Is there an AI bubble and has it started to burst?

The artificial intelligence boom appears unstoppable, but a growing number of investors and other observers fear it is a bubble about to burst.

After climbing more than 50% from April lows, the tech-heavy Nasdaq Composite Index is down nearly 5% this month. Investors fear it will take longer than expected to realize big profits from the billions of dollars they have invested in the next wave of technology.

Those who have been around long enough say some of this enthusiasm resembles the dot-com boom and bust of the turn of the century. Optimists say this time it’s different.

AI chipmaker Nvidia led the stock rally, becoming the world’s most valuable company thanks to enthusiasm for AI. The Santa Clara-based company makes specialized computer chips that technology companies use for training AI models, data centers, robotics and more.

AI bulls and bears have been waiting to hear what Nvidia had to say about the state of the company in its quarterly earnings report released Wednesday. The company helped to maintain this hope by announcing that its profits and future prospects were better than analysts expected. Its shares jumped more than 4% in early after-hours trading.

“There’s been a lot of talk about an AI bubble. From our perspective, we’re seeing something very different,” Nvidia Chief Executive Officer Jensen Huang said in a call following the results. “As a reminder, Nvidia is unlike any other accelerator. We excel in every phase of AI.”

From social media to self-driving cars, Huang highlighted how AI, capable of generating content and performing tasks without human intervention, will affect every sector.

Nvidia’s results could help reignite the AI ​​rally. However, investors and analysts still worry about whether high stock prices are justified for all the different companies in the AI ​​race. After the dotcom bubble, many companies disappeared, but those that survived are now among the largest and most profitable companies in the world.

Sky-high valuations for the Silicon Valley tech giant and other big players in the AI ​​race have investors wondering if and when their bet on the future of technology will pay off. The fates of tech companies have become increasingly intertwined as they invest hundreds of billions of dollars in each other, as well as in data centers, AI research and lucrative employee compensation packages.

In September, Nvidia announced plans to invest up to $100 billion in OpenAI, the maker of ChatGPT, to fund the massive construction of data centers housing the equipment used to store and process the trove of information needed for AI systems. OpenAI has also committed to purchasing Nvidia chips for at least 10 gigawatts of AI data centers.

Total capital spending needed to support OpenAI’s computing needs could reach $130 billion by 2027, according to an October analyst note from New Street Research. That means OpenAI alone could spend $52 billion on Nvidia technology.

Despite its sky-high valuation of around $500 billion, OpenAI is losing billions of dollars as it continues to spend on infrastructure, computing power and other expenses.

“Whether we spend $500 million a year, $5 billion a year, $50 billion a year, I don’t care. I really don’t care,” Sam Altman, CEO of OpenAI, said at a conference at Stanford. talk last year. “It’s going to be expensive. It’s definitely worth it.”

But as losses pile up, investor anxiety increases.

About 45% of global fund managers surveyed by Bank of America said there was an “AI bubble” that could negatively impact the economy and markets.

The debate over whether there is an AI bubble will continue.

Samuel Hammond, chief economist at the Foundation for American Innovation, said he doesn’t think AI investments are happening in a bubble. But there could be winners and losers.

“Companies that get a massive valuation just for integrating AI into their mission statement but fail to achieve them could still go to zero,” he said. “But most of the stock market’s growth is driven by large-cap tech stocks like Nvidia and Google.”

Tech companies are largely financing these massive data center projects with equity — not debt — which reduces the likelihood that it’s a bubble about to burst, Hammond said.

Strategists at investment bank Goldman Sachs said in an October article that while there is a risk that technology companies will overinvest, they have generated profit growth and have strong balance sheets.

“Even if the success of dominant technology companies is evident, this does not necessarily mean that there is a bubble in the market that is likely to burst imminently,” the paper said.

Gary Smith, an economics professor at Pomona College and author, has warned of an AI bubble. He highlighted OpenAI’s losses, circular financing among tech companies and the limitations of AI.

“OpenAI is certainly in a very fragile situation,” he wrote in an opinion piece with retired author and professor Jeffrey Funk on MarketWatch. “When the AI ​​bubble bursts, this will be one of the first casualties.”

Some analysts have compared data center construction to the telecommunications boom of the 1990s, when companies invested 500 billion dollars to lay fiber optic cables to accommodate rapid Internet adoption, creating a glut of dark fiber that sat unused for years as physical capacity exceeded actual traffic.

Google CEO Sundar Pichai told the BBC that there were times when the tech industry “overshot”.

“We can look at the internet right now. There’s clearly been a lot of excess investment,” he said. “But none of us would question the depth of the Internet.”