U.S. AI and semiconductor stocks staged a sharp but uneven rebound on Monday, Oct. 19, ending an eight-session slide in the Philadelphia Semiconductor Index after Goldman Sachs upgraded cloud infrastructure provider CoreWeave. The index closed 1.6% higher after gaining more than 3% during the session, while heavily sold hardware names including SanDisk, SK Hynix and Micron posted some of the day’s strongest advances.
The recovery offered relief after a rapid unwind in AI-linked equities, but positioning data from Goldman Sachs and UBS Wealth Management indicates the move has yet to resolve concerns over leverage, expensive infrastructure financing and the durability of the AI spending cycle.
Goldman raised CoreWeave to “buy” from “neutral,” even as it reduced its price target to $106 from $139. The bank said its revised target represented more than 40% upside from the previous Friday’s closing price. Bloomberg-compiled analyst estimates put the average CoreWeave target at $135.23.
CoreWeave rose as much as 7.4% intraday before reversing to close down 0.4%. That reversal captured the market’s mixed response: traders were willing to buy shares after an aggressive decline, but remained reluctant to hold the company through persistent questions about debt, capital expenditure and the financing requirements of AI data centers.
Semiconductor rebound follows concentrated selling
SanDisk gained 5.5% on Monday, while SK Hynix’s U.S.-listed shares rose 4.6%, Micron advanced 3%, and Intel added 2.4%. The strongest gains were concentrated in companies that had fallen most sharply during the preceding week, suggesting short covering and bargain buying played a substantial role in the rebound.
Goldman Sachs said hedge funds had been net sellers of global AI equities for five consecutive trading days. According to the bank, cumulative net buying of AI-related stocks since the start of 2025 had been erased during the selling wave.
The scale of the flow reversal was significant. Goldman estimated last week’s net selling in global AI shares at negative 2.1 standard deviations, a measure indicating activity well beyond a normal weekly range. Long sales outpaced short selling by roughly three to one, meaning managers were primarily liquidating existing bullish holdings rather than building major new bearish positions.
Net allocation to global AI equities fell to 26.1% from a June high of 31.7%, according to Goldman. The bank placed that level at the 49th percentile of the prior year, while the long-to-short ratio dropped to 3.27 from 4.34, reaching the 7th percentile.
Those figures point to a reduction in conviction after months in which AI infrastructure had been among the market’s most crowded growth trades. Yet gross exposure has remained elevated. UBS Wealth Management said hedge fund net exposure had fallen to the 30th percentile of the past five years, while gross exposure remained above the 90th percentile. That combination can leave funds vulnerable to further deleveraging if price swings remain large.
CoreWeave financing remains under scrutiny
CoreWeave has become a focal point because its business model requires substantial spending on the servers, chips and data-center capacity used to supply computing power for AI workloads. Demand for those services can remain strong while the cost of funding the infrastructure becomes more difficult to manage.
The company’s bonds fell last week to prices commonly associated with distressed credit conditions, while the cost of credit-default protection rose intraday to a record high, according to market commentary cited in the supplied materials. Those moves added pressure to the equity market’s debate over whether AI infrastructure can continue to be financed on terms that support current valuations.
The question reaches beyond CoreWeave. AI developers and cloud providers have committed enormous sums to advanced chips, power capacity, networking equipment and new data centers. UBS Wealth Management raised its 2026 forecast for global AI capital spending to $900 billion, representing an 84% increase from its prior-year comparison, and projected $1.2 trillion for 2027. UBS linked the revisions to cost increases across the AI server supply chain.
Higher projected spending supports demand for semiconductor companies, memory suppliers and data-center equipment makers. It also raises the stakes for companies funding expansion through debt or relying on capital markets to sustain increasingly large build-outs.
Options flows may have intensified the decline
CoreWeave’s options market showed how quickly leveraged enthusiasm can disappear during a selloff. Call options expiring in November with a $95 strike price fell from a high of $18 to around $1, according to the supplied market data.
When call-option values collapse, market makers that had bought stock to hedge their exposure can reduce those hedges. That process can add selling pressure to the underlying shares, particularly in volatile stocks with active options markets. It does not alter a company’s operations, but it can magnify short-term price moves and accelerate an already crowded exit.
CoreWeave’s shares fell 37.4% from Sept. 8 in the basket cited in the supplied materials. Over the same period, SK Hynix fell 25.8%, SanDisk lost 23.8%, Bloom Energy declined 18.1%, and Intel dropped 9.4%, while AMD gained 9.1%. The divergence illustrates that traders have not treated the AI supply chain as a single trade: companies with differing balance sheets, product exposure and revenue visibility have begun to separate.
Earnings will test the spending assumptions
The next major test will come from earnings reports by large technology companies, including Meta and Tesla, where markets will focus on capital expenditure plans, demand for computing capacity and any changes to data-center timelines.
A sustained semiconductor recovery would require more than a reversal in heavily sold names. Companies deploying AI infrastructure will need to show that rising capital costs are matched by durable demand, while hardware suppliers will need to demonstrate that orders remain supported beyond inventory rebuilding and near-term project commitments.
Rising interest rates add another layer of pressure. Higher Treasury yields raise the benchmark cost of borrowing across corporate credit markets, making debt-funded server and data-center construction more expensive. That places greater attention on cash flow, customer contracts and the credit quality of companies trying to scale AI capacity quickly.
Monday’s rally reduced some immediate selling pressure, but the underlying market has shifted from broad enthusiasm toward a closer examination of financing structures, positioning and spending discipline. The upcoming earnings season will show whether the AI trade can regain that confidence through operating results rather than a technical bounce.
Navigate AI volatility smarter—explore automated strategies with AI copy trading to mirror top performers in real time.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
