By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
While Alphabet's need for third-party compute validates near-term demand for neoclouds like Nebius (NBIS) and CoreWeave (CRWV), panelists express concerns about their high valuations, heavy cash burn, and potential commoditization risk as hyperscalers scale their own capacity.
Risk: Commoditization risk and potential supply glut due to improved AI model training efficiency or accelerated hyperscaler buildout.
Opportunity: Multi-year take-or-pay deals that amortize capex and raise utilization floors if hyperscalers outsource at scale in the near term.
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 →
Google parent Alphabet (GOOG) (GOOGL) recently released its latest earnings report, which wasn't just about stronger AI demand or another impressive increase in capital spending. Buried within the company's commentary was a signal that could have even bigger implications for a fast-growing corner of the AI infrastructure market – neocloud providers.
The tech giant once again raised its 2026 capital expenditure forecast to between $195 billion and $205 billion, underscoring just how aggressively hyperscalers are investing to keep up with the artificial intelligence (AI) boom. But perhaps the more important takeaway came from management's admission that even Google can't build capacity fast enough on its own.
During the earnings call, CFO Anat Ashkenazi said Google plans to expand its use of third-party computing capacity as a temporary bridge while additional in-house infrastructure comes online. In other words, demand for AI compute is growing faster than hyperscalers can currently supply it.
That comment immediately caught Wall Street's attention because it reinforces one of the biggest investment themes in AI today. Companies offering specialized cloud infrastructure are becoming increasingly important as enterprises race to secure GPU capacity. Instead of waiting years for hyperscalers to expand their own data centers, customers – and even the cloud giants themselves – are increasingly turning to neocloud providers to help bridge the gap.
Investors were quick to connect the dots. Shares of Nebius Group (NBIS) and CoreWeave (CRWV) climbed following Alphabet's earnings call and management's comments, as many viewed them as a vote of confidence in the growing role neocloud companies are playing in the AI ecosystem.
For investors looking to capitalize on the booming neocloud opportunity, NBIS and CRWV stocks could be wise buys now.
Stock #1: Nebius
Headquartered in Schiphol, the Netherlands, Nebius Group is an AI infrastructure company building a full-stack cloud platform for AI applications. The company provides large-scale GPU clusters, AI cloud services, and developer tools that help enterprises train and deploy AI models. Beyond its core AI business, Nebius owns TripleTen, a technology reskilling platform, and Avride, an autonomous driving and delivery robotics company.
Backed by a strategic partnership with Nvidia (NVDA) and led by founder and CEO Arkady Volozh, Nebius is rapidly expanding its AI supercomputing and cloud infrastructure, serving customers across industries and boasting a market capitalization of roughly $55.9 billion.
Nebius has been one of the standout performers in the AI infrastructure space, rewarding investors who stayed patient through the volatility. Over the past 52 weeks, NBIS stock has soared 277.4%, while gaining 135.2% year-to-date (YTD). The rally accelerated after the company delivered blockbuster first-quarter fiscal 2026 earnings in May, and momentum strengthened further when Nebius was added to the Nasdaq-100 Index in June, bringing even more attention from investors.
Like many high-flying AI stocks, though, the rally eventually paused. After climbing to an all-time high of $299.86 on June 22, NBIS pulled back as investors locked in profits amid broader concerns that AI-related stocks had become too expensive after their massive run.
More recently, however, sentiment has started to improve again. NBIS stock has increased 10.76% over the past five trading sessions after a regulatory filing revealed that Nvidia now owns a 9.3% stake in the neocloud company. Investors viewed the disclosure as another strong vote of confidence from the AI chip leader, reinforcing Nebius' position as a key partner in next-generation AI cloud infrastructure. A fresh round of bullish analyst calls also helped fuel the rebound.
Technically, the picture is improving as well. The stock's 14-day RSI has recovered to 44.07, suggesting bearish momentum has eased, and the shares are no longer in either overbought or oversold territory.
