As October began, The Washington Post business columnist Bethany McLean cited these two seemingly contradictory quotes about the future of AI:“AI could eventually cure cancer, end poverty, and even bring world peace.”— Anthropic founder Dario Amodei (on his blog one year ago)And:“The amount of money being spent is not proportionate to the money that’s coming in.”— Dr. Sasha Luccioni, AI researcher at Hugging Face (talking to The New York Times)Then, she reminded readers that “Two things can be true at the same time. Bets on AI are going to cause financial calamity. And AI is going to change the world.”And this two-headed reaction seems to be taking hold. While citing hopes AI could “reshape multiple industries, cure diseases and generally accelerate human progress,” Bloomberg wrote earlier this month that “never before has so much money been spent so rapidly on a technology that, for all its potential, remains somewhat unproven as a profit-making business model.” (The headline? “Why Fears of a Trillion-Dollar AI Bubble Are Growing.”)In August, CNBC published quotes from The Verge’s interview with OpenAI CEO Sam Altman in August. “Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes,” Altman said. “Is AI the most important thing to happen in a very long time? My opinion is also yes.”Even Amazon’s Jeff Bezos says we’re in an “industrial bubble,” while adding that the technology is “real” (in an onstage interview at the Italian Tech Week conference). Bezos sees stock prices “disconnected from the fundamentals” (according to a report from CNBC). But Bezos quickly went on to say, “AI is real, and it is going to change every industry.”“The [bubbles] that are industrial are not nearly as bad, it can even be good, because when the dust settles and you see who are the winners, societies benefits from those inventions. That is what is going to happen here, too. This is real, the benefits to society from AI are going to be gigantic.”Tom Orlik, global chief economist for Bloomberg Economics, nicely sums up our current dilemma.“Artificial intelligence might be a bubble. It is also a juggernaut.”The Case for and Against an AI BubbleIn an October blog post, Nobel Prize-winning economist Paul Krugman explored what’s behind concerns about a bubble. “Reinforcing this suspicion is the fact that big tech companies, which generate billions in cash flow, are spending more on AI than their gushers of dollars can support. So now they’re taking on lots of debt.” (Krugman links to a Bloomberg article noting debt financing isn’t limited to OpenAI, but also companies like Meta and Vantage Data Centers.)The Associated Press also reports that investors are closely scrutinizing “a series of intertwined deals over recent months” — for example, OpenAI’s deals with chipmakers NVIDIA and AMD — which seem to link the companies’ fortunes together.But last week, Yahoo Finance executive editor Brian Sozzi determinedly argued the case against a tech bubble. “It’s just not what’s happening out there yet.” Instead, Sozzi presented his counterargument, that AI “is a real technology being deployed in real ways inside of Corporate America.” To the naysayers, he says “the spending on AI infrastructure doesn’t strike me as reckless. … The tech giants making the biggest AI investments are fueling their ambitions by cash on hand — not loading up balance sheets with debt.” Even the smaller upstarts are “well-funded” and “working on tangible technology that has actual orders behind it.”Things were worse in the 1990s, when the biggest seven companies had a median “forward price to earnings ratio” that was double, writes Sozzi (citing new research from Goldman Sachs). Company strategist Peter Oppenheimer argued today’s valuations are high, “but, in our view, generally not at levels that are as high as are typically seen at the height of a financial bubble.”And, “In a perfect world, the bubble doesn’t pop, but instead gets a small leak,” added TNS analyst Lawrence Hecht. “That best-case scenario would result in a soft landing.”Will AI Replace Human Workers?After all, AI’s untapped potential is huge, some have argued, because it could even replace human laborers in the future.But what if it can’t? Is that what dashes all those hopes for vast future profits?Krugman is still seriously concerned about AI eliminating large categories of jobs — but remains skeptical about mass layoffs. “People have been predicting mass unemployment caused by automation since the 1930s, and it keeps not happening.”And Hecht agrees that here again, two things can be true. “I think the potential for AI has created ‘rational’ exuberance that extends beyond technophiles and futurists.” But at the same time, “I think there is a huge bubble in terms of expectations about productivity increases. From about 2006 to 2020, productivity growth had slowed after the initial rapid growth due to the internet and other IT innovations. It appears that has increased again, but I have not seen any evidence that productivity growth will be greater than it was in the late 90s.”Of course, there still remain many other paths to profitability. One example? OpenAI’s ChatGPT seems to be hoping to monetize its free users with advertising, according to recent job listings spotted by Adweek.What Happens After the AI Bubble Bursts?But the voices worrying about a bubble are growing louder. Last Sunday, 60 Minutes even broadcast an ominous report titled “Booms, Busts and Bubbles” featuring Andrew Ross Sorkin, the co-host of CNBC’s Squawk Box. “I think it’s hard to say we’re not in a bubble of some sort,” Sorkin said to millions of viewers. “The question is always, when is the bubble going to pop?”Lesley Stahl: Do you think that we will have a crash or not?Andrew Ross Sorkin: The answer is we will have a crash; I just can’t tell you when, and I can’t tell you how deep. But I can assure you, unfortunately, I wish I wasn’t saying this, we will have a crash.TNS analyst Hecht notes the definition of a “crash” is debatable, but suggests it’s when prices decline at least 10% in a matter of days, triggering panic selling that causes another big drop.And economist Krugman is worried about what comes after — about how a bursting bubble could then impact the greater economy. “My guess is that the current tech boom, like the ’90s boom, will end in a painful bust.”So the real question may be: What would remain after that bust was over?“I don’t have a clear view of what the post-bubble landscape will be,” Hecht warns me. But he believes history might offer some hints — like the aftermath of the “frenzied over-investing” between 1999 and 2001, where many promising dotcoms suddenly failed. Hecht saw two major consequences that he’d expect to repeat after an AI-industry crash:“Investors became gun-shy and unwilling to invest in risky businesses they didn’t understand.”“It took a couple of years for those laid off from dotcoms to find jobs that paid as well.”But there’s one more consequence. “Companies with lots of cash on hand, free cash flow and access to credit markets will be able to buy assets for pennies on the dollar. … Vulture capitalism will thrive for the first time in over a decade.”Hecht does acknowledge that there are worries in the world about how a bursting bubble might impact the larger economy, like the doldrums that followed the collapse of the dotcom bubble. But “I’m more worried about other black swan events and geopolitical risks,” Hecht says, “including a tariff war.”So in short, Hecht says, a market crash may be bad for some, but it isn’t an apocalypse for the entire technology sector forever. “When a bubble bursts, investors always get hit hard if they aren’t diversified. … Yet, companies with real customers and market power will recover eventually.“The best-case scenario, and one that seems to be happening now, is that market winners will weather any burst in this bubble and become even stronger as they buy up struggling startups.”The post Is AI a Trillion-Dollar Bubble or a World-Changing Juggernaut? appeared first on The New Stack.