Related video: What Business Leaders Are Saying About AI TAKEOVER In The Job Market
Concerns about an artificial intelligence (AI) bubble across the U.S. technology sector are simmering ahead of the second-quarter earnings report from Nvidia, to be released after financial markets close on Wednesday.
The maker of processors for data centers and server farms is seen as a bellwether for the AI subsector of the tech industry, which is facing concerns over its long-term profitability after huge investments in the wake of the 2022 release of OpenAI’s ChatGPT chatbot.
Traders are bracing for swings upwards of 6 percent, or $260 billion, in the company’s stock value, according to options pricing data, compared to moves of about 3 percent over the past several quarters.
“The big event in corporate earnings will be Nvidia’s results on Wednesday, which come as tech stocks had seen their biggest five-day pullback since April prior to Friday’s rally,” Deutsche Bank’s Peter Sidorov and others wrote in a Monday note to investors.
The pullback followed comments from OpenAI CEO Sam Altman about how AI might be experiencing a bubble, which is a period of sustained overvaluation that eventually pops when investors get cold feet or don’t get the returns they expected.
“Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes,” Altman told The Verge.
Others in the field of AI and analytics have expressed similar concerns about a possible bubble following the rapid adoption of the new technology.
“It’s had by ten times, fifty times the fastest adoption of any type of computer application ever. I don’t know how much more it could bubble beyond the obvious success that it’s already had,” Jed Dougherty, head of AI architecture at Dataiku, an AI platform company, told The Hill.
But despite AI’s widespread usage and the aggressive marketing push from companies and tech investors, the prospects for profitability are far from a lock.
A recent report from the Massachusetts Institute of Technology (MIT) found that 95 percent of organizations using AI are not making any money off of it despite the widespread promotion of the technology.
“Just 5 percent of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable [profit and loss] impact. This divide does not seem to be driven by model quality or regulation,” the researchers found.
The report described the status of AI’s commercial applicability as “high adoption, low transformation,” noting that only two of eight major economic sectors were showing any “meaningful structural change.”
Media and professional services were the industries in which workflows were substantially affected by AI while healthcare, pharmaceuticals, retail, finance, manufacturing and energy saw little or no change related to the technology.
Dataiku’s Dougherty said that companies are nervous to relinquish control of essential business operations like sales and supply chain management to AI.
“Having chatbots give advice is great, but having them replace workflows is really scary,” he said. “Companies are much more concerned about putting into place truly autonomous systems that are making decisions about their business, even if those decisions are relatively low-level.”
Tech companies famously prioritize user adoption and market share over revenue growth, going into debt on the assumption that they’ll be able to monetize themselves after they establish a customer base.
But with 95 percent of companies not making any money off AI three years after its highly publicized arrival, the trend is nerve wracking for many investors. Facebook, which took five years to initially turn a profit, is considering downgrading its AI unit, the New York Times reported earlier this month.
Total investment in the sector is difficult to estimate, since much of it is private, though investment firm Warden Capital estimated direct spending in AI-related chips and data centers at about $220 billion earlier this year. Whether future profitability can match and exceed that expenditure, or it falls short and exposes a bubble, remains to be seen.
“It’s kind of like the dot-com [era],” Jason Bishara, financial practice leader with NSI Insurance Group, told The Hill, speaking of general market conditions responding to surges in the AI and cryptocurrency sectors. “When the dot-coms collapsed and the vast majority of them disappeared, the good ones rose to the top and a lot of people made fortunes on them,” he said.
What’s been lost in the marketing, lobbying and deregulatory push surrounding AI, tech workers say, is the fact that “AI” is actually a catch-all phrase that describes many different types of software products, some of which may be more commercially viable than others.
Large language models based on machine learning can be structurally very different from the facial and voice recognition software based on neural networks, which are a type of filtering algorithm. Some of these formulas are better thought of as developments of older analytics software than a fundamentally new technology.
“There’s really a spectrum,” Dougherty said.
Even in the cases where it’s commercially viable, AI could — and has — run afoul of the law, further limiting its use and making companies vulnerable to legal action.
AI has exhibited racial bias when it comes to offering mortgage loans to aspiring homeowners, and it’s been used in an algorithmic price-fixing scheme, according to the Justice Department. There are all sorts of other potential legal violations that the software could unwittingly commit as it seeks to optimize commercial returns.
Another sign of potential stress in the tech sector came in the form of an unexpected 10-percent, $10 billion government stake in chipmaker Intel, which the Trump administration announced last week.
Intel warned of “adverse reactions” from shareholders as a result of the government stake.
The Company may experience other adverse consequences resulting from the announcement or completion of the transactions,” the company wrote in an SEC filing. “Given the scarcity of recent US precedents … it is difficult to foresee all the potential consequences.”