Valuation-wise, NBIS stock is priced at 16.69 times forward sales, representing a premium to the sector average.
Nebius' fiscal Q1 2026 earnings report, released in May, was impressive, with revenue surging 684% year-over-year (YOY) to $399 million. Its core AI business – excluding Avride and TripleTen – grew even faster, with revenue soaring 841% to $390 million. The growth was driven by rapid data center expansion, strong utilization rates, and healthy pricing across its AI cloud platform. Investors welcomed the results, sending NBIS stock up 22.4% over the two trading sessions following the earnings release.
Scaling that quickly does not come cheap, though. Heavy investments in new data centers, AI platform development, engineering talent, and strategic acquisitions pushed the company's adjusted net loss up 20% annually to $100.3 million. Still, adjusted EBITDA climbed to $129.5 million, representing a 32% margin, while the core AI business expanded its adjusted EBITDA margin to 45%.
Nebius strengthened its balance sheet, raising $4.3 billion through convertible senior notes and securing Nvidia's $2 billion strategic investment. The quarter ended with $9.3 billion in cash and cash equivalents, while operating cash flow jumped to $2.26 billion, giving the company plenty of financial flexibility to continue expanding its AI infrastructure.
Additionally, management unveiled a transformative five-year, $27 billion agreement with Meta Platforms (META), including a $12 billion dedicated compute commitment and $15 billion in optional capacity, providing greater revenue visibility while preserving flexibility to serve higher-margin AI cloud customers. Nebius also strengthened its partnership with Nvidia, earning Nvidia Exemplar Cloud status for GB300 training workloads and gaining access to future GPU platforms, including Vera Rubin.
Meanwhile, a customer pipeline that grew 3.5x sequentially prompted the company to expand contracted power capacity beyond 3.5 GW, with a target of more than 4 GW by 2026, supported by a new 1.2 GW Pennsylvania AI factory and acquisitions that broaden its full-stack AI platform.
Looking ahead, management estimates annualized run-rate revenue to be between $7 billion and $9 billion in 2026, with total group revenue projected between $3 billion and $3.4 billion and an adjusted EBITDA margin of roughly 40%. The company raised its 2026 capex forecast to $20 billion to $25 billion.
The company is all set to release its second-quarter earnings report for fiscal 2026 soon. Analysts monitoring the company anticipate Q2 revenue of $582.8 million, with losses coming in at $0.67 per share. Looking further ahead, fiscal 2026 loss per share is expected to widen by 7.9% YOY to $1.91, before narrowing down by 63.9% annually to -$0.69 in fiscal 2027.
Baird initiated coverage of NBIS with an "Outperform" rating and a $250 price target, citing the company's expanding role in AI inference. The brokerage firm believes Nebius is well positioned for long-term growth, supported by its strong software capabilities, growing customer base, experienced management team, and strategic acquisitions that have broadened its AI platform and service offerings as enterprise AI demand continues to evolve.
Overall, NBIS stock carries a consensus "Moderate Buy" rating. Among the 16 analysts in coverage, 11 suggest a "Strong Buy," and five analysts recommend a "Hold." The stock has a mean price target of $254.92, implying upside potential of 32.1% from the current price levels. The Street-high target of $410 – set by Northland Securities – suggests that NBIS could rise as much as 112.4%.
Stock #2: CoreWeave
Founded in 2017 and headquartered in Livingston, New Jersey, CoreWeave is a cloud infrastructure company purpose-built for AI workloads. Its GPU-optimized cloud platform combines proprietary software with high-performance GPU and CPU computing resources to help enterprises train, fine-tune, and deploy AI models at scale.
Beyond compute, the company offers AI-focused storage, infrastructure management, and machine learning tools that simplify the development and deployment of large-scale AI applications. As demand for generative AI continues to grow, CoreWeave has emerged as a key player in AI infrastructure, currently having a market capitalization of $36.3 billion.
