CoreWeave Inc. (CRWV) stock fell almost 20% in a single session after the company reported disappointing fourth-quarter earnings.
Wall Street was taken by surprise.
I was not.
Weeks ago, I flagged that CoreWeave insiders were selling shares at an alarming pace. The Chief Development Officer unloaded nearly one million shares, while the Chief Strategy Officer and CEO sold another million between them. Many of those sales happened almost immediately after shares were awarded, a very bearish sign.
The news is now catching up to what the insiders already knew: though growth remains impressive, CoreWeave’s underlying economics are deteriorating.
The disappointment came on several fronts. Gross margin collapsed to 67.6%, down from 73% in the prior quarter. Operating margins fell to 6% from 16%, cutting adjusted income by two-thirds. And a quarter of every dollar in sales went straight to interest payments.
Most worryingly, guidance came in far worse than expected. CoreWeave now expects to spend $32.5 billion — the midpoint of its guidance — to generate $12.5 billion in 2026 revenues. That’s $2.60 of capex per dollar of sales – far higher than the roughly $2.07 figure the Street expected.
It’s getting harder for CoreWeave to make money… and the market finally listened.
CoreWeave’s Business Model Has a Fatal Flaw — Just Ask MARA Investors
Back in 2020, I wrote a piece arguing that investors should own bitcoin options rather than shares of MARA Holdings Inc. (MARA), a bitcoin mining company. MARA was producing a commoditized product (bitcoin) where the only real edge was access to specialized ASIC mining rigs. That’s a purchasing relationship, not a moat. And every four or five years, every miner starts over from scratch as they replace every aging processor.
The scoreboard resets. Every time.
That’s why it’s no surprise to me that MARA’s stock has vastly underperformed bitcoin, even though it was supposed to be a bullish leveraged bet on the cryptocurrency. Shares of MARA have risen 180% since I wrote that article, while bitcoin has surged from roughly $11,600 to north of $66,000… a 470% rise. Bitcoin options have done even better.
CoreWeave has the same business problem. At its heart, the firm operates large-scale data centers stocked with Nvidia (NVDA) processors, which it rents to customers on a per-hour or take-or-pay basis (meaning customers pay whether they use the capacity or not) under long-term contracts. Unlike generalist hyperscalers like Amazon Web Services or Google Cloud, CoreWeave makes no attempt to serve general-purpose enterprise workloads. Every architectural decision is optimized for GPU-intensive AI training and inference.
That makes CoreWeave much like MARA Holdings: a capital-intensive, energy-burning company focused on a very specific product. Though CoreWeave will almost certainly succeed thanks to insatiable AI demand, returns will be dragged down by the company’s unending need to replace aging chips, buy expensive electricity, and compete with rivals on price.
There are far better ways to make money from the AI business.
CoreWeave Stock Is a Financing Company, not a Tech Stock
Here’s the framing I keep coming back to: CoreWeave isn’t really a cloud software company. It’s a capital-markets-driven infrastructure lender that happens to operate AI data centers.
Think of it as a rental business, but for AI hardware. CoreWeave borrows money, buys NVIDIA GPUs, leases them to AI companies, and uses that “rental income” to pay back loans. Whatever’s left over goes to reinvestment and shareholder returns.
The financials make this obvious:
- Total Debt — $21.6 billion against $3.3 billion in book equity, 6.5x debt-to-equity ratio
- Free Cash Flow — Negative $7.3 billion in 2025, even as revenue almost tripled
- Capital Expenditures — $32.5 billion budgeted for 2026, nearly the company’s entire market cap a year ago
Real tech companies get cheaper to scale as they grow. CoreWeave does not. Its business requires perpetual, massive capital infusions simply to maintain its position. And that spread between borrowing costs and GPU rental income? It’s narrowing fast.
CoreWeave’s Biggest Risk: Too Much Revenue From Too Few Customers
The market for large-scale AI computing is dominated by a small number of high-value AI organizations — a fact that CoreWeave’s customer list reflects. Its IPO filing disclosed that its two largest buyers (Microsoft Azure and OpenAI) accounted for 77% of total revenue in 2024. Subsequent deal flow suggests that figure remains above 70%.
