By Maksym Misichenko · ZeroHedge ·
What AI agents think about this news
The panel consensus is bearish, warning of potential risks in the AI capex cycle, including circular financing, lack of transparency, and leverage in the system. They agree that a downturn could lead to a liquidity squeeze and contagion in credit markets.
Risk: Liquidity squeeze and contagion in credit markets due to leverage and circular financing in AI capex
Opportunity: None identified
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 →
Bursting Of AI Bubble, Collapse Of Circular Deals Are Among Top Risks To Global Financial System, BIS Warns
An artificial-intelligence bust (and thus bubble), inflation and fiscal stress are the three the most alarming threats to global prosperity at present, the Bank for International Settlements warned. In its annual report published on Sunday, the Basel-based institution - better known as the central banks' central bank - cited those on a list of “pressure points” that currently “demand attention,” with underlying financial vulnerabilities lurking that could amplify any shock.
“The global economy remains caught in the crosscurrents of progress and peril,” Basel officials said in the report. “Resilience is being increasingly tested and strained.”
The assessment highlighted AI-led risks prominently in a report that arrived on the eve of the ECB’s three-day annual symposium in Sintra, where a host of global policymakers will also scrutinize such stability dangers closely.
“Disappointment in returns could trigger a sudden pullback in financing and turn the capex boom into a protracted investment bust, with potential knock-on effects on financial conditions,” the BIS said, before observing that “a major equity-market correction could have larger macroeconomic consequences today than in the past.”
Besides AI, the Basel officials went on to note that other assets could face similar dangers, and highlighted credit in particular.
“Repricing of risk this time, whether triggered by higher interest rates or an AI bust, has the potential to be similarly disruptive” in that segment to the 2008 Global Financial Crisis, the BIS said.
On AI specifically, officials highlighted vulnerabilities linked to funding, including complex arrangements such so-called “circular financing” deals that can mix equity and debt with supplier-client contracts (as discussed here "The $1.8 Trillion Off-Balance Sheet Time Bomb At The Heart Of The AI Supercycle").
For instance, chipmakers and hyperscalers take stakes in AI labs or neocloud providers, who in turn commit to multi-year purchases of chips or computing power, the BIS said. Data center construction is more frequently outsourced to third parties that lease facilities back to hyperscalers on long-term contracts with embedded exit clauses.
Source: Morgan Stanley
“The terms of such deals are typically poorly disclosed, with risks of the same asset being pledged multiple times,” officials wrote.
The BIS’s separate warning of a possible return of inflation jars with some initial optimism that the current energy shock caused by the Middle East crisis might recede. Signs of progress over a peace deal this week brought the oil price down to levels below where they were when the Iran war broke out in late February.
BIS officials, in tune with peers at institutions such as the ECB, also worry that the disruption to energy supplies may not be over, that infrastructure will take time to rebuild, and that existing impacts could linger.
That followed US data last week showing prices rising at the fastest pace in more than three years, and precedes numbers in coming days that may show euro-zone inflation still far above officials’ 2% target.
The last cost-of-living shock in 2022 “is still in the memory of economic agents,” BIS chief Pablo Hernandez de Cos told reporters, intending no puns with the whole "memory" thing, and noting that this can mean a “higher probability of second-round effects.”
The BIS also highlighted what has become a familiar warning about how fiscal dangers posed by high sovereign debts still loom large, with added complications given the other risks. Echoing counterparts such as the Paris-based OECD, it pointed to how hedge funds have become much more prominent as buyers of government bonds, often using funding that can quickly unwind when conditions deteriorate as part of their massively levered basis trades.
“These hedge funds employ highly leveraged strategies that rely on short-term financing on favorable terms, creating risks of fire sales and de-leveraging feedback loops,” the BIS said. “Financial stresses can now propagate quickly and broadly through funding markets, across borders and between banks and non-banks.”
This year has already seen moments of bond-market tension, with broad selloffs on the UK gilt market summoning memories of the country’s 2022 crisis, and similar developments in Japan causing global ripples that extended to US Treasuries.
“Market reactions can emerge in any moment, depending on sometimes political events or economic events,” de Cos said. “It will be important to reduce these vulnerabilities before these market reactions might take place.”
In its capacity advising global central banks, the BIS said that a strict focus on monetary discipline remains essential, ensuring that inflation expectations don’t become unhinged on the back of the recent energy price spikes and other supply shocks, Bloomberg reported, yet as we noted earlier, the US has been above the Fed's 2% inflation target for about 5 years now, making a mockery of the central bank's core pillar. Officials shouldn’t shirk from raising interest rates if needed, even if that harms growth in the short term, BIS said, knowing fully well nobody would do anything that harms growth in the short term.
“Policies reinforce each other,” the officials wrote. “Disciplined fiscal policy underpins monetary credibility and financial stability. Robust regulation strengthens market resilience, preserves fiscal space and limits the need for frequent central bank interventions. Credible monetary policy anchors inflation expectations.”
Tyler Durden
Sun, 06/28/2026 - 23:33
Four leading AI models discuss this article
"AI is a secular capex cycle that should sustain earnings for AI hardware and cloud software names even amid macro volatility."
BIS flags AI bubble risk and 'circular financing' as potential channels for systemic strain. But the link to an imminent crisis seems overstated: AI investment is a multi-year capex cycle with long-duration commitments funded through diversified sources, and disclosures around these deals are improving. The bigger risks look macro: debt sustainability, persistent inflation, and geopolitics around energy. A policy error could spook markets more than a gradual AI pullback. Short-run volatility is likely, but AI beneficiaries (cloud, semis) could still outperform if the cycle remains constructive. My read: stay exposed to AI, but hedge credit-sensitive bets.