CRWV stock has mirrored the AI trade almost perfectly – explosive rallies, sharp pullbacks, and plenty of volatility along the way. The company made its public debut in March 2025 at $40 a share, and investor enthusiasm quickly sent the stock soaring. By June, CRWV had climbed to an all-time high of $106.54, fueled by a steady stream of AI cloud contracts, strategic partnerships, acquisitions, and an expanding presence in the federal market.
But like many high-growth AI stocks, the momentum did not last forever. Investors hit the brakes after the company issued weaker-than-expected guidance for the first quarter of fiscal 2026, sending the stock tumbling 18.5% in a single session.
The stock rallied again to $138.25 in May 2026, but has since retreated and now trades around $81.10. Overall, CRWV is down 52% from its October peak and has declined 38.7% over the past 52 weeks. Although up 2.65% YTD, the stock fell 33.26% over the past three months.
More recently, however, sentiment has begun to improve. Encouraged by bullish analyst calls and renewed optimism about long-term AI infrastructure demand, buyers have started stepping back in. Over the past five trading days alone, CRWV stock has gained marginally 0.41%, suggesting renewed optimism surrounding long-term demand for AI infrastructure.
On the charts, green volume bars point to sustained buying interest, while the 14-day RSI hovers at 36.49, leaving room for further upside.
Valuation-wise, CRWV stock is trading at 3.50 times sales, outpacing the sector average, suggesting investors are willing to pay up for the company's outsized AI-driven growth prospects and long-term expansion potential.
CoreWeave reported first-quarter fiscal 2026 results on May 7, and it delivered a top-line beat, even though losses came in wider than what Wall Street had expected. However, management said it recorded the strongest bookings quarter in the company's history, highlighting just how quickly customers are snapping up AI computing capacity.
Revenue more than doubled, soaring 111.6% YOY to $2.1 billion, fueled by surging demand for its specialized AI cloud platform. Yet, profitability remains the tradeoff for that breakneck growth. CoreWeave posted a net loss of $740 million, or $1.40 per share, as it continued pouring billions into expanding its infrastructure, compared with a loss of $315 million, or $1.49 per share, a year earlier.
However, adjusted EBITDA jumped 90.9% YOY to $1.2 billion, while the company maintained a healthy 56% EBITDA margin.
Perhaps the biggest headline, though, was the backlog. CoreWeave ended the quarter with a record $99.4 billion revenue backlog, including more than $40 billion in new commitments, marking a staggering 284% annual growth. Management said demand is being fueled by the shift from AI model training toward inference, AI agents, and enterprise-scale applications, while early Vera Rubin deployments and continued monetization of Blackwell, Hopper, and Ampere GPUs are bringing in fresh business. The company also crossed 1 gigawatt of active power, a milestone achieved by only a handful of cloud infrastructure providers.
Looking ahead, management is not thinking small. CoreWeave plans to scale its active power capacity to more than 8 GW by 2030 and exceed $30 billion in annualized revenue by the end of 2027, with roughly 75% of that target already backed by long-term contracts.
For Q2, management expects revenue of $2.45 billion to $2.60 billion and $7 billion to $9 billion in capex. For fiscal 2026, management forecasts $12 billion to $13 billion in revenue alongside $31 billion to $35 billion in capex as it races to stay ahead of AI demand.
Wall Street knows CoreWeave is not out of the woods just yet. For fiscal Q2 2026, analysts project loss per share to widen 142.6% YOY to $1.31. For the full year, losses are anticipated to deepen 70% YOY to $4.57 per share. But the silver lining is that they are expected to narrow 29.3% in FY2027 to a $3.23 per share loss.
Wall Street has become increasingly optimistic about CRWV's growth potential. Recently, Truist Securities upgraded the stock from a "Hold" to a "Buy," arguing that the company's growth opportunities outweigh its financial risks. The brokerage firm believes rising enterprise adoption of open-source AI models and the growing push for sovereign AI are expanding CoreWeave's addressable market. With roughly 1 gigawatt of active power already online and a sizable contracted revenue backlog, Truist sees the company as well positioned to capitalize on that demand. The brokerage firm set a price target of $126.