When your largest customers together generate over two-thirds of your revenue, you don’t set prices. They do.
That’s a problem that MARA Holdings had with bitcoin. And it’s one that CoreWeave is beginning to face. During prepared Q4 earnings call remarks, CEO Michael Intrator admitted that AI computing pricing didn’t rise in 2025, despite accelerating demand. This comes even as prices for other data center components like RAM and hard drives spike. Operating margins are expected to hit just 8% in 2026 (vs. 16% in Q3, 2025).
I’m also skeptical that CoreWeave can build a proprietary software stack to change this equation. No major AI lab will willingly lock itself into a single data center provider if it can avoid it, since it will limit bargaining power down the road.
What Is CRWV Stock Actually Worth?
Let me be clear: CoreWeave is not a fraud, and it is probably not going to zero.
The company’s management has been phenomenal at inking deals with major firms; its $66.8 billion revenue backlog now represents at least three years of strong, growing demand.
AI computing demand will also keep growing. The technology gets better with more processing power, and every new breakthrough seems to pull more demand forward.
Like MARA Holdings in 2020, CoreWeave has found itself in the right business at the perfect time.
But “a great business” and “a great investment” are two different things.
My base-case fundamental value for CRWV stock is roughly $100 per share, close to where it traded before the 20% post-earnings drop. To achieve a suitable return (given the business risk of owning a capital-intensive company), a justified buy price is closer to $65.
I wouldn’t buy shares above that price.
That said, could CRWV trade above $200? Absolutely. Retail-driven euphoria pushed MARA to similarly unsustainable heights during the 2021 crypto mania — and it could easily do the same for CoreWeave.
But that’s speculation, not investment. And I can think of far more reliable ways to find 3x returns than hoping a retail frenzy materializes on schedule.
Better AI Stocks to Buy Instead of CoreWeave
If you’re bullish on AI (and you should be), the question is where to deploy capital for the best risk-adjusted return.
In my August 2025 article, ‘Missed Out on Nvidia?’ I laid out the bull case for two stocks that offer genuine AI upside without the leverage trap and customer concentration risk baked into CRWV — Ondas Holdings (ONDS) and WeRide (WRD).
The former, an AI-powered drone maker, has already risen 150% since that recommendation. I continue to see strong upside as AI technologies mature, though the upside from here is naturally more modest.
But the latter has yet to surge… creating an opportunity for risk-seeking investors looking for large AI-driven returns.
WeRide is China’s largest pure-play robotaxi firm. It’s now building proprietary end-to-end ADAS systems in partnership with Bosch and is quickly expanding robotaxis across the globe. This is the kind of AI investment that can compound over years, not just spike during GPU shortages. You can read more about WeRide here.
The Bottom Line on CoreWeave Stock (CRWV)
CoreWeave caught lightning in a bottle. It pivoted from crypto mining to GPU rentals at precisely the right moment, secured landmark contracts with OpenAI and Microsoft, and rode the AI infrastructure wave to a $187 stock price less than a year after its $40 IPO.
But the 20% single-day collapse after Q4 earnings is the market finally doing the math. Depressed gross margins… near-zero operating income… billions of dollars of debt… insiders selling as fast as shares vest. There’s also Magnetar, one of CoreWeave’s largest investors. The secretive hedge fund has been steadily exiting its position while using options to protect its downside.
As InvestorPlace senior analyst Luke Lango cautioned in a December 2025 investor note, CoreWeave is yesterday’s trade.
True, the AI revolution is still in its early chapters. But the next chapter will be written by companies with proprietary technology, durable software moats, and business models that don’t require perpetual capital infusions to survive. CoreWeave, as currently constructed, is not that company.
Don’t get stuck holding the bag on the AI era’s version of MARA.
Editor’s Note: Forbes calls $1 billion fund manager Louis Navellier “the king of quants.” Today, he’s stepping forward to reveal why he’s investing $358 million of his own firm’s money in the next stage of Artificial Intelligence… a technological sea-change that could erase millions of jobs, solve humanity’s biggest mysteries, and spark a wave of moneymaking opportunities — both in and outside the stock market.
