MW The blogger who was right about Nvidia is now panning a $25 billion AI company

By Steve Gelsi

Jeffrey Emanuel, known for turning Wall Street's attention to the risk DeepSeek might pose to Nvidia, warns that CoreWeave's IPO is way overvalued

Jeffrey Emanuel was a little-known crypto executive and blogger when he sat down to pen a cautionary post on Nvidia Corp. earlier this year. His warning about the cost efficiencies exhibited by Chinese artificial-intelligence startup DeepSeek helped shave $600 billion off Nvidia's market capitalization in a single day.

When Emanuel wrote that post, it caught the attention of prominent thinkers from Wall Street to Silicon Valley. Now he's hoping they listen again - this time to his warnings on New Jersey-based CoreWeave Inc. (CRWV), the cloud-services provider expected to go public in the coming weeks.

CoreWeave saw its revenue balloon 747% in 2024 by running data centers for artificial-intelligence companies.

While Emanuel remains bullish on AI overall, he said CoreWeave is far from being the booming industry's next hot investment.

Instead, he said CoreWeave more closely resembles to WeWork - the office-sharing pioneer that went from a $47 billion valuation to bust in four years after it had loaded up on debt to pay for its real estate while demand for office space dropped.

Emanuel cautioned that CoreWeave similarly could face a steep drop-off in demand if companies cut back on the massive budgets they're currently dedicating to AI chips. That's partly because CoreWeave doesn't own its data centers - it leases them and then curates the technology for its AI clients.

If revenue doesn't keep up with its data-center lease payments, CoreWeave could face a cash crunch.

"Just as WeWork ran into debt payment problems when it rented long-term leases and lost its short-term lease holders during COVID, CoreWeave's business may not be able to continue to grow its revenue as rapidly," Emanuel said.

On Thursday, CoreWeave set an estimated price range of $47 to $55 a share for its initial public offering, and an offering size of 49 million shares. Based on the midpoint of the estimated range, CoreWeave's IPO would raise $2.5 billion for the company.

That would make its IPO the fourth since 2022 to raise at least $2.5 billion, after Lineage Inc. (LINE), Arm Holdings PLC $(ARM)$ and Kenvue Inc. (KVUE), according to Renaissance Capital.

With a total of 480.83 million Class A and Class B shares expected to be outstanding after the IPO, CoreWeave would have a market cap of about $25 billion based on the midpoint of the estimated pricing range.

That estimated market cap for CoreWeave comes in about $10 billion less than many were expecting, Emanuel noted.

But even at that reduced level, CoreWeave would be still be way overvalued, Emanuel added, saying he wouldn't want to invest in it for a long-term holding unless it was valued at $3 billion or less.

To be sure, his view on CoreWeave remains contrarian to many on Wall Street who are more receptive to the deal.

Renaissance Capital analyst Matthew Kennedy said CoreWeave will rank as the largest pure-play IPO from an AI company ever.

"CoreWeave is part of a new wave of IPOs tapping into the explosive demand for AI computing power," Kennedy said in an email to MarketWatch. "The vast majority of pure-play AI companies are still private, so I expect many investors will take note."

While CoreWeave has a lot of debt and capital-expenditure needs and its industry will grow more competitive, "for next three years at least they're in exactly the right place to benefit from the explosive growth in AI," Kennedy said.

Emanuel told MarketWatch he doesn't own a positive or negative position in CoreWeave, which filed to go public on March 3. And he does not work for an investment bank looking to win advisory business from CoreWeave.

Beyond working as an executive at blockchain platform Pastel Network, Emanuel also charges consulting fees to hedge funds and other clients interested in hearing his insights on technology and AI. He said he's "particularly eager to help less-informed retail investors" from "being invested in what I view as an egregiously overvalued 'pig in a poke.'"

Since Microsoft Corp. $(MSFT)$ currently accounts for 62% of the database company's sales, CoreWeave faces big revenue risks due to higher customer concentration, Emanuel noted.

That concern around customer concentration was echoed by D.A. Davidson analyst Alexander Pratt.

"We would flag this as a concern that customer diversification is decreasing, not growing, with the company expecting to garner a significant percentage of revenue from limited customers for the foreseeable future," Pratt said.

Microsoft is using CoreWeave to handle overflow from its own cloud-computing service, Azure. The software giant also plans to invest $80 billion in 2025 to build its own AI-focused data centers, Pratt said.

Looking ahead, "either $(A)$ Azure infrastructure expands to meet demand levels - negative for CoreWeave - or (B) demand as a whole rolls over, in which Azure won't need overflow - again, negative for CoreWeave," Pratt noted.

While CoreWeave claims to have proprietary technology and software around managing AI servers, it hasn't revealed anything "unique or particularly impressive/difficult" about it, according to Emanuel.

"There are plenty of open-source alternatives to any software like that, which they do have," Emanuel said. "There's no enduring business franchise value there - no real competitive moat from better technology, unlike with Nvidia $(NVDA)$, which at least does have the technology edge at the moment."

While CoreWeave recently unveiled a contract with OpenAI valued at up to $11.9 billion by 2030, Emanuel said the announcement lacks details such as the minimum cash that OpenAI will be required to spend each year regardless of usage, as well as what rights OpenAI has to get out of the contract and any penalties it faces if it backs out.

"Without these details, I would be just as skeptical about this deal as I think you should be about the headline number of '$500 billion' for the Stargate project between OpenAI and SoftBank," he said.

CoreWeave didn't reply to MarketWatch's request for comment on Emanuel's thesis. It's currently in a quiet period required after a company files to go public.

While CoreWeave saw its revenue grow to $1.9 billion last year from just $229 million the year before, it also posted a larger loss of $863 million in 2024, compared with a loss of $594 million the year before.

CoreWeave said in its prospectus that its cloud platform "consists of our proprietary software and cloud services that deliver the software and software intelligence needed to manage complex AI infrastructure at scale.

"Our platform supports the development and use of groundbreaking models and the delivery of the next generation of AI applications that are changing the way we live and work across the globe," it added.

CoreWeave's 242-page prospectus also contains about 55 pages of risk factors. These risks include how "recent growth may not be indicative of our future growth," and how "a substantial portion of our revenue is driven by a limited number of our customers."

The filing also cites the prospect that CoreWeave's offerings become less competitive "if we fail to efficiently enhance our platform and develop and sell new solutions and services and responses effectively to rapidly changing technology."

The company said it offers benefits around overall performance, time to market and reduced cost of ownership. Its customers "make large, long-term initial commitments" and grow those commitments to CoreWeave over time, per the filing.

CoreWeave reported $15.1 billion of remaining performance obligations as of Dec. 31, up 53% from $9.9 billion in the year-ago period. The metric tracks contracted business that has yet to be recognized as revenue.

Emanuel said he's been getting queries from his clients about CoreWeave's true value for several weeks. CoreWeave fetched a $24 billion valuation on the private market back in November, and Emanuel said was "flabbergasted" at that number.

"This IPO is a cynical attempt to capitalize on the AI hype and to foist this company on the uninformed investing public at an unsupportable valuation," he said.

That gets back to the WeWork comparison. WeWork was valued at as much as $47 billion by the private market, but its valuation fell to $9 billion when it went public in 2021 through a merger with a blank-check company. The company filed for bankruptcy in 2023.

-Steve Gelsi

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

March 20, 2025 13:24 ET (17:24 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.