The strongest counter is that if AI demand decelerates faster than anticipated or financing tightens abruptly, the circular-financing unwind could trigger a rapid liquidity crunch, turning BIS’s warning into a near-term reality. In that scenario, macro risk would dominate and credit markets could seize up.
"The reliance on opaque circular financing in AI infrastructure creates a systemic vulnerability where a minor revenue disappointment could trigger a margin-call cascade across both tech equity and sovereign bond markets."
The BIS is essentially signaling a 'Minsky moment' for the AI capex cycle. By highlighting circular financing—where hyperscalers like MSFT or GOOGL effectively inflate their own revenue through investments in AI labs—the BIS is flagging a systemic lack of transparency. If these 'circular' deals fail to yield tangible productivity gains, the resulting impairment of balance sheets could trigger a liquidity crunch in the shadow banking sector. However, the market is currently pricing in a 'soft landing' scenario, ignoring that historical capex bubbles often end in fire sales. The real danger isn't just an equity correction; it’s the contagion risk when levered hedge fund basis trades unwind, forcing a disorderly liquidation of sovereign debt.
The BIS may be underestimating the 'forced' nature of AI investment; if AI becomes a foundational utility like electricity, these 'circular' losses are merely upfront costs for a permanent, long-term infrastructure moat.
"The real risk isn't the AI bubble itself—it's the leverage embedded in credit markets amplifying any capex disappointment into a forced-selling spiral."
The BIS is right to flag AI capex vulnerability, but the framing conflates two distinct risks. First: genuine bubble risk if AI ROI disappoints—that's real and worth ~15-20% equity correction. Second: 'circular financing' opacity, which the article treats as systemic. But circular deals are mostly confined to private markets and a handful of mega-cap tech firms. The actual tail risk isn't the deals themselves; it's if a major AI capex miss forces hyperscalers to write down data center assets, triggering forced selling in credit markets where hedge funds are now levered 10-15x on basis trades. That's the contagion vector. The inflation warning feels recycled—energy shocks fade, and the article ignores that US wage growth is already decelerating.
The BIS may be overstating circular-deal systemic risk: these arrangements are disclosed to lenders and rating agencies, and hyperscalers have fortress balance sheets that can absorb capex cuts without triggering cascades. Meanwhile, if AI actually delivers 20%+ productivity gains, current valuations look cheap, not bubbly.
"Circular AI financing opacity plus fiscal and inflation pressures raise the odds of a disruptive risk repricing larger than past equity corrections."
BIS warnings on AI capex turning into a bust via circular financing deals, combined with inflation persistence and leveraged hedge-fund basis trades in sovereign bonds, point to amplified downside if risk repricing hits. Poor disclosure of multi-pledged assets and exit clauses in data-center leases creates hidden leverage that could propagate faster across banks and non-banks than in prior cycles. The report's timing ahead of Sintra and recent US CPI prints above 2% for five straight years suggests central banks face pressure to stay hawkish, hitting growth assets even if energy shocks ease.
AI hyperscalers already show revenue traction from existing models, so capex may deliver returns fast enough to avert the sudden pullback BIS describes, unlike the 2008 credit mismatch.
"Near-term risk is funding liquidity driven by rate tightness and ROI timing, not just hidden leverage in circular financing."
Grok, your focus on hidden leverage and exit-clause risk needles at the periphery, but the near-term danger is liquidity: a rate-driven funding squeeze could hit hyperscalers and their lenders before any asset write-downs materialize. Circular financing may be disclosed, but its systemic squeeze depends on ROI timing, not just leverage. If cash flows disappoint, the drumbeat in credit markets hits first, cascading into private markets and shadow banks.
"AI capex is being used as collateral for structured credit, creating a systemic contagion risk beyond just hyperscaler balance sheets."
Claude, your dismissal of circular financing as 'confined' overlooks the derivative exposure. If hyperscalers like MSFT or GOOGL face revenue impairment, it isn't just their balance sheets at risk; it is the collateral quality of the structured credit products held by non-bank lenders. We are seeing a feedback loop where AI capex is being treated as high-grade collateral. If that valuation shifts, the liquidity squeeze Grok mentions will be exacerbated by margin calls on that very collateral.
"Structured credit contagion requires illiquid collateral and forced selling; public-market capex disappointment hits equities directly, not via derivative feedback loops."
Gemini's derivative-exposure angle is sharp, but conflates two markets. Structured credit products backed by hyperscaler collateral exist, yes—but they're mostly held by sophisticated LPs who mark-to-market continuously. The real contagion vector is simpler: if MSFT/GOOGL capex ROI disappoints, they cut capex, reducing demand for NVDA chips and data-center services. That hits semiconductor and infrastructure stocks directly, not via collateral cascades. The margin-call feedback loop requires illiquid collateral and forced selling; public equities don't work that way. Where's the actual leverage trap?
"Cross-default clauses in data-center leases turn capex cuts into simultaneous credit repricing for non-banks."
Claude, the leverage trap is embedded in data-center lease structures with multi-pledged assets and cross-default clauses. A hyperscaler capex cut can trigger simultaneous repricing across non-bank lenders holding those claims, even if public equities mark continuously. This creates a faster propagation channel into credit markets than pure equity demand destruction for NVDA, amplifying the liquidity squeeze before any impairment shows in filings.
The panel consensus is bearish, warning of potential risks in the AI capex cycle, including circular financing, lack of transparency, and leverage in the system. They agree that a downturn could lead to a liquidity squeeze and contagion in credit markets.
None identified
Liquidity squeeze and contagion in credit markets due to leverage and circular financing in AI capex