By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel is largely bearish on CoreWeave (CRWV), citing persistent negative EBITDA, massive capex runway, execution risk, customer concentration, and potential insourcing by hyperscalers like Meta and Microsoft. While some panelists (Grok, Gemini) argue that power constraints could create a strategic asset for CoreWeave, others (Claude, ChatGPT) question whether this is a durable moat.
Risk: Potential insourcing by major customers (Meta, Microsoft) and the risk of massive capex becoming a stranded asset.
Opportunity: Potential power constraints creating a strategic asset if CoreWeave controls high-density compute sites that hyperscalers can't replicate quickly.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
CoreWeave (NASDAQ: CRWV) has been an up-and-down investment since the company went public in April 2025. It started as one of the hottest stocks in the market and notched one new high after another. Now it's down around 50% from its peak. The question is, can CoreWeave return to those highs, or is it just a value trap at this time?
I think investors need to start looking at the stock a bit differently than they have before.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
CoreWeave is a neocloud company that provides AI-focused cloud computing resources. Two of its major clients are Meta Platforms (NASDAQ: META) and Microsoft (NASDAQ: MSFT), which use its data centers to bolster their AI computing capacity. One of its biggest backers is Nvidia (NASDAQ: NVDA). Nvidia crosses paths with a lot of companies in the AI world, so its decision to invest in CoreWeave was notable. According to its most recent 13-F form, Nvidia held around $3.6 billion worth of CoreWeave stock as of the end of the first quarter. Relative to CoreWeave's market cap of about a $50 billion, that's no small position.
However, there has been major media coverage recently concerning one of CoreWeave's biggest clients: Meta. Although Meta hasn't confirmed it, reports say that it's forming a cloud infrastructure business of its own -- something that CEO Mark Zuckerberg had previously stated would happen only if the company felt it had excess AI computing power to lease. One way for it to remedy the problem of having excess computing power would be to terminate its contract with CoreWeave, which is why the neocloud stock has sold off so much in recent days. However, with the world still being in a compute-constrained state, having access to excess computing power isn't a bad thing, so it's likely that Meta's contract with CoreWeave will play out as originally expected.
So is CoreWeave still a good investment? From a revenue growth standpoint, it's expected to deliver some incredible results. Wall Street analysts expect 147% top-line growth this year and 98% growth next year. The biggest question mark with CoreWeave doesn't concern revenue; it concerns profit. Wall Street expects the company to lose $ 4.64 per share this year as it spends freely to build as much computing capacity as possible, as rapidly as possible. Those capital expenditures will continue to result in major losses as its infrastructure build-out continues, but the company could flip the switch to profitability if and when it decides that it has captured as much market share as it wants. That probably won't happen for some time, though, so shareholders will need to be OK with the company booking further losses.
This creates some added execution risk for investors, and the market has priced CoreWeave's stock with skepticism. It's trading at 7.1 times sales, which is a pretty cheap valuation. If CoreWeave could snap its fingers and deliver a 30% profit margin, it would be trading for 24 times earnings. If one bases its valuations on next year's revenue projections, the stock is valued at just 1.9 times sales. Using that same hypothetical 30% profit margin, that would price CoreWeave at 6.3 times forward earnings.
So if CoreWeave can maintain its business and turn profitable, it could be a worthwhile investment. However, it may take some years before it turns that corner, and investors should use where the stock trades right now as their anchor point, not the higher prices it saw in the first couple of months after its IPO.
Before you buy stock in CoreWeave, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and CoreWeave wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $397,351! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,304,257!
Now, it’s worth noting Stock Advisor’s total average return is 937% — a market-crushing outperformance compared to 211% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of July 16, 2026. *
Keithen Drury has positions in Meta Platforms, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Meta Platforms, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"At 7.1x sales with years of heavy losses still ahead and rising customer vertical integration risk, CRWV is more value trap than bargain despite impressive topline growth projections."
The article portrays CRWV's 50% drawdown from its April 2025 IPO highs as a potential bargain at 7.1x current-year sales and 1.9x next-year sales, given explosive 147% and 98% expected revenue growth. Yet it glosses over persistent negative EBITDA, massive capex runway (potentially years of losses), and execution risk in a compute-constrained but rapidly commoditizing AI infrastructure market. Nvidia's $3.6B stake (on a ~$50B market cap) and contracts with META/MSFT provide credibility, but Meta's reported in-house cloud push signals customer concentration risk the piece downplays. Valuation assumes a clean flip to 30% margins; history with hyperscaler-adjacent plays shows that rarely happens on schedule.
The strongest case against bearishness is that if AI training demand remains structurally supply-constrained through 2027-28, CoreWeave's early-mover GPU cluster scale and Nvidia partnership could drive actual revenue far above consensus, compressing the path to positive free cash flow and justifying a multiple re-rating to 12-15x sales before profitability.
"CoreWeave’s valuation assumes sustained pricing power in a market where its largest customers are actively incentivized to commoditize the very cloud services they currently rent."
The article frames CoreWeave as a simple 'growth vs. profit' play, but it misses the structural risk of being a middleman in a hardware-commoditized market. At 7.1x trailing sales, the valuation looks attractive only if you ignore the massive depreciation cycle of H100/B200 GPUs. If Meta or Microsoft pivot to internal 'neocloud' infrastructure, CoreWeave loses its primary revenue moat. The 147% revenue growth is irrelevant if the company is merely a pass-through entity for Nvidia’s hardware sales. Investors are essentially betting on a perpetual compute shortage that may evaporate as hyperscalers achieve vertical integration, turning CoreWeave’s massive capex into a stranded asset liability.
