By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel generally agrees that neocloud providers face significant challenges, including high capital expenditure, potential margin compression, and debt servicing risks. However, they differ on the extent to which these challenges are insurmountable.
Risk: Liquidity crunch due to debt servicing and potential asset sales or strategic pivots before margins improve.
Opportunity: Improved utilization and pricing power could move neoclouds toward meaningful margins and potential monetization beyond pure compute leasing.
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 →
Strategists at BCA Research recommend short selling a basket of neocloud stocks, including CoreWeave, while taking a long position in four hyperscalers, including Alphabet. - MarketWatch photo illustration/iStockphoto
Neocloud companies are providers of computing power that are benefiting from massive investment by hyperscalers as the artificial-intelligence infrastructure build-out continues. But Noah Weisberger, chief strategist at BCA Research, believes the neoclouds may "destroy capital" as they invest borrowed cash to grow what will turn out to be a low-margin commodity service.
In a recent research report, Weisberger and colleague Dishaan Pandey presented data indicating that six neocloud companies' heavy capital expenditures would lead to much lower returns on invested capital than the hyperscalers would derive from their own capex.
Neocloud stocks have been volatile. For example, shares of CoreWeave CRWV were up 39% for the first half of 2026 but fell 28% in July. The stock is now up 27% for the year.
"The neoclouds benefit from spending higher up the food chain," Weisberger told MarketWatch, "but they are spending a lot of capital to get those revenues in the door."
He also said that although there are "legitimate questions about the pace of capex" for hyperscalers, this group — which includes Microsoft MSFT, Alphabet GOOGL, Meta Platforms META and Amazon AMZN — is earning well above its cost of capital, while the neoclouds are showing negative returns on invested capital, on average.
BCA is recommending a two-legged trade, with 80% being a long position in those four hyperscalers, weighted by market capitalization. The remaining 20% would be the short sale of a basket of six neocloud companies, also weighted by market capitalization. The six companies for the short trade recommended by BCA are CoreWeave CRWV, Nebius Group NBIS, Applied Digital APLD, Iren IREN, TeraWulf WULF and Cipher Digital CIFR.
"Since hyperscalers have a beta of approximately 1.2 to the S&P 500 and neoclouds are much more volatile, with a beta of 2.9 (calculated over the last 2 years), we are structuring our trade to target a net beta of 2 to the S&P 500," the analysts wrote in the report.
Weisberger acknowledged that the neoclouds are "still in the ramp-up phase," but said they would eventually compete in a generic space for commodity cloud-computing services, and have relatively low valuations to revenue and earnings as a result.
"In the worst-case scenario they destroy capital," he said, while in the best case, the neoclouds will be "cyclical leasing companies."
Here's a chart from BCA showing how high capital spending is projected to be in 2026, relative to revenue, for 13 technology companies, based on consensus estimates among analysts polled by FactSet. The six neoclouds are on the left:
The six neocloud companies at the left of this chart have been "paying a lot for every dollar of revenue they are bringing in," according to Noah Weisberger, chief strategist at BCA Research. - BAC Research
Despite the BCA team's bleak outlook for the neocloud companies' profit margins and stock valuations over the long term and its short-sale recommendation, analysts for brokerage firms see plenty of room to ride for this group over the next year.
Analysts polled by FactSet have majority buy or equivalent ratings on all 10 stocks mentioned in this article, along with the additional three (Oracle, Equinix and Digital Realty Trust) shown on the chart above.
With the 13 companies listed in the same order as on the chart, here is a summary of 12-month price targets, along with forward price/sales ratios.
| | | | | | | --- | --- | --- | --- | --- | | Company | Forward price/sales | Aug. 19 price | Consensus price target | Implied 12-month upside potential | | Cipher Digital CIFR | 11.2 | $15.89 | $32.17 | 102% | | TeraWulf WULF | 11.8 | $15.50 | $36.61 | 136% | | Iren IREN | 4.8 | $42.84 | $82.80 | 93% | | Nebius Group NBIS | 6.9 | $223.90 | $288.88 | 29% | | Applied Digital APLD | 7.5 | $28.23 | $74.21 | 163% | | CoreWeave CRWV | 2.3 | $90.87 | $144.12 | 59% | | Oracle ORCL | 4.2 | $143.81 | $249.45 | 73% | | Equinix EQIX | 9.7 | $1,077.08 | $1,235.07 | 15% | | Meta Platforms META | 4.9 | $546.03 | $745.98 | 37% | | Digital Realty Trust DLR | 9.4 | $192.35 | $223.55 | 16% | | Microsoft MSFT | 8.9 | $484.31 | $563.92 | 16% | | Alphabet GOOGL | 7.4 | $344.72 | $426.68 | 24% | | Amazon.com AMZN | 3.2 | $265.84 | $329.63 | 24% | | Source: FactSet | | | | |
The table shows forward price/sales ratios because the consensus among analysts is that the neocloud companies will show net losses for the next 12 reported months combined. The forward price/sales ratios compare with weighted valuations of 7.4 for the S&P 500 information technology sector and 3.2 for the full S&P 500 SPX.
Click on the tickers for more about each company, exchange-traded fund or index.
Read: <a href="https://www.marketwatch.com/story/these-are-the-most-important-things-to-check-on-a-stocks-quote-page-before-deciding-whether-to-buy-or-sell-11630783155">Tomi Kilgore's detailed guide to the information available on the MarketWatch quote page</a>
Don't miss: <a href="https://www.marketwatch.com/story/9-stocks-to-ride-the-industrial-sectors-blazing-rally-8f23092f?mod=article_inline">9 stocks to ride the industrial sector's blazing rally</a>
Four leading AI models discuss this article
"The article’s ROIC-focused bear case overemphasizes capital destruction; with durable AI demand, neoclouds could achieve meaningful margin uplift and re-rate, making the current short thesis too simplistic."
