Alphabet, Microsoft, Amazon.com, and Meta Platforms have already committed to spending hundreds of billions of dollars on artificial intelligence as demand for the tech continues to grow.
But building out the infrastructure needed to power AI requires massive amounts of money, and all of that spending has begun to raise investor eyebrows.
Still, even more AI spending is expected to be reflected in the earnings reports of the big tech companies, which will kick off with Alphabet on July 22.
Megacap tech companies were at one point free cash flow generating machines. Wall Street initially celebrated these companies taking that capital and spending it on AI, but that optimism has since shifted to uncertainty as spending doesn't seem to be slowing down anytime soon. Investors are constantly looking for proof that those hefty investments are leading to valuable returns.
As investors look for tangible proof that AI spending is paying off, experts believe that capex expectations will go even higher through 2026.
"We believe hyperscalers will prioritize capacity availability over near-term FCF optimization, as demand for AI training, inference, cloud workloads and internal AI product deployment remains supply constrained, " BofA Securities analyst Justin Post wrote on Tuesday.
Alphabet will be the first megacap tech company to report earnings. Analysts surveyed by FactSet expect second-quarter capex to be $44.9 billion, which would be a 100% increase from the $22.4 billion in the prior year period.
Alphabet will be followed by Microsoft on July 29 and Apple on July 30. Meta and Amazon have yet to officially announce what day they will report.
"We do think capex probably does have an upwards tilt since there are continuous higher component costs, and we saw that Micron will be raising prices in Q3, and they have continuous pricing power," says Ken Mahoney, CEO of Mahoney Asset Management.
Memory is a critical component used to help power AI. Demand for memory and other AI hardware is far outpacing supply, which has led to a massive rise in costs. That's been a boon for Micron and SK Hynix, but has put pressure on customers that need to buy the hardware as they build out data centers.
Goldman Sachs analyst Eric Sheridan agrees that elevated component costs are going to lead to higher capital spending in the near-term.
Sheridan raised his capex estimates for Amazon Web Services on Tuesday due to the "continued constraints between the demand for AI and the wider supply of compute coupled with an inflationary pressure around key capex inputs (overall data center construction, 3rd party chips & memory)."
Sheridan, who rates Amazon as a Buy with a $335 price target, now expects the e-commerce giant to spend a cumulative $827 billion between 2026 and 2028.
This could be a problem for hyperscaler stocks as rising spending puts pressure on margins and free cash flow. However, Post believes there's reason for optimism ahead.
"New AI-driven revenue streams could improve investor sentiment by providing clearer evidence that elevated infrastructure investment is translating into incremental revenue," he said.
Comments