- Tech companies plan to spend over $1 trillion on artificial intelligence.
- But the payback period could be long and disappointing, Goldman Sachs said.
- Some experts say the AI may not be good enough to justify its prohibitive costs.
Tech companies are investing heavily in the AI boom, but it will be some time before we see results.
As companies prepare to spend more than $1 trillion on artificial intelligence, a new Goldman Sachs report examines a big question: will this massive spending pay off?
This huge investment will go towards the data centers, power grids, and AI chips needed to run AI, but a shortage of these AI components could lead to disappointing returns for companies.
“AI technologies are very expensive, and to justify their cost, they must be able to solve complex problems, which is what AI was not designed for,” Jim Covello, head of global equity research at Goldman Sachs, said in the report.
“The starting point of costs is also so high that even if costs were to come down, they would need to come down substantially for AI to automate tasks to be affordable,” he added. “In our experience, even basic summarization tasks often result in indecipherable and gibberish.”
He's not wrong: Google scaled back its use of AI in search after bots started making strange suggestions, like telling a Business Insider reporter to put glue on pizza to stop the cheese from sticking.
The tech industry is also “too complacent in assuming that the costs of AI will come down significantly over time,” Covello said, especially since that assumption appears to be dependent on a race to unseat Nvidia's dominance in AI chips.
Other experts cited by Goldman Sachs were more enthusiastic.
“AI technologies are undoubtedly expensive today, and the human brain is 10,000 times more efficient per unit of power than generative AI at performing cognitive tasks,” said Kash Langan, senior equity research analyst at Goldman Sachs. “But the cost equation for technology will change, as it always has.”
Eric Sheridan, another senior equity research analyst at the firm, compared it to the initial lukewarm reactions to technology developments like the iPhone and Uber.
“Before smartphones, Uber and Airbnb, people didn't think they needed them, but today it's inconceivable that people would resist such technological advances, and that will almost certainly be true of generative AI technologies,” Sheridan said.
