, models, and enterprise tools.
SoftBank’s ambitions don’t stop there. The company is a key player in the proposed Stargate Project, a joint venture with OpenAI, Oracle, and investment firm MGX, which aims to invest a staggering $500 billion in AI infrastructure projects over the next four years. Such a commitment requires significant financial maneuvering, and the sale of Nvidia shares is a crucial part of freeing up capital for these new ventures.
Despite SoftBank’s high-profile divestment, industry observers caution against interpreting the move as a bearish signal for Nvidia or the AI sector at large. “SoftBank’s sale may be large in dollar terms (around $5.8 billion), but it’s a drop in the bucket for a company worth over $4.3 trillion (as of Nov. 21),” The Motley Fool noted, emphasizing that individual investors should make decisions based on their own goals and risk tolerance rather than following corporate giants blindly. Nvidia’s own performance has been nothing short of spectacular, with its stock surging over 1,280% in the past five years and recent earnings reports showing total revenue up 62% and data center revenue up 66%.
For SoftBank, the decision to pivot away from Nvidia and toward OpenAI and other AI software initiatives is consistent with its history of seeking out the next big thing. The company’s portfolio is stacked with high-profile names like ByteDance (TikTok’s parent), DoorDash, Fanatics, Chime, and IonQ. By shifting focus from hardware infrastructure to the software and applications layer, SoftBank is betting that the future of AI will be shaped not just by the chips that power models, but by the platforms and tools that put those models to work.
Mark Cuban, for his part, sees both peril and opportunity in this rapidly evolving landscape. He believes the current environment, with companies “anticipating for at least another decade spending every penny they have,” is ripe for disruption. “I mean, if that’s not ripe for disruption to come up with better ways, I don’t know what is,” Cuban observed. His advice, drawn from the lessons of the dot-com era, is to remain wary of hype and to watch for the unexpected—because, as history has shown, it’s often the unforeseen innovation that changes everything.
For investors and industry watchers alike, the message is clear: the AI arms race is far from over, and the winners may not be the ones spending the most today. With giants like SoftBank and outspoken voices like Cuban shaping the conversation, the next decade promises both extraordinary advances and, perhaps, a few cautionary tales.