Meanwhile, Baird recently initiated coverage with an "Outperform" rating and a $100 price target. Analyst Rob Oliver called CoreWeave a "pioneer and leader" in purpose-built AI infrastructure, pointing to its rapidly expanding compute capacity, proprietary technology platform, and close partnership with Nvidia. Baird also believes ongoing GPU supply constraints should favor experienced operators like CoreWeave that have already proven they can execute at scale.
The consensus rating on CRWV stock remains at "Moderate Buy." Among 35 analysts covering the stock now, 20 advise a "Strong Buy," 14 analysts are playing it safe with a "Hold" rating, and the remaining one analyst remains skeptical with a "Strong Sell."
Even with the stock sliding, analysts are bullish. The average target of $136.15 signals 85.4% upside potential, and the Street-high target of $250 suggests that the stock could rise as much as 240.4% from the current level.
On the date of publication, Sristi Suman Jayaswal did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Barchart.com
Four leading AI models discuss this article
"Google’s comments validate near-term neocloud demand but do not guarantee long-term structural market share once hyperscalers catch up on capex."
Alphabet’s admission that it cannot scale internal capacity fast enough is a clear near-term tailwind for neoclouds, validating demand spill-over. Nebius (NBIS) shows 841% core-AI revenue growth, a $27B Meta deal, Nvidia’s 9.3% stake, and $9.3B cash; CoreWeave (CRWV) boasts a $99B backlog and 56% EBITDA margins. Yet both trade at 16.7× and 3.5× forward sales while still burning cash heavily and guiding to multi-billion capex. The article correctly flags GPU scarcity as a bridge opportunity, but glosses over execution risk, customer concentration, and the fact that hyperscalers will eventually internalize most incremental demand.
Once hyperscalers finish their 2026-2027 build-outs, the very capacity bridge that now props up neocloud valuations could collapse, leaving NBIS and CRWV with stranded assets and collapsing multiples.
"Neocloud providers are currently priced as permanent infrastructure winners, but they are more likely to be treated as temporary, high-cost bridge capacity by hyperscalers."
The market is conflating 'capacity shortage' with 'long-term pricing power' for neoclouds like Nebius (NBIS) and CoreWeave (CRWV). While Google’s reliance on third-party compute validates the immediate demand surge, it also highlights the commoditization risk; hyperscalers are merely using these firms as a stopgap until their own custom silicon and data centers scale. With NBIS trading at 16.69x forward sales and CRWV burning cash to fund massive capex, the margin for error is razor-thin. If AI model training efficiency improves—reducing the raw compute intensity per parameter—these capital-heavy business models could face a brutal supply glut just as their massive infrastructure projects come online.
If these neoclouds successfully lock in multi-year, take-or-pay contracts with hyperscalers, they effectively become the 'utility' layer of the AI economy, insulating them from cyclical fluctuations in GPU demand.
"Neocloud demand is real, but NBIS and CRWV's valuations price in perfect execution, sustained AI capex growth, and minimal hyperscaler self-sufficiency—three assumptions that aren't guaranteed."
The article conflates two distinct narratives: (1) hyperscalers need third-party capacity NOW, which is real and bullish for neocloud, and (2) NBIS and CRWV are the answers, which is murkier. NBIS trades 16.69x forward sales with $100M+ adjusted losses despite 684% revenue growth—classic hypergrowth math, but the Meta deal ($12B committed, $15B optional) is 40-50% of their $7-9B 2026 revenue guidance. Concentration risk. CRWV's $99.4B backlog sounds massive until you realize it's 7-8x annualized revenue—contract duration matters enormously. Both burn cash aggressively ($20-25B and $31-35B capex respectively). The article never asks: what if hyperscalers accelerate their own buildout, or AI demand growth moderates? Or if GPU supply normalizes?
