AI-linked stocks rebounded on Monday as falling government bond yields and easing crude oil prices offered relief after a bruising selloff last week, but renewed gains did little to resolve investor concerns about the debt burden of AI infrastructure companies, rising financing costs and the valuations attached to the sector.
CoreWeave rose 2.4% in volatile trading after losing roughly a quarter of its market value last week, when its $50 billion debt burden triggered a crisis of confidence in the financing model behind AI computing infrastructure. The company’s difficulties contributed to a broader selloff across technology and AI-related stocks, exposing the sensitivity of the sector to borrowing costs and investor appetite for risk.
The rebound extended beyond CoreWeave, with rival neocloud provider Nebius Group climbing 3.4% after falling more than 31% over the previous two weeks. Data center operators Applied Digital and IREN advanced 8.5% and 5%, respectively, as investors bought into companies that had suffered steep losses.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
However, the recovery remained fragile. Investors are increasingly weighing the enormous capital requirements of AI against the prospect of sustained earnings growth, particularly as companies building computing capacity depend on continued access to debt and equity financing.
“It’s going to continue to be a factor, it’s going to be a constant push-pull over time,” said Thomas Hayes, chairman at Great Hill Capital.
CoreWeave Upgrade Offers Relief, but Debt Remains a Concern
CoreWeave received support from Goldman Sachs, which upgraded the company’s stock to “buy” from “neutral”. The brokerage argued that constrained computing capacity and CoreWeave’s performance advantage over competing neocloud providers could increase annual recurring revenue per megawatt over the next two years. It also described the recent share-price decline as a buying opportunity.
The upgrade offers a more constructive assessment of CoreWeave’s underlying business prospects at a time when the market is questioning whether its financing requirements could undermine its growth potential. Demand for AI computing remains an important source of support for specialized cloud providers, but the ability to convert that demand into sufficient revenue and cash flow is becoming crucial to investors.
Developments in the company’s bond market provide a clearer indication of the pressure. The yield on CoreWeave’s bond maturing in 2031 fell 19.5 basis points to 15.13% on Monday, according to Tradeweb data, retreating from levels above 16% last week.
Although the decline offered some relief, the yield remained elevated, indicating that creditors continued to demand substantial compensation for the risks associated with lending to the company. High borrowing costs can increase the expense of funding additional data center capacity, potentially putting pressure on profitability and limiting the financial flexibility needed to sustain rapid expansion.
For companies whose growth depends on building or securing large quantities of computing infrastructure, access to financing is as important as customer demand. Expanding capacity requires substantial upfront investment, while the revenue generated by that capacity must eventually justify the capital committed to it.
Hayes cautioned against interpreting Monday’s recovery as evidence that the underlying financing concerns had disappeared.
“You could certainly see a dead cat bounce today and maybe some continuation tomorrow, but it doesn’t change the fact that the bond market is very skeptical about continuing to finance this without demanding higher and higher yields,” he said.
His assessment highlights the distinction between a short-term recovery in share prices and a meaningful improvement in credit conditions. Equity investors may respond quickly to a brokerage upgrade or a decline in bond yields, but creditors ultimately assess whether a company’s future cash flows can support its debt obligations.
Interest-Rate Fears Deepen the AI Market Correction
The previous week’s selloff was intensified by a hotter-than-expected inflation reading, which increased expectations of interest-rate hikes and raised concerns that heavily indebted AI companies could face more expensive financing.
Higher interest rates can affect the sector through several channels. They raise the cost of borrowing, increase the returns investors can obtain from lower-risk assets, and reduce the present value of profits expected further into the future. These pressures can weigh heavily on companies valued on expectations of rapid growth and future earnings rather than established, near-term cash generation.
The impact was especially visible in semiconductor stocks. The Philadelphia Semiconductor Index fell more than 11% last week, with chipmakers among the hardest-hit groups in the broader AI selloff. The index lost approximately $2 trillion in market value during an eight-session losing streak through Friday.