If the AI compute shortage remains structural rather than cyclical, CoreWeave’s specialized, low-latency infrastructure could command a permanent premium that hyperscalers cannot replicate without significant operational friction.
"CoreWeave's 147% revenue growth is real, but the path to profitability assumes customer stickiness and margin expansion that are both under active threat from internal competition by its largest clients."
The article frames CRWV as a 'look forward' story, but conflates revenue growth with investability. Yes, 147% top-line growth is impressive—but the company is losing $4.64/share while burning cash on capex, and the article admits profitability is 'some time' away. The valuation math (7.1x sales today, hypothetically 6.3x forward earnings at 30% margins) assumes two massive unknowns: (1) that CoreWeave sustains hypergrowth without margin compression as competition intensifies, and (2) that Meta doesn't actually cannibalize demand by building its own infrastructure. The Meta risk is hand-waved away with 'compute-constrained state'—but that constraint is precisely why Meta is building internal capacity. The article also omits: capex intensity (what % of revenue?), customer concentration risk (how much revenue from Meta + Microsoft?), and whether Nvidia's $3.6B stake signals conviction or hedging.
If Meta and Microsoft—CoreWeave's two largest clients—are both building internal AI infrastructure, CoreWeave faces structural demand destruction that no amount of revenue growth can offset; the 'compute-constrained' argument actually strengthens the case for hyperscalers to internalize rather than outsource.
"CoreWeave's profitability trajectory is highly uncertain and the current valuation already discounts aggressive growth that may never materialize if customers insource or competition compresses margins."
CoreWeave hinges on rapid revenue growth with near-term losses expected (147% growth this year, 98% next) while burning capital to scale AI compute. The article omits near-term burn rate and cash runway risks, plus customer concentration with Meta, Microsoft, and Nvidia backing which could erode if any customer shifts in-house or renegotiates terms. The stock trades at about 7.1x sales—cheap only if growth and durable margins materialize. Yet a crowded hyperscale market, energy/capex costs, and potential insourcing threaten a clear path to profitability, making a multi-year turnaround scenario plausible but far from assured.
The strongest counterpoint is that Meta or Microsoft could aggressively insource or push meaningful price concessions, undermining CoreWeave’s revenue base just as capex remains high; in that case, the ‘years to profitability’ thesis collapses and the 7x sales multiple looks like a trap.
"Nvidia stake creates switching costs and energy access is the unmentioned bottleneck risk."
Claude correctly flags omitted capex intensity and customer concentration, but underweights Nvidia's $3.6B stake as strategic alignment rather than mere hedging. That equity position plus GPU supply priority creates switching costs for META/MSFT that their internal builds won't fully replicate before 2027. The real unmentioned risk is energy availability—CoreWeave's cluster scale depends on securing power contracts amid national grid constraints nobody quantified.
"Grid power availability, not just GPU access, creates a defensive moat that favors CoreWeave's existing infrastructure over hyperscaler internal builds."
Grok, your focus on power constraints is the missing piece. While others debate hyperscaler insourcing, they ignore that building internal data centers is even more power-constrained than renting from CoreWeave. If grid capacity remains the primary bottleneck, CoreWeave’s existing, permitted, and powered-up clusters become a massive strategic asset, not just a commodity. The 'stranded asset' risk Gemini fears is mitigated if they control the scarce resource: power-ready, high-density compute sites that hyperscalers can't replicate quickly.
"Power availability, not compute demand, may be CoreWeave's actual moat—but the article and panel lack hard data on grid constraints, permitted capacity, or hyperscaler power-build timelines to validate this."
Gemini and Grok just pivoted the entire thesis from demand-side risk to supply-side moat. That's material. But neither quantified power availability—CoreWeave's permitted capacity, grid headroom, or whether Meta/Microsoft can solve power constraints faster than compute constraints. If power is truly the bottleneck, CoreWeave's valuation should reflect scarcity premium, not commodity multiples. The burden is now on them to prove power is the binding constraint, not just plausible.
"Exclusive, long-term GPU commitments with Nvidia create a true moat; power constraints alone won't sustain pricing power if hyperscalers expand internal capacity."
Responding to Gemini: I'd push back on the 'power bottleneck makes CoreWeave a moat' thesis. If grid constraints ease, hyperscalers can build near cheap energy centers and bypass CoreWeave with internal capacity. The real moat would be exclusive, long-term GPU commitments with Nvidia's supply. The article's stranded-asset fear ignores that energy cost dynamics alone don't guarantee pricing power or durable margins if supply expands.
The panel is largely bearish on CoreWeave (CRWV), citing persistent negative EBITDA, massive capex runway, execution risk, customer concentration, and potential insourcing by hyperscalers like Meta and Microsoft. While some panelists (Grok, Gemini) argue that power constraints could create a strategic asset for CoreWeave, others (Claude, ChatGPT) question whether this is a durable moat.
Potential power constraints creating a strategic asset if CoreWeave controls high-density compute sites that hyperscalers can't replicate quickly.
Potential insourcing by major customers (Meta, Microsoft) and the risk of massive capex becoming a stranded asset.