The MarketWatch piece highlights a BCA Research view that neoclouds burn capital chasing growth while hyperscalers enjoy returns above their cost of capital. That dichotomy is informative but incomplete: ramping capex is a feature of a late-cycle AI build-out, not a permanent moat destroyer. If utilization and pricing power improve, neoclouds can move toward meaningful margins and potential monetization beyond pure compute leasing. Valuations look stretched on net losses, but growth expectations and funding conditions could keep a few names solvent and even cash-flow positive in time. The risk is leverage and energy/commercial-environment costs; a sector-wide crash in AI demand would hit hard, but a milder, durable demand trajectory could surprise to the upside.
The strongest counter is that AI demand could sustain higher utilization and lead to improving margins for neoclouds, potentially allowing several to reach positive ROIC even before hyper-scale consolidation occurs.
"Neoclouds are essentially levered leasing companies that will face severe margin erosion as AI compute shifts from a supply-constrained luxury to a commoditized utility."
The BCA Research thesis correctly identifies the 'commodity trap' for neocloud providers like CoreWeave and IREN. Their business model relies on massive, debt-fueled GPU procurement, effectively functioning as high-beta derivatives on Nvidia's roadmap. While hyperscalers (MSFT, GOOGL) have diversified revenue streams—cloud, search, ads—that allow for high ROIC, neoclouds are pure-play infrastructure plays with zero moat once H100/B200 supply chains normalize. If the AI capex cycle cools, these companies face a brutal 'liquidity crunch' as they struggle to service debt on depreciating hardware. The 100%+ upside targets from sell-side analysts look like classic 'hope-based' modeling that ignores the inevitable margin compression of utility-like compute services.
The bull case for neoclouds hinges on the 'specialized cloud' narrative: hyperscalers are too bureaucratic to optimize for specific AI workloads, allowing agile neoclouds to capture high-margin, niche enterprise demand that the giants ignore.
"Neoclouds' negative ROIC is real, but the article conflates cyclical ramp-phase losses with permanent capital destruction, and ignores that specialized workload stickiness could sustain margins hyperscalers can't undercut."
BCA's short thesis rests on a clean story—neoclouds burn capital at negative ROIC while hyperscalers earn above cost of capital. The data on capex-to-revenue ratios is real and damning. But the article omits critical context: (1) neocloud margins may improve dramatically as utilization rises and they move past ramp phase; (2) the 'commodity' assumption ignores specialized GPU/inference workloads where neoclouds have structural advantages hyperscalers can't easily replicate; (3) consensus price targets embed 59-163% upside for neoclouds over 12 months, suggesting the market hasn't priced in the BCA thesis despite public data. The trade's 80/20 long hyperscaler/short neocloud structure also assumes hyperscaler capex concerns are overblown—a bet that may not hold if AI ROI disappoints.
If neoclouds achieve even 40-50% utilization rates within 18 months and pricing holds, their unit economics flip sharply positive, making the negative ROIC snapshot a timing artifact rather than structural doom. Hyperscalers' own capex efficiency could deteriorate faster than BCA models if AI infrastructure becomes oversupplied.
"N/A"
[Unavailable]
"Refinancing risk and credit stress could derail the neocloud upside even if utilization improves"
Gemini understates refinancing risk. Even with utilization gains, debt-laden neoclouds face tighter credit conditions and covenant stress if AI demand softens or capex cycles elongate. The 100%+ upside sell-side targets assume benign financing and stable pricing; a material liquidity crunch could force asset sales or strategic pivots before margins improve. Until debt maturity ladders, interest coverage and vendor credit terms are tested, the 'specialized cloud' narrative may stall.
"Neoclouds face an insurmountable switching-cost moat from hyperscalers that utilization rates alone cannot overcome."
Claude, you’re missing the 'vendor lock-in' reality. Hyperscalers like MSFT and GOOGL aren't just selling compute; they are selling the entire software stack—Kubernetes, data lakes, and security—that makes switching to a neocloud a massive operational headache for enterprises. Even if neoclouds hit 50% utilization, they remain 'price-takers' in a race to the bottom against incumbents who can subsidize compute losses with high-margin SaaS revenue. The neocloud model is structurally fragile regardless of utilization targets.
"Vendor lock-in matters only if customers value it more than price; the neocloud bet hinges on whether they don't."
Gemini conflates two separate problems: vendor lock-in (real, but applies to all cloud) and structural margin compression (the actual neocloud risk). Hyperscalers' SaaS cross-sell is powerful, but it doesn't prove neoclouds can't capture price-insensitive, latency-sensitive workloads willing to accept narrower ecosystems. The real test: do enterprise AI teams actually value integrated stacks enough to pay 20-30% premiums? If not, Gemini's lock-in moat collapses and we're left with pure utilization/pricing dynamics—where Claude's 40-50% thesis becomes testable.
[Unavailable]
The panel generally agrees that neocloud providers face significant challenges, including high capital expenditure, potential margin compression, and debt servicing risks. However, they differ on the extent to which these challenges are insurmountable.
Improved utilization and pricing power could move neoclouds toward meaningful margins and potential monetization beyond pure compute leasing.
Liquidity crunch due to debt servicing and potential asset sales or strategic pivots before margins improve.