Hold on, technology enthusiasts: Goldman Sachs has just thrown a curveball into the ongoing AI-generated saga. In its latest report, the investment giant appears to have shifted its stance from cautious optimism to skepticism about the immediate benefits of AI investments.
A June 2024 report by a global investment bank titled “Gen AI: Too Much Spending, Too Little Benefit?” is a reality check for AI bulls. Despite the tech industry frenzy and a staggering $1 trillion predicted to be spent on AI infrastructure, the report highlights a sobering observation from MIT's Daron Acemoglu and Goldman Sachs' Jim Covello.
Acemoglu predicts that AI will automate less than 5% of tasks over the next decade and contribute just 0.9% to GDP growth. Covello goes further, arguing that current AI technologies are not capable of solving complex problems cost-effectively enough to justify investing at least $1 trillion into them.
“AI bulls seem to believe that use cases will proliferate as the technology evolves,” he said in an interview quoted in the report, “but 18 months after generative AI was introduced into the world, we have yet to see a single truly transformative, let alone cost-effective, application.”
Image: Goldman Sachs
For Jim Covello, the issue bears some similarities to the dot-com bubble (and cryptocurrency for the rest of us). But the impact of future AI winters will depend on the real-world use cases the technology can serve.
“If AI technology has fewer use cases and lower adoption rates than the current consensus expects, [a bust of the AI bubble] “For many of the companies currently investing in this technology, this won't be an issue,” he said.
Covello believes the AI craze could last another 18 months before investors start to lose interest due to a lack of truly meaningful progress.
The issue of the market being in a bubble has been raised before by leading figures in the AI community. Just a year ago, Emad Mostaque, now former CEO of Stability AI, warned that generative AI was an overvalued technology. “I think this will be the biggest bubble of all time,” Mostaque said in a conversation with analysts at UBS. “I'm calling it the 'dot AI' bubble, and it hasn't even started yet.”
But the longer-term outlook remains bullish: US software analyst Rush Langan, quoted in the report, argues that today's AI companies are building the foundations for the next generation of profitable industries.
“We're still in the infrastructure-building phase of the AI cycle, so it's going to take us longer to find the killer application, but I believe we'll get there,” said Langan, who is less bullish on the short term and believes the clock is ticking for AI startups.
“If we don't see any large-scale consumer applications in the next six to 18 months, that could raise even more concerns,” he added.
Profitability is in the eye of the beholder
Contrast this with Goldman Sachs' May 2024 report, which, despite its cautious tone, waxed almost lyrical about AI's potential. Led by Joseph Briggs, the report predicted that generative AI could eventually boost U.S. productivity by 9% and GDP by 6.1% over the course of a decade. The report painted an optimistic picture, with AI laying the foundation for future efficiency gains and economic growth, and highlighted early positives from AI adoption in certain sectors.
“Early signs of future productivity gains are very encouraging,” Briggs wrote. If generative AI proves profitable, it could begin to have a meaningful impact on the economy no later than 2027, the report said.
This also mimics JPMorgan's outlook for the AI industry as a whole, which said, “Despite AI stocks trading at fairly low prices as measured by P/E, Wall Street believes today's AI leaders will achieve higher earnings growth than expected from dot-com leaders.”
Image: JP Morgan
For JPMorgan, AI has the potential to have a major impact on industries like logistics and finance, which the company is calling a transition to “AI 2.0.”
So why the sudden change in tune? A June report pointed to high costs, slow returns, and significant infrastructure challenges that could hinder AI growth.
Private investment in generative AI exceeded $20 billion last year, according to Stanford University, but investment is expected to fall significantly by 2024, with Ernst & Young predicting that around $12 billion will be pumped into AI this year.
Image: Ernst & Young
Not surprisingly, analysts are expecting a decline in AI stocks. So, bubble or not, most forecasters are predicting the AI market is slowing down.
The next few quarters will be crucial in determining whether generative AI can live up to those expectations, or if the industry truly heads into a cooling period, which could leave many of the high-performance GPUs that were once essential for generative AI startups to create cute cat pictures and anime waifus sitting idle while the market recalibrates expectations.
