Get ready... trillions to be "burned" in Artificial Intelligence – Barclays sees a war of attrition, debts skyrocket

The global race for dominance in artificial intelligence (AI) is developing into a ruthless "war of attrition," as the massive cost of building infrastructure, the explosive increase in capital expenditures, and the growing reliance on borrowing create an exceptionally dangerous environment for the world's largest technology companies. The big question is no longer which company possesses the best technology or which will dominate the AI market. It is which one commands the financial resources to endure a long-term investment battle without exhausting its liquidity and without confronting a dangerous financing crisis. Barclays highlights a particularly alarming dimension of the AI race: demand for artificial intelligence remains robust, but financing these massive investments is becoming increasingly difficult. Meanwhile, limited availability of electricity favors companies with the strongest balance sheets, creating a new competitive landscape in which financially weaker players risk getting shut out, impacting AI capital expenditure strategy.

Barclays: AI is transforming from Wall Street driver into a source of risk

According to Julien Lafargue, Head of Market Strategy at Barclays Private Bank, the investment landscape for artificial intelligence has shifted noticeably since mid-October. AI, which until recently served as one of the strongest growth drivers for stock markets, is now turning into a source of anxiety. Barclays underscores that there are no clear indications that demand for artificial intelligence is receding. What investors are now questioning is who will finance the construction of vast infrastructure projects and under what terms. This shift is crucial, as the market begins evaluating not only the future prospects of artificial intelligence but also the actual cost of its deployment. Indicative of the shifting sentiment is that the Nasdaq 100 traded 9.5% below its October high, recording three single-day declines between 2.5% and 3.5% within three weeks. Barclays notes that the intensity of selling pressure is not the main story in itself; the critical factor is that liquidations are concentrated in AI-linked companies, while even positive corporate announcements fail to reverse negative sentiment, reshaping Wall Street tech valuations.

The terrifying problem: Investments rise, but free cash flows disappear

At the heart of the anxiety are the hyperscalers—the tech giants building and operating massive networks of data centers and cloud infrastructure. The largest tech companies continue to ramp up capital expenditures (Capex) to secure the computing power required for AI development. However, according to Barclays, all major hyperscalers have signaled further investment increases for next year, in some cases at an exceptionally aggressive pace. At the same time, Free Cash Flow is turning negative at the group level. This means companies are spending ever larger sums on infrastructure buildouts without generating corresponding free cash flow from operations. Microsoft appears as an exception, but Barclays clarifies that this is largely due to third-party infrastructure deployment lagging behind original schedules, with related expenditures expected to materialize in upcoming quarters. Consequently, the problem is not disappearing; it may simply be deferred, altering hyperscaler free cash flow.

A dangerous vicious cycle of financing is forming

Barclays describes a self-reinforcing risk mechanism that could escalate into a severe problem for the entire artificial intelligence ecosystem. The sequence is specific: tech giants increase investments in data centers and AI hardware; free cash flow compresses or turns negative; companies are forced to seek additional capital from bond and capital markets; if funding grows more expensive or scarce, fewer data centers get built; lower available computing power restricts revenues for both hyperscalers and their suppliers; slowing profitability makes investors less willing to fund the next investment cycle. This represents a vicious cycle capable of turning the AI investment boom into a financing trap. Barclays warns that as long as this mechanism remains active, markets have little reason to price tech companies based on promised future growth, highlighting AI infrastructure funding risks.

AI bonds yielding up to mid-teens – Warning signal from debt markets

Messages emerging from the corporate bond market are particularly concerning. According to Barclays, bonds issued by the most heavily indebted AI infrastructure companies now offer double-digit yields, reaching into the mid-10% range. Concurrently, high-yield bond issuers linked to data centers are trading at credit spreads of roughly 600 to 1,600 basis points. Barclays points out that while this represents a relatively small segment of the overall market, it carries immense significance because this exact segment finances the expansion of additional data center capacity. In other words, the issue extends beyond corporate equity valuations; it is taking root in borrowing costs for enterprises at the core of the AI investment surge. Furthermore, Barclays points to reports indicating that Anthropic's valuation ahead of a potential initial public offering (IPO) was revised downward from around $2 trillion to $1.5 trillion. A lower valuation could constrain capital raised, thereby reducing funds available for infrastructure development while pressuring valuations across the broader sector, impacting data center credit spreads.

Energy becomes the ultimate weapon – Advantage for deep-pocketed tech giants

Another critical element of this new competitive phase is access to electrical power. Scaling data centers requires massive amounts of electricity, making power infrastructure availability an increasingly restrictive bottleneck. In this context, the primary insight highlighted in the ZeroHedge report is that electricity scarcity favors companies possessing the strongest balance sheets. Tech giants with vast cash reserves, capital market access, and the ability to finance long-term power purchase agreements hold a distinct competitive edge. Conversely, smaller AI infrastructure providers and startups reliant on external financing risk being pushed out of the market. Thus, the artificial intelligence race is gradually transforming from a technological rivalry into a test of financial stamina, power sufficiency, and capital access. Survival will not necessarily belong to whoever possesses the best technology, but to whoever can fund their infrastructure for the longest duration, driving data center power availability.

Not a simple sector rotation – Investors abandoning AI altogether

Barclays identifies another troubling pattern in investor behavior. During previous correction phases, investors rotated capital from one artificial intelligence segment to another—for instance, exiting certain semiconductor stocks to buy software companies or tech platforms. This time, however, the picture is different. Shares of semiconductor firms and the so-called Magnificent 7 dropped simultaneously for three consecutive trading sessions in mid-October. According to Barclays, this phenomenon had occurred only four times throughout 2026. Simultaneously, the Philadelphia Semiconductor Index (SOX) logged eight consecutive losing sessions—its longest negative streak since 2008—followed by a single-day drop exceeding 6%. This demonstrates that investors are not merely reallocating capital among AI sub-sectors, but are reducing overall exposure to the artificial intelligence investment thesis, accelerating semiconductor index sell-offs.

Barclays sees risk of further correction, but conditions for strong recovery

Despite these concerns, Barclays does not view the AI investment narrative as over. On the contrary, it estimates that the best trading sessions often follow the worst, and powerful upward relief rallies can emerge. However, it warns that such recoveries will struggle to sustain momentum unless convincing answers are provided regarding financing. The bank also notes that investor positioning has not yet been fully cleared, leaving room for additional selling pressure. At the same time, a self-correcting mechanism exists: as valuations pull back, companies will be forced to scale back expenditures to levels they can genuinely fund. This could yield slower short-term growth but foster a healthier long-term investment environment characterized by higher free cash flows and reduced capital market reliance. Barclays assesses that once investors identify a viable pathway for funding AI infrastructure, the investment thesis can regain momentum, refining tech market correction forecasts.

The grand AI gamble: Who will survive the war of attrition?

The core of the warning is that artificial intelligence is entering a new, far more demanding phase. The initial period was characterized by euphoria, astronomical valuations, and aggressive capital deployment. The next phase will be judged by corporate capacity to translate capital investments into real cash flows and earnings. The question is whether tech giants will manage to sustain current investment cadences without overleveraging their balance sheets. In this new war of attrition, access to capital, energy sufficiency, and financial resilience may prove far more decisive than technological supremacy itself. And that remains the truly sobering message for markets: while demand for AI may remain robust, it is by no means certain that every company investing today will possess the financial stamina to complete the race, determining AI long-term profitability.

www.bankingnews.gr