Nvidia’s earnings could impress — But not reverse AI worries

US and European equity markets rebounded yesterday as global investors digested the latest US tariff shake-up and the AI fear trade eased. LegalZoom, for example, which had been heavily hit by fears that Anthropic’s Claude could wipe out its business, jumped 2.5% from its lowest levels on fresh news that Claude unveiled features that could be integrated into the company’s products — making them more customizable and AI‑friendly.

That’s the brighter side of the medal: AI tools can enhance some of these software products — instead of replacing them — potentially helping these companies thrive alongside AI rather than disappear. The iShares Expanded Software ETF rebounded close to 2%.

The software stress is probably not over, but opportunities are clearly emerging.

In the Big Tech space, it was a good day too. Meta announced plans to buy AMD chips — a lot of AMD chips — to deploy up to 6 gigawatts of computing power. One gigawatt is roughly the output of a full-scale nuclear reactor and could power a small city. Meta is talking about the equivalent of six nuclear reactors — just for AI computing. Six gigawatts is enough to power 4–5 million homes. That’s the size of the deal Meta signed with AMD. The deal is worth tens of billions of dollars in revenue for AMD and pushed AMD’s stock price up by nearly 9% yesterday. That’s the good.

The bad is that Meta is spending money against investors’ will and taking on debt — something investors are growing increasingly uncomfortable with: more spending, increasingly on debt.

The ugly is that Meta is receiving warrants that would allow it to buy shares in AMD if the company hits certain performance targets. Meta could ultimately end up owning up to 10% of AMD. Again, the circular nature of AI deals attracts attention. Everybody’s hand is in everybody’s pocket, and if one of the companies stumbles, the whole group — increasingly running on debt — could wobble.

For now, Nvidia and AMD will continue to grow given the hyperscalers’ huge spending plans in AI infrastructure involving chips, despite investors’ reluctance. The risk is, what happens if the big spenders are forced — or decide — to spend less? Could a gap left by Big Tech be filled by smaller spenders? Time will tell. The world is migrating to AI, but the spending increase from outside Big Tech is obviously more granular.

Today, Nvidia will provide clarity on how many chips it sold last quarter and how much it earned. Q4 revenue is projected near ~$65.6–66.1 billion, which would represent nearly 70% year-over-year growth compared with $39.3bn earned in Q4 last year. Note that before the AI boom, this company’s quarterly revenue was around $6–7bn, and growth is expected to continue.

Now, all of this sounds great — but investors won’t just cheer the headline. Last quarter was a good reminder. Nvidia delivered strong top-line numbers, yet the stock didn’t fully ride the wave. Why? Because investors zoomed in on the details — specifically the widening gap between revenue booked and cash actually collected. And that matters in a world of rising leverage and massive AI capex. In this environment, investors don’t just want contracts. They want cash in the door.

So yes — Nvidia has built a track record of meeting and beating expectations in this AI cycle. But this time again, it won’t just be about revenue growth or EPS beats. The devil will be in the details — cash flow, receivables, margins and forward guidance. I’m afraid Nvidia alone may not be able to reverse the fears. Yes, AI spending remains strong, but investors are pulling back — not from Nvidia, but from its biggest clients. Given that hyperscalers make up almost 50% of Nvidia’s client book, the AI stress is likely not over just yet.

Broadly, Nvidia’s results will pretty much mark the end of the earnings season, and the numbers look strong. In Q4 2025, the S&P500 posted 13.2% earnings growth — its fifth consecutive quarter of double-digit earnings growth. Other good news is that the rally has spread to non-tech sectors as well, with the S&P500’s equal-weighted index mostly closing its gap with the tech-heavy, market-cap weighted version.

The S&P 500 looks toppish since the start of the year due to the AI fear trade — both in Big Tech and the software space. The latest AAII investor sentiment index confirms that the bears took the upper hand over the bulls for the first time since November — hinting that much of the selling may already be done. Still, the major US indices need technology to perform well, as the Mag 7 companies make up to a third of the index. Without their help, gains from rotation would be marginal.

Looking at the data, the Conference Board pointed to improved sentiment in February, while Home Depot warned that its customers — middle- and upper-income homeowners — are worried about housing affordability, job stability, and higher financing costs.

Meanwhile, Donald Trump, at the longest State of the Union speech ever, said America is “bigger, better, richer and stronger than ever before. ‘We’re winning so much…’” he added.

Alas, the US dollar weakened in Asia, pushing the EURUSD to the 1.18 level. The USDJPY retraced losses near 156, but Takaichi voiced apprehension regarding further Bank of Japan (BoJ) rate hikes— which could temper the impact of her expansionary fiscal plans aimed at boosting growth — and nominated two reflationist academics to join the BoJ policy board, suggesting that the yen will remain under pressure. Elsewhere, the AUDUSD jumped past 0.71 on stronger-than-expected CPI numbers, reviving Reserve Bank of Australia (RBA) hawks.

Today, euro area CPI figures are expected to confirm softening inflation that could allow the European Central Bank (ECB) to act if trade tensions flare and threaten economies. For now, the Stoxx 600 doesn’t look like it needs any help, trading near an all-time high, supported by stronger-than-expected economic data, and ignoring the fresh tariff uncertainty.