Dan Ives: Bears are ‘underestimating scale and scope’ of the AI revolution
As valuations for AI companies continue to skyrocket — including Nvidia (NVDA) briefly becoming the first ever $5 trillion company — a debate is swirling about whether the multiples for AI companies still make sense, or whether they’re pointing to the emergence of a bubble for the sector.
The question of whether the AI industry has entered a bubble was even recently posed to Federal Reserve Chairman Jerome Powell, who drew a distinction between the dot-com bubble of the 1990s and what is happening now.
Powell pointed to the fact that AI companies today “that are so highly valued actually have earnings,” whereas the internet companies during the dot-com era “were ideas more than they were companies.”
Dan Ives, managing director and global head of technology research for Wedbush Securities, is squarely on the side of AI bulls, and as one of Wall Street’s most prominent tech analysts, investors tend to pay attention to his views on the industry.
In a post on X on Wednesday, he took the skeptics to task for what he sees as a shortsighted view of the long-term impact that AI companies today will be making.
“Bears have NEVER understood this tech AI driven bull market since early 2023 and all kicked off with Nvidia May 2023 blow out quarter and MSFT/OpenAI,” he said, referring to when Microsoft ramped up its investments in OpenAI in 2023. “The Street is underestimating scale and scope of this 4th Industrial Revolution. We own AI winners and buy tech on sell-off.”
One of the companies that has become central to the debate over a potential bubble is software company Palantir (PLTR), which is drawing increased skepticism from analysts over how much its valuation dwarfs that of other companies in the software sector, with a market cap of just over $448 billion.
Even investors have perhaps gotten skeptical of its value, which is why Palantir’s stock plunged nearly 8% on Tuesday despite a blowout quarterly earnings this week.
But Ives said on Monday that Wedbush was raising its price target on the company even before its earnings, noting “we believe earnings/guidance will be another major step in the right direction as Karp & Co. build out this AI juggernaut,” referring to Palantir co-founder and CEO Alex Karp.
“While investors will continue to fret about valuation we believe this will be a trillion dollar market cap (in the) next 2 years,” Ives added.
CNBC commentator Jim Cramer seems to agree with Ives, but recognizes that there are likely many investors who don’t.
“Yes, a lot of soul-searching about Karp,” Cramer said on X. “I think his company's results speak for itself but it would not surprise me if there were a ring of short sellers betting against the company.”
Doubts raised about how AI will be monetized
However, Peter Berezin, chief global strategist and director of research at BCA Research, questioned on Wednesday the notion of there being “massive excess demand for AI compute,” which is one of the arguments made to justify how valuable AI companies are now.
“Yes, that’s true, just like there would be massive excess demand for cars and homes if cars and homes were free,” Berezin said.
He also compared AI companies to social media companies, saying that the reason the latter have become so profitable is because of “network effects,” which means that people spend time on these platforms because other people spend time on them as well.
“People use Instagram because that’s what others use,” Berezin said. “This does not apply to AI. When you use AI you interact with the AI, not with other users of AI. AI will likely become a commoditized product with commoditized margins.”
McKinsey & Co. is estimating that it will cost about $5.2 trillion by 2030 to have enough compute power to handle AI processing loads, which has raised questions about how AI companies will be able to turn a profit when so much money is being spent on their data centers.
“There just isn’t enough revenue and there never can be enough revenue,” Praetorian Capital CIO Harris Kupperman wrote earlier this year. “The world just doesn’t have the ability to pay for this much AI.”
A user on X suggested to Berezin that the costs of compute will eventually come down, which would presumably make it easier for companies to turn a profit.
“That is no doubt true, just like the cost of broadband has come down rapidly over time,” Berezin said. “But last I checked, companies like Verizon trade at less than 10 times earnings.”
Nvidia (NVDA) is currently trading at nearly 58 times its earnings.
But when Nvidia founder and CEO Jensen Huang was asked on Wednesday whether he thought the AI sector was heading toward a “Big Short”-style collapse, he insisted the industry was “a long, long way from that.”
"I would say that we're in the beginning of a very long build out of artificial intelligence," he said.