Big Tech’s AI Numbers Failed to Impress. It’s But a Blip.

While the headline numbers were generally solid, they weren’t good enough to impress investors given the stocks’ big runs.

Microsoft
had the best quarter of the bunch, reporting earnings per share of $2.93, well ahead of the analyst consensus for $2.76.
Alphabet
 beat profit estimates, posting EPS of $1.64 versus the consensus of $1.59. AMD’s profit was in line with the estimates, but the company’s revenue outlook was disappointing.

All three stocks were down in mid-day trading Wednesday. Alphabet shares dropped 6%, AMD slipped 3%, and Microsoft was down 1.4%. The tech-heavy Nasdaq Composite was off 1.6%.

The main problem with the reports wasn’t the numbers but the expectations going in. Take AMD’s AI chip outlook. On last night’s conference call with investors, CEO Lisa Su said that AMD now expects revenue for its AI data center MI300 GPU products to surpass $3.5 billion in 2024—up from a $2 billion forecast just three months ago. While the guidance is up significantly, some Wall Street analysts had estimates of up to $8 billion.

Investors would be wise to largely overlook these day-to-day stock movements. The technology companies’ conviction over future AI demand is more important. And, given the latest commentary about capital expenditure budgets, the robust trend is intact.

Microsoft said its expects capex to “increase materially” in the current quarter, and it intends to invest aggressively in the coming quarters. Alphabet said its capex would be “notably larger” in 2024 versus the prior year. Both companies said infrastructure investments are being driven by trends in AI demand.

There’s other evidence the AI arms race is still on beyond the comments from Microsoft and Alphabet. On Monday, Super Micro—a leading independent manufacturer of high-end AI servers for data centers— easily beat expectations and raised its full-year revenue guidance by nearly 40%. Last week,
Nvidia
CEO Jensen Huang told reporters in Taiwan that demand for AI GPUs is still outstripping supply, while adding 2024 is going to be a “huge year.”

Finally,
Meta
CEO Mark Zuckerberg boasted on social media earlier this month that his company will have 350,000 Nvidia H100 GPUs—and almost 600,000 H100 equivalent GPUs based on total computing power—by the end of this year.

We’ll find out more when Meta,
Amazon,
and
Apple
report on Thursday, but all signs suggest that AI spending is still accelerating—no matter what stocks said on Wednesday.

This Week in Barron’s Tech

Write to Tae Kim at [email protected] or follow him on X at @firstadopter.