On Monday, the index gained 2.4%, putting it on course to end its longest losing streak since October 2008.
The recovery spread across major semiconductor companies. Nvidia rose 1.5%, Advanced Micro Devices gained 3.6%, and Broadcom advanced 2.4%, following sharp declines the previous week. Memory-chip manufacturer Micron Technology climbed 4.2% after losing nearly 15% last week. Data-storage companies SanDisk, Western Digital and Seagate also advanced after suffering substantial losses.
Intel gained 4.5% after JPMorgan upgraded the stock to “neutral” from “underweight”. The brokerage cited growing workloads from agentic AI systems and tight supplies of advanced chips as factors that could support upward revisions to earnings estimates through 2027.
The upgrade suggests that the investment case for parts of the semiconductor industry remains intact despite the broader market correction. Increasing use of AI agents could generate additional demand for computing resources, while supply constraints may support pricing and earnings for companies positioned to benefit.
Nevertheless, the sharp decline in semiconductor valuations indicates that strong demand alone may not protect stocks from a reassessment of their prices. Investors must also consider how much future growth is already embedded in valuations and whether expected returns justify the capital required to expand production and infrastructure.
The market correction comes as spending on AI infrastructure is expected to increase substantially. Morgan Stanley forecasts that total AI investment will reach $1.6 trillion in 2027 and $1.8 trillion in 2028, pointing to the scale of the capital commitments expected across the industry.
That investment outlook presents a contradiction for the market. Rising expenditure can create opportunities for semiconductor manufacturers, data center operators and specialized cloud providers, but it also increases the importance of financing conditions and the returns generated by new capacity.
The contrast between investment growth and investment profitability is central to the current debate. Companies can continue spending aggressively because they expect AI adoption to accelerate, yet investors may become less willing to finance that expansion if revenue growth fails to keep pace with costs or if the resulting earnings do not justify prevailing valuations.
The concern is acute for infrastructure providers carrying large debt loads. They must finance facilities and computing equipment before all the associated revenue is realized, leaving them exposed to changes in interest rates, capital-market conditions and customer demand. If financing becomes more expensive while equity valuations decline, raising additional capital can become more difficult on favorable terms.
The same pressures extend to chipmakers, whose earnings prospects depend partly on sustained infrastructure spending by cloud providers and other large technology companies. A slowdown in AI investment could affect suppliers even if demand for AI applications continues to increase.
Big Tech Earnings Become the Next Test
Investors are now looking to quarterly results from major technology companies and hyperscalers for evidence that the AI spending boom can translate into earnings growth sufficient to support premium valuations.
Meta Platforms, the parent of Facebook, and Tesla are scheduled to report results later this week, alongside SpaceX. Semiconductor-related companies Lam Research, Texas Instruments, and Western Digital are also due to release quarterly results.
The reports could help determine whether the recent selloff represents a temporary repricing driven by interest-rate concerns or a more significant reassessment of the economics of AI investment. Investors will be looking beyond revenue growth to assess profitability, capital expenditure, cash generation, and the expected returns from AI-related spending.
For major technology companies, continued investment may be justified if AI products generate new revenue streams, improve operating efficiency or strengthen their competitive positions. But sustained spending without corresponding evidence of financial returns could intensify scrutiny of the sector’s capital allocation decisions.
The results from chipmakers and storage companies will also provide insight into demand for the hardware underpinning AI development. Strong orders and earnings forecasts could reinforce expectations of continued infrastructure expansion, while weaker guidance could raise questions about whether spending plans are sustainable at current levels.
“The question this quarter is not whether AI will continue to grow, it’s whether investors are still willing to pay a premium for that growth,” said Jay Woods, chief market strategist at Freedom Capital Markets.
The AI sector’s long-term growth prospects remain a major source of investor interest, but the market is becoming more sensitive to the cost of pursuing them.