The market is pricing two futures at once

Pricing two futures at once

The tape feels like a market caught between analysts’ wary sermons and spreadsheet reality.

We entered the week with China closed for the Lunar New Year while U.S. desks were returning overnight from a long weekend, leaving the order book thin and conviction apparently even thinner. The Goldilocks glow we left on Friday, with a solid jobs report earlier in the week and cooler inflation in the Friday CPI, never really caught fire again. The majors finished barely changed today. A headline about constructive US-Iran talks steadied the index, but we spent the day doing what markets do best when clarity is lacking. We spun in place. A 0.10 percent gain is not a rally. It is an index going nowhere with good posture.

Under the surface, however, the real story was not the index but the internal cross currents. Financials and Real Estate quietly caught a bid, the rate-sensitive complex behaving as if duration is still a friend. Consumer Staples and Energy lagged, crude cracked, gold and silver rolled over, and Bitcoin extended its February hangover. That is not classic risk-on or risk-off. That is a real-time market repricing of assumptions.

The assumption under review concerns artificial intelligence.

Software names were clubbed again, while the mega-cap generals managed to hold formation. The market split the AI trade into two camps overnight. The infrastructure kings are being granted some patience. The application layer is being forced to prove it still deserves to be in the game. It is a brutal divide.

What makes this episode so unstable is that investors are wrestling with two fears that point in opposite directions.

The first fear is creative destruction. If AI is as powerful as advertised, then entire industries are at risk of being hollowed out. Wealth managers, insurance brokers, logistics firms, and even parts of real estate services have been treated like ice cubes in a warming room. Investors are selling first and asking questions later because the cost of being wrong on disruption is existential. No one wants to own the next casualty of code.

The second fear is almost the mirror image. What if the hundreds of billions pouring into AI capex from Amazon, Microsoft, Alphabet and Meta do not generate a timely return? What if the hyperscalers are building cathedrals before there is a congregation? In that world, margins compress, free cash flow gets deferred, and the most crowded trade in the market suddenly looks like an expensive science project.

So we have a market that simultaneously believes AI will destroy everything and, at times, deliver nothing. That tension is why single stocks are being whipsawed like penny names even though we are talking about trillion-dollar balance sheets. It is also why the S&P 500 can erase a 1% drawdown and still feel heavy into the close.

AI was the market’s patron saint six months ago. Today it is being priced like a loaded weapon left on the table. The irony is that earnings from many AI-exposed companies have not collapsed. But it is hard to disprove a negative. You cannot easily prove that disruption will not happen, nor can you easily prove that capex will pay off. That uncertainty becomes a meaningful barrier to re-engagement. In markets, doubt is often more powerful than data.

Layer on top the policy fog, and the plot thickens.

There is a hawkish Fed camp whispering that if AI meaningfully boosts productivity, it could lift the neutral rate. Stronger structural growth implies a higher resting heart rate for the economy. In that framework, bonds are mispriced, and the market underestimates how high the policy floor ultimately is.

Then there is the Warsh argument, which flips the script. If AI is profoundly disinflationary, crushing labour costs and enhancing output, then the Fed could cut rates swiftly and deeply because the economy can run cooler without stalling. Productivity as a permission slip for easier policy.

And then there is the Daly camp, which in practice means humility. We simply do not know yet. Productivity data lags reality. Neutral rate is a theoretical construct. Policymakers are feeling their way through a fog bank while markets demand instant clarity.

This is why crude and crypto are cracking while mega-cap tech is being selectively reaccumulated. This is why gold can tumble even with duration well bid. This is why the index can finish flat while more than a trillion dollars of market value swings around underneath it over a fortnight. Simply put, everyone is working from a very foggy crystal ball.

The market is not confused because it is irrational. It is confusing because it seeks to price both a technological and a policy-regime shift simultaneously.

My read is simple. Until we get hard evidence that AI capex is translating into durable revenue acceleration or measurable productivity gains, volatility will remain sectoral and violent. The broad index can mask it, but under the hood, capital is being reallocated with a machete, not a scalpel.

This is not a bubble bursting. It is a narrative being stress tested.

When a theme moves from euphoria to scrutiny, price action becomes the battlefield. Right now, AI is no longer a one-way escalator. It is a high-wire act without a net. And in that environment, flat indices lie while dispersion tells the truth.