AI capital expenditure is draining market liquidity: a quiet "reverse QE"

By Plur Daddy

Translated by AididiaoJP, Foresight News

We are facing a fundamental shift in market dynamics, caused by capital expenditure cycles in the artificial intelligence sector leading to a shortage of financial capital.

This will have a profound impact on asset prices, after all, for a long time capital has been abundant. The Web 2.0 and SaaS models that drove market prosperity in the 2010s required very little capital, which led to excess funds flowing into various speculative assets.

Yesterday, while discussing the current market situation, I suddenly realized some things. This might be the most insightful article I’ve written in a long time. Next, I will gradually break down the underlying logic.

There are parallels between AI capital expenditure and government fiscal stimulus, which help in understanding their mechanisms.

In fiscal stimulus, the government issues government bonds, which are taken up by the private sector. The government receives funds and invests them into the economy. This money circulates within the real economy, creating a multiplier effect. Due to this multiplier, the ultimate impact on financial asset prices is positive.

In AI capital expenditure, large tech companies finance themselves through debt issuance or by selling government bonds (and other assets). The private sector takes on the long-term debt, and the companies then invest the proceeds into projects. These funds also circulate within the real economy and generate multipliers, positively influencing financial asset prices.

As long as there is idle capital in the economy, this process can run smoothly. Its effects are significant, capable of boosting the market comprehensively. This has been the case in recent years, where AI capital expenditure acted as an additional economic stimulus, boosting both the economy and the markets. But the problem is: once the idle funds are exhausted, every dollar invested in AI must be diverted from other areas. This will trigger a fierce capital competition. When capital becomes scarce, people are forced to carefully evaluate its most effective uses, and the cost of capital (i.e., market interest rates) rises accordingly.

I want to emphasize again: when funds are tight, there will be clear differentiation among assets. The most speculative assets will suffer disproportionate losses, just as they previously gained disproportionate gains during times of capital surplus and a lack of productive investment opportunities. From this perspective, AI capital expenditure actually plays a role of “reverse quantitative easing,” causing a negative rebalancing effect on investment portfolios.

Fiscal stimulus rarely faces this dilemma because the Federal Reserve usually becomes the ultimate buyer of government bonds, thus avoiding crowding out other capital uses.

The term “funds” here can be interchangeable with “liquidity.” The word “liquidity” is often confusing because it has many different meanings.

Let me give an analogy: funds or liquidity are like water. You need a higher water level in the bathtub for the floating rubber ducks (financial assets) to rise. There are several ways to achieve this: either increase the total amount of water (cut interest rates or implement quantitative easing), unblock the inflow pipe (such as current reverse repurchase operations—these “pipe unclogging” measures), or reduce the drain of the bathtub.

Currently, most discussions about liquidity in the economy focus only on the money supply. However, the demand for money is equally critical. What we are facing now is excessive demand, which causes crowding out effects.

Some media reports suggest that the world’s wealthiest investors—such as Saudi Arabia’s sovereign wealth fund and SoftBank—are nearly out of funds. Over the past decade, global investors have “feasted” on a large amount of assets. Let’s analyze what this means: when Ultraman reaches out to fulfill previous funding commitments, unlike in the past when funds were abundant, they now have to sell some assets first to raise cash. What will they sell? Likely those holdings with less confidence: some recently underperforming Bitcoin, SaaS stocks facing industry disruption, or redeeming underperforming hedge fund shares. To meet redemption demands, hedge funds also have to sell assets. Falling asset prices will undermine market confidence, tighten financing conditions, and trigger more sell-offs across sectors… This chain reaction will propagate through the financial markets layer by layer.

More complexly, Trump has chosen Wosh. This is especially concerning because he believes the current problem is too much money, when in fact it is the opposite. This is also why a series of market changes since his nomination have been accelerating.

I have been trying to understand why stocks like SNDK, MU—manufacturers of storage chips (DRAM / HBM / NAND)—perform far better than others. Of course, rising product prices are part of the reason. But more importantly, these companies are currently and recently highly profitable, even though everyone knows profits are cyclical and will eventually decline. When capital costs rise, discount rates increase as well. Long-duration, future cash flow-based speculative assets are suppressed, while assets that generate cash flows in the near term are favored.

In this environment, cryptocurrencies, as a liquidity-sensitive indicator, naturally suffer heavy losses. That’s why their recent decline seems bottomless.

Highly speculative retail favorite stocks struggle to maintain gains, and even sectors with improving fundamentals are finding it difficult.

Due to demand exceeding supply, yields on sovereign bonds and credit bonds are rising.

It is no longer possible to be blindly optimistic and solely bullish.