The market runs back into AI despite higher Oil and yields

Back into AI

World stocks surged as traders made a spirited return to the market’s commanding centre of gravity, piling back into semiconductors, AI leaders and momentum, with the enthusiasm of a crowd rushing through the one door it still trusts.

The Nasdaq led Wall Street higher, semiconductor shares jumped more than 5%, and momentum staged its strongest rebound in years. The supposed broadening trade was shoved aside as capital returned to the same handful of companies that powered the original boom. Investors had spent several sessions talking about and dipping their toes into rotation, diversification and the virtues of looking beyond AI. But once the violent momentum selling pressure lifted, they went straight back to the old AI neighbourhood.

The rotation, it turns out, may have been more of a temporary evacuation than a permanent change of address.

The speed of the rebound says as much about positioning as it does about conviction. Momentum exposure had been cut hard, semiconductor sentiment had deteriorated sharply, and investors had spent much of the previous week reducing risk in the market’s fastest-moving names. The rubber band had been stretched so far that when the selling eased, it snapped back with enough force to drag the entire index higher.

Short covering added fuel. Negative gamma left dealers chasing the move rather than leaning against it, while zero-day options helped sustain the advance through the session. What began as buying in beaten-down chipmakers quickly became a scramble to rebuild exposure before the next wave of earnings.

That does not make the rally artificial. Positioning is part of the market, not something that sits outside it. But it does explain why the rebound was so violent even though the broader fundamental landscape had not changed nearly as much as the price action suggested.

The market had become too negative, too quickly, in the part of the equity complex that still carries the greatest long-term earnings promise. Once the selling became crowded, the same concentration that had magnified the decline worked in reverse.

The narrowness, however, remains impossible to ignore. The S&P 500 excluding the main AI leaders was broadly flat, leaving the index once again dependent on a small group of companies to do the heavy lifting. The market may wear the costume of diversification, but underneath it still moves like one enormous semiconductor position.

That is why the coming earnings season matters so much.

Nearly one-fifth of the S&P 500 by market value is due to report this week, with Alphabet providing the clearest test of whether the AI investment cycle still has enough momentum to justify the market’s enthusiasm. Investors are no longer satisfied with broad promises about artificial intelligence changing the world. They want revenue growth, stronger guidance and evidence that the billions being poured into data centres, chips and computing infrastructure can eventually produce returns.

The burden of proof has moved.

During the early phase of the boom, companies were rewarded simply for announcing larger investment plans. Capital expenditure was treated as a sign of ambition, and ambition was treated as a substitute for earnings. That equation is becoming less generous. A higher spending guide may still support chip demand, but it also raises questions about free cash flow, funding needs and how much of the future is already embedded in valuations.

Earnings season is where the AI story must stop selling blueprints and start showing tenants.

The equity market is still willing to admire the skyscraper. The credit market has begun to assess whether the foundations can support another twenty floors.

That divergence is becoming more important. Hyperscalers, data-centre operators and technology companies have stepped up borrowing at a remarkable pace, with jumbo issuance increasingly concentrated in the technology complex. Equity investors see expanding AI infrastructure and another leg of demand. Bond investors see a funding requirement that is growing faster, lasting longer and becoming more expensive.

The buildout is still in its early innings, but the financing burden is no longer a footnote. It is becoming part of the central investment argument. The AI boom was born in an environment where capital appeared abundant, and the cost of money was treated as manageable. The next phase may have to prove it can survive with oil above $90, Treasury yields pushing higher and credit markets becoming more selective.

That is where the macro backdrop begins to press against the equity celebration.

Brent crude climbed back above $90 as the US-Iran conflict entered another round of strikes and shipping risks spread across two of the world’s most important energy corridors. Hormuz remains under pressure, while renewed Houthi threats in the Red Sea have placed further strain on the Bab el-Mandeb route, the very passage Saudi exporters have relied upon when Gulf shipping becomes more difficult.

The oil market is no longer dealing with one blocked doorway. It is watching smoke gather around both exits.

Traffic through the Red Sea had already begun to fall before the latest threats, while attacks on tankers and regional infrastructure have added a fresh layer of uncertainty. The market is being forced to price not only the loss of barrels but the growing cost, delay and risk involved in moving them.

Oil has returned as the tax collector the market thought it had already paid.

The immediate effect is visible in the bond market. Treasury yields pushed back toward two-month highs, with the short end leading as traders increased the probability that the Federal Reserve may have to remain restrictive for longer. The softer inflation data that had briefly calmed the market last week was quickly pushed aside as higher energy prices reopened the debate over headline inflation and policy credibility.

Stocks were celebrating upstairs while the bond market quietly raised the rent.

That matters because the same technology companies driving the rebound are among the most sensitive to changes in the discount rate. A higher oil price alone does not end the AI trade. A sustained rise in oil prices, yields, and funding costs is more dangerous because it attacks the valuation structure from several sides at once.

The equity market can absorb Brent at $90. It has done so before. Brent nearer $120 would be a different conversation. At that point, oil would stop being treated as a geopolitical subplot and start becoming a global demand shock, a margin squeeze and a policy problem rolled into one.

For now, traders are still compartmentalizing the risk. AI earnings sit in one box, Middle East escalation in another, and bond yields in a third. The market can maintain that separation for a while. It becomes harder when the boxes begin leaking into one another.

The currency market is already showing signs of strain.

The dollar strengthened broadly as higher oil prices pushed US yields higher, while the yen fell below 163 per dollar for the first time in four decades. Japan is being squeezed by the combination of expensive imported energy, a cautious central bank and a widening credibility gap between policy intentions and market outcomes.

The yen has become the pressure gauge on a machine running too hot.

Intervention risk is now obvious, but intervention alone may buy only time. Tokyo can step into the market, slow the move and force speculative positions to retreat. It cannot permanently erase the yield differential, reduce Japan’s energy bill or restore confidence in the policy mix with a single burst of yen buying.

That is why the move through 163 matters. It is not merely another technical level. It is a warning that the market is beginning to test how much weakness Japanese officials are prepared to tolerate before words become action.

Elsewhere, the cross-asset picture was equally revealing. Gold rallied despite a stronger dollar and rising yields, while bitcoin also pushed higher. Oil rose, equities rose, and the dollar strengthened simultaneously.

Every asset class appeared to be carrying a different map of the same battlefield.

Gold was trading geopolitical and fiscal uncertainty. Bitcoin was trading liquidity and momentum. AI stocks were trading cleaner positioning and earnings hope. The dollar was trading rate differentials and energy stress. None of these moves was individually irrational, but together they showed that the market was no longer operating under a single clean macro regime.

That may be the defining feature of the current tape. Investors are not choosing between risk and safety. They are buying momentum, protection, scarcity and yield all at once.

The sharp rebound in momentum probably means the unwind has entered its later stages. Positioning is cleaner, the short side became crowded, and the long-term AI story remains intact. But a durable recovery will need more than a squeeze. Earnings must validate the spending cycle, credit markets must absorb the borrowing and oil must stop pushing bond yields higher.

The bulls have recovered the steering wheel, but the road has narrowed. On one side sits an oil shock. On the other sits a bond market losing patience.

For now, AI remains the fastest lane. It is no longer the safest one.