Why the two most important events of 2025 have been completely shrugged off by the stock market

The two biggest catalysts for US stocks this year — the DeepSeek-inspired rout that began to tip over the AI momentum trade and the Rose Garden reciprocal tariffs announcement that accentuated recession fears and crippled consumer-facing stocks — are so far shaping up to be complete nothingburgers in the eyes of the market.

The iShares MSCI USA Momentum Factor ETF, which includes many AI market darlings, ended Thursday a whisker shy of all-time highs. And a basket of consumer stocks compiled by Goldman Sachs as being particularly vulnerable to tariffs has erased its drop since the April 2 Rose Garden tariff announcement, and then some.

But the circumstances surrounding these respective about-faces appear to be completely different. For AI stocks, it’s that the sum of all fears added up to zero; for consumer stocks, it’s that the sum of all fears turned from something unquantifiable into something traders could wrap their heads around.

It’s amazing to see how there was virtually no break whatsoever in the upward trend in earnings estimates for stocks that play different roles in the AI data center supply chain, like Nvidia, Arista Networks, or Constellation Energy, following the DeepSeek realization moment or the increase in recession fears that came along with tariff talk and action.

At the time of the massive sell-off across the tech complex in late January, I wrote that DeepSeek was next year’s nightmare for Nvidia, today, in that it gave reason to question the magnitude and duration of the capex binge that has been oh-so-beneficial for the chip designer’s bottom line. At some point, the bumper growth in profits driven by the boom would diminish somewhat, and DeepSeek offered a reason to think that might happen sooner rather than later.

In taking stock of the situation, Jefferies analyst Edison Lee flagged that tech companies faced a fork in the road: “1) still pursue more computing power to drive even faster model improvements, or 2) refocus on efficiency and ROI, meaning lower demand for computing power as of 2026.”

What’s happened since? Meta actually upped its already ample 2025 capex budget. And all the concern about Microsoft pulling away from some data center opportunities may be a case of missing the $80 billion forest for the trees. 

The earnings and capex revisions so far show DeepSeek was only ever a threat to tech in narrative terms, not fundamental ones.

Megacap tech companies are taking the more trodden path, and that has made all the difference. Jevons Paradox rules, DeepSeek-esque efficiency campaigns drool, and the nightmares have been fictions, not realities.

As earnings estimates have gone up while the price is off the peak, most of these AI-linked momentum stocks are actually cheaper (based on forward price-to-earnings ratios) than they were in late January — particularly as other AI-adjacent names that aren’t really a part of the physical supply chain, like Palantir, have boomed.

By contrast, we’ve seen profit estimates for consumer-facing firms that would suffer under the weight of onerous tariffs come under the knife, before and after the imposition (and dialing down) of reciprocal tariffs. The likes of Nike and General Motors are more expensive than they were before the Rose Garden announcement.

But tariffs that started as a threat turned reality with a range of outcomes that could span the Pacific Ocean — an ocean that had far fewer cargo ships traveling from China to the US — and that range of possible outcomes has seemingly narrowed significantly.

“The UK has one of the least imbalanced relationships with the US and now has a universal tariff rate of 10%. China has one of the most imbalanced relationships and now has a tariff rate of 30%,” Deutsche Bank strategist George Saravelos wrote earlier this week. “It is reasonable that these two numbers now set the bounds of where American tariffs will end up this year, a material increase in visibility from just last week.” 

It may seem odd on the surface, but it’s completely reasonable for stocks to see valuations increase at the same time as earnings get revised lower. That’s usually because a) the downdraft in the stocks that preceded the hit to earnings was the market beating analysts to the punch on this front, and b) with more confidence in being able to assess a potential trough in earnings, it’s easier to look through a weak patch (if you can deem it to be temporary).

Of course, the dust has not even half-settled on the 2025 macro outlook, and you can look at either part of these pockets of the market with a glass half-full or half-empty view based on the prevailing data.

Some AI-linked stocks are off their highs and cheaper? Well, surely safe to play for a return all the way back to the highs, right? AI stocks are off their highs and cheaper? Well, still, each passing day brings us closer to a time when aggregate AI-centric capex is going to inflect lower, and at that point they’ll all deserve lower multiples…

For tariff-sensitive stocks, you can either make the judgment that what’s been priced in with respect to the magnitude and duration of the hit to profits is appropriate, or not.

What’s common to the recoveries in both is the diminution of recession fears, largely due to the pausing of reciprocal tariffs and the trade truce with China. Tariffs have turned from an existential threat to potentially everything into a manageable, localized hit to profits in the market’s eyes, and that’s something you see priced in broadly across financial markets.

What these episodes and relative recoveries in these once battered, now surging parts of the market serve to underscore is that the AI momentum rally is the market rally. And this rally faces two threats: collapsing under the sheer scale of being unable to build on its past successes, or a broad downturn in the business cycle that leaves no company unscathed.

Those are both threats that have appeared and been dismissed in the first half of 2025.