Google's capex guidance ($195-205B) is THEIR OWN spend, not outsourced. If hyperscalers can self-fund at that scale, why would they permanently outsource to neocloud providers rather than view them as temporary gap-fillers? The backlog metrics assume contract conversion and sustained utilization.
"Near-term upside for neocloud stocks hinges on outsized capex and long-term contracts, but profitability remains highly uncertain and valuations may compress if AI demand cools or hyperscalers accelerate internal capacity."
Alphabet's bigger capex forecast and its admission that it needs third-party compute highlight a shared industry crunch that could extend the neocloud thesis. The article frames NBIS and CoreWeave as near-term beneficiaries of AI demand, backed by multi-billion partnerships and expanding capacity. But these are wildly capital-intensive growth stories with questionable near-term profitability. NBIS reports losses with sizable capex, even as EBITDA margins look strong on the core AI business; CoreWeave posts large losses despite a $99.4B revenue backlog and an 8 GW 2030 target. Valuations imply sustained hypergrowth (NBIS ~16x forward sales; CRWV ~3.5x). A macro pullback, GPU pricing shifts, or hyperscalers accelerating internal capacity could stall the rally.
Risks include demand normalization or accelerated hyperscaler self-build that erodes pricing and contract visibility. If financing costs stay high or backlogs fail to convert to sustained revenue, the upside could be limited.
"Outsourcing momentum likely locks in longer contracts, extending the capacity bridge beyond 2027."
Claude correctly flags concentration but underplays the second-order effect: if hyperscalers like Google outsource at scale in 2025-26 to hit AI timelines, NBIS/CRWV gain multi-year take-or-pay deals that amortize their capex and raise utilization floors. The 'temporary bridge' narrative ignores how custom ASIC cycles lag GPU availability by 18-24 months, extending the neocloud window.
"Neocloud viability is tied more to interest rate-sensitive debt refinancing than to the duration of the GPU supply-demand mismatch."
Grok, your 'take-or-pay' thesis assumes these neoclouds have the balance sheet to survive the 24-month ASIC lag. You ignore the interest rate sensitivity; if these firms need to refinance their massive capex debt in a higher-for-longer environment, the cost of capital will cannibalize their margins long before hyperscalers internalize demand. They aren't just selling compute; they are leveraged bets on credit markets. If liquidity tightens, these 'bridges' collapse under their own debt-servicing requirements regardless of demand.
"Leverage risk is real, but only if neoclouds haven't already front-loaded long-duration debt financing—which we don't know from the article."
Gemini nails the refinancing risk—I hadn't weighted it enough. But there's a counterpoint: NBIS raised $9.3B cash, CRWV has backing from Blackstone/Brookfield. If they lock in 3-5 year contracts NOW at fixed rates before rates normalize, they hedge refinancing risk entirely. The real question isn't 'can they survive higher rates?' but 'did they already lock in cheap debt before the window closed?' That changes the timeline materially.
"Backlog conversion risk undermines the notion that fixed-rate debt and take-or-pay contracts guarantee durable margins."
Gemini, refinancing risk is real but not the only pressure. The bigger flaw in the take-or-pay defense is conversion risk: CRWV’s $99.4B backlog is 7-8x annualized revenue, meaning a lot of that revenue depends on multi-year contract closure and sustained utilization. If hyperscalers accelerate self-build or demand normalizes, fixed-rate debt won’t rescue margins—backlog may deform into idle capacity or underutilization, depressing returns despite lower financing risk.
While Alphabet's need for third-party compute validates near-term demand for neoclouds like Nebius (NBIS) and CoreWeave (CRWV), panelists express concerns about their high valuations, heavy cash burn, and potential commoditization risk as hyperscalers scale their own capacity.
Multi-year take-or-pay deals that amortize capex and raise utilization floors if hyperscalers outsource at scale in the near term.
Commoditization risk and potential supply glut due to improved AI model training efficiency or accelerated hyperscaler buildout.