“Despite its expensive price tag, the technology is nowhere near where it needs to be in order to be useful,” Jim Covello, Goldman Sachs’s most senior stock analyst and a 30-year veteran of covering tech companies, said in a recent report about AI. “Overbuilding things the world doesn’t have use for, or is not ready for, typically ends badly.”
Covello’s comments are in sharp contrast to a different Goldman Sachs report from just over a year ago, in which some of the bank’s economists said AI could automate 300 million jobs around the world and increase global economic output by 7 percent in the next 10 years, spurring a spate of news coverage about the disruptive potential of AI.
Barclays said Wall Street analysts are expecting Big Tech companies to spend around $60 billion a year on developing AI models by 2026, but reap only around $20 billion a year in revenue from AI by that point. That kind of investment would be enough to power 12,000 products of a similar size to OpenAI’s ChatGPT, Barclays analysts wrote in a recent report.
OpenAI released ChatGPT in November 2022, kicking off a race in Silicon Valley to build new AI products and get people to use them. Big Tech companies are spending tens of billions of dollars on the technology. Retail investors have bid up the price of those companies and their suppliers, especially Nvidia, which makes the computer chips used to train AI models. Year to date, shares of Google parent Alphabet are up 25 percent, Microsoft is up 15 percent, and Nvidia shares are up 140 percent.
Venture capitalists have also poured billions more into thousands of AI start-ups. The AI boom has helped contribute to the $55.6 billion that venture investors put into U.S. start-ups in the second quarter of 2024, the highest amount in a single quarter in two years, according to venture capital data firm PitchBook.
Tech executives insist that AI will change whole swaths of modern life, in the same way the internet or mobile phones did. AI technology has indeed improved drastically and is already being used to translate documents, write emails and help programmers code. But concern over whether the tech industry will be able to recoup the billions of dollars it’s investing in AI anytime soon — or ever — has risen among some firms that only last year were heralding the boom.
“We do expect lots of new services … but probably not 12,000 of them,” Barclays analysts wrote. “We sense that Wall Street is growing increasingly skeptical.”
In April, Meta, Google and Nvidia all signaled their commitment to going all in on AI by telling investors during quarterly earnings calls that they would ramp up the amount of money they’re spending on building data centers to train and run AI algorithms. Google reiterated Tuesday it would spend more than $12 billion a quarter on its AI build-out. Microsoft and Meta are due to report their own earnings next week and may give further indication about their AI road maps.
Pichai said Tuesday that it would take time for AI products to mature and become more useful. He acknowledged the high cost of AI but said even if the AI boom slows down, the data centers and computer chips the company was buying could be put to other uses.
“The risk of underinvesting is dramatically greater than the risk of overinvesting for us,” Pichai said. “Not investing to be at the front here has much more significant downsides.”
A spokesperson for Microsoft declined to comment. A spokesperson for Meta did not respond to a request for comment.
Unrealistic expectations
Vinod Khosla, who co-founded computer network systems company Sun Microsystems and is one of Silicon Valley’s most influential venture-capital investors, compared AI to personal computers, the internet and mobile phones in terms of how much it would affect society.
“These are all fundamentally new platforms. In each of these, every new platform causes a massive explosion in applications,” Khosla said. The rush into AI might cause a financial bubble where investors lose money, but that doesn’t mean the underlying technology won’t continue to grow and become more important, he said.
“There was a dot-com bubble, according to Goldman Sachs, because prices went up and prices went down. According to me, internet traffic didn’t go down at all.”
As AI changes the way people work, do business and interact with one another, many start-ups will fail, he said. But overall the industry will make money on AI. He predicts there will eventually be multiple trillion-dollar businesses in AI, such as humanoid robots, AI assistants and programs that can completely replicate the work of highly paid software engineers.
But so far, AI is not contributing to an increase in venture capital getting a return on those investments. The amount of money made in venture capital exits, which represent initial public offerings or acquisitions of tech start-ups, fell to $23.6 billion in the second quarter, down slightly from $25.4 billion the previous quarter, according to PitchBook.
The tech industry would need to generate around $600 billion in revenue a year to make up for all the money being invested in AI right now, yet it is far from close to that number, David Cahn, a partner at venture firm Sequoia Capital, wrote in a blog post last month.
“Speculative frenzies are part of technology, and so they are not something to be afraid of,” Cahn said. “But we need to make sure not to believe in the delusion that has now spread from Silicon Valley to the rest of the country, and indeed the world. That delusion says that we’re all going to get rich quick.”
Microsoft’s and Google’s revenue are growing, especially in their cloud businesses where they sell access to AI algorithms and the storage space to use them. Executives from the companies say AI is driving new interest in their products and will become a major moneymaker in the future. But some analysts are pointing out that there have been very few hugely successful stand-alone products, besides OpenAI’s ChatGPT and Microsoft’s coding assistant GitHub Copilot.
“Wall Street is growing increasingly skeptical given that ChatGPT and GitHub Copilot are the two breakout successes in consumer and enterprise thus far 20 months in,” the Barclays analysts wrote in their report.
The cost of developing and running AI programs will come down as other companies compete with Nvidia and the technology becomes more efficient, said Vineet Jain, CEO of Egnyte, an AI and data management company. For now, the cost of providing AI products is too expensive, and he doesn’t expect to make any AI-specific revenue this year. But as costs go down and demand continues to rise, that will change, Jain said.
“The value proposition is absolutely there, but the expectation right now is still unrealistic,” he said, referring to the frenzy to sell AI products to consumers and businesses.
Some start-ups have already come down from the heights of the early part of the AI boom. Inflection AI, a start-up founded by veterans of Google’s famous DeepMind AI lab, raised $1.3 billion last year to build out their chatbot business. But in March, the company’s founders left for jobs at Microsoft, taking some of their top employees with them to the tech giant. Other AI companies, like Stability AI, which was one of the first companies to build a widely popular AI image generator, have had to lay off workers. The industry is also facing lawsuits and regulatory challenges.
Bigger companies like Google and Microsoft will be able to keep spending money until demand for AI products increases, but smaller start-ups that have taken on a lot of venture capital might not survive the transition, Jain said.
“It’s like a soufflé that keeps popping up and popping up, it has to come down a bit.”
