Global financial markets are currently experiencing a rare, synchronized sell-off across asset classes. Its defining feature is not a valuation reset in any single sector, but a systemic contraction in liquidity under extreme stress. Unlike previous bouts of panic, when money flowed into gold or U.S. Treasuries as safe havens, this sell-off has seen stocks, cryptocurrencies, precious metals and even industrial metals fall in tandem, alongside a weaker U.S. dollar and lower Treasury yields (meaning higher bond prices). This unusual combination, in which there is “no buyer of last resort,” reveals a key fact: money is not rotating between assets, but is being pulled directly out of risk assets and flowing back into cash or highly liquid reserves. Take Bitcoin, for example. Although it has long been viewed as an uncorrelated asset, it has recently shown a correlation close to 1 with U.S. tech stocks, largely due to a chain reaction driven by leverage. When accounts holding highly valued AI stocks and crypto assets face margin calls, investors tend to sell the most liquid assets first—those that can be converted to cash around the clock—rather than gold, which is costly to hold, or less liquid physical assets. James Butterfill, head of research at CoinShares, noted that the crypto market’s decline was driven more by a liquidity crisis triggered by the sell-off in AI stocks than by deteriorating fundamentals within the industry. This mechanism means that once the same collateral has been amplified through three layers of leverage—repos, margin and futures—a stronger dollar or higher financing costs can trigger forced liquidations that instantly wipe out purchasing power on paper, and that money does not automatically flow into other asset classes. The current market is therefore not simply a sector rotation, but a liquidity deleveraging process driven by a strong dollar and rising real interest rates. The slight rise in crude oil prices, against the trend, also supports this view: if the sell-off were driven by fears of a demand-side recession, oil prices should be leading the decline. Their relative resilience suggests instead that the main cause is tightening financial conditions, not a collapse in economic fundamentals.
Three forces are driving the current shock: expectations for macro policy, cracks in the AI narrative, and mechanical selling. First, persistent inflation has made the repricing of rate-hike expectations the key variable. Core CPI rose 0.3% month over month for the third consecutive month in October, while PPI rose 6.5% year over year, its highest level in nearly four years. In addition, Philadelphia Fed President Paulson publicly said he was open to consecutive rate hikes, prompting markets to sharply raise expectations for tighter policy. As of November 20, futures markets priced in a 90.2% probability of a 25-basis-point hike in December, taking rates to the 4.00%–4.25% range. Rising rate expectations directly increase the opportunity cost of holding non-yielding assets such as gold, reduce the present value of future cash flows for highly valued growth stocks, and drain global dollar liquidity, weighing on all risk assets. Second, cracks have appeared in the AI sector’s narrative. Anthropic’s IPO did not produce the expected market-clearing effect; instead, its shares opened sharply higher before retreating on the first day and fell 8.5% the next day, draining huge sums from the market. At Nvidia, Jensen Huang’s resignation as CEO and director, the board’s refusal to let him return, and media reports about a regulatory review of its accounting practices have shaken investors’ confidence in the certainty of its long-term growth. Richard Saperstein of Treasury Partners described this phenomenon as a repricing of AI assets, reflecting the reality of slower data-center construction, tighter project financing channels, and lower earnings expectations for tech companies. Finally, mechanical selling has created a self-reinforcing negative feedback loop. Across the crypto market, $1.893 billion in positions were liquidated over 24 hours, with 292,000 traders forcibly liquidated; long positions accounted for 91%. U.S. spot Bitcoin ETFs saw net outflows of $3.468 billion from November 2 to 19, with IBIT recording a record single-day outflow of $541.1 million on November 19. Notably, the nature of ETF outflows has changed fundamentally: the basis-trading strategy commonly used by hedge funds in the past—long ETFs, short futures—has become unattractive because annualized futures spreads have compressed to around 5%, offering little risk premium compared with the 10-year Treasury yield of 4.9%–5.15%. As a result, these direction-neutral funds have systematically withdrawn. Meanwhile, Goldman Sachs’ Panic Index rose to 9 out of 10, trend followers sold a combined roughly $57 billion in stocks over the past month, and the VIX closed at 35.89 after briefly touching 40.03 intraday. Together, these three forces point to one overarching switch: a stronger dollar and rising real interest rates.
To determine when this shock will bottom, three key gates need to be watched. The first is the completion of deleveraging, with a time frame of several days to two weeks. This will mainly determine whether the market gets a technical breather. The process depends solely on positioning, not fundamentals. Watch for the VIX retreating from elevated levels to the 20–25 range and its term structure returning to normal; open interest in crypto markets contracting and spot ETFs shifting from net outflows to net inflows; the S&P 500 reclaiming its 200-day moving average and CTA strategies shifting from net short positions to covering; and Goldman Sachs’ Panic Index falling back from extreme levels. Although the liquidity drain caused by Anthropic has passed, and seasonal factors after Thanksgiving, corporate buybacks, and positions that were heavily cleared out earlier all create conditions for a short squeeze, while Goldman Sachs, Morgan Stanley, and other institutions have also flagged the possibility of a year-end rally, BofA stresses that the breadth rule’s trigger signals only short-term relief for oversold stocks, not a trend reversal. Therefore, the current rebound is more a technical recovery after leverage has been flushed out than a fundamental reversal. The second gate is the repricing of the policy path, with the Fed decision on December 9 as the key event that will determine the market’s direction. Major data releases ahead of the meeting include October PCE and the second estimate of Q3 GDP on November 25, November nonfarm payrolls on December 4, and the FOMC decision and dot plot on December 8–9. The probability of a rate hike in December has fallen from 100% a week ago to 90.2%, suggesting that expectations are beginning to crack. A bottoming would require one of two paths: either weaker data cause rate-hike expectations to fade and the dot plot no longer signals further tightening in 2027; or hawkish policy is delivered as expected, but the bad news is fully priced in and the statement adds no further signals of tightening. BofA’s macro confirmation checklist includes the 10-year Treasury yield falling below 4.5%, WTI crude dropping below $70, and the dollar index peaking. If these indicators improve in tandem, that would signal that rate-hike concerns have peaked. The third gate is validation of returns on AI capital, over a time frame of several quarters. This will determine whether the current decline is a correction or a bear market. If this is merely a liquidity shock, the market can stabilize once deleveraging ends; but if returns on AI capital spending are disproven, the risk of a structural bear market looms. Evidence to distinguish between the two lies in credit and cash flow: watch high-yield bond spreads, cloud providers’ capital-spending guidance and free cash flow, the availability of data-center financing, Nvidia’s succession plans and the outcome of investigations, as well as changes in the ARR growth trajectory and computing costs as a share of revenue in Anthropic’s upcoming quarterly reports. UBS has already warned clients to watch for the risk that AI concerns could evolve into a cross-asset growth scare.
Follow me so you don’t miss the next quick market update.