Big Tech’s AI spending is surging, though not necessarily because companies are building significantly more capacity.
Loading audio narration…
This week, industry executives pointed to the same underlying issue: sharply rising component costs, especially for memory chips, are inflating capital expenditures spent on data centers and other equipment.
“We are increasing our infrastructure capex forecast for this year. Most of that is due to higher component costs, particularly memory pricing,” Meta CEO Mark Zuckerberg told analysts on Wednesday.
Microsoft echoed that trend, with CFO Amy Hood saying that roughly $25 billion of its projected $190 billion in 2026 capex is tied to higher component prices.
Amazon didn’t raise its capex forecast at all this week. Even then, CEO Andy Jassy said memory costs have “skyrocketed,” and the company is trying to keep these costs under control.
Memory prices are soaring as AI demand strains supply. Research firm TrendForce expects DRAM prices to jump as much as 63% in the second quarter of 2026, while NAND flash prices could surge 75%. NAND and DRAM are different semiconductor technologies used to hold data.
That’s reshaping how investors should interpret Big Tech’s spending boom. Take a theoretical example: If you buy 100 AI components for $1,000 each, that’s $100,000 in capex. If the cost of each component goes up 25%, they now cost $1,250 each. Your capex is now $125,000, even though you’re adding zero extra capacity.
When stripping out this pricing impact, Big Tech companies may not have boosted their AI buildout plans much at all this week.
Take a look at this table from BNP Paribas. Microsoft’s capex plans exceeded Wall Street estimates by $32 billion. That seems big, until you consider $25 billion of that increase comes from higher prices, not a more ambitious buildout. Meta upped its capex forecast by $10 billion. Higher memory prices could account for most, or all, of that increase.
RBC Capital analysts spotted this phenomenon in early February. Back then, they estimated that higher memory prices could account for about 45% of the total capex increase by top cloud providers this year.
The result is a more nuanced picture of the AI arms race. While investment remains strong, the headline numbers may overstate the pace of expansion.
Instead, Big Tech is increasingly paying a premium to secure scarce components, highlighting memory pricing as a key bottleneck in the industry’s next phase of growth.
Sign up for BI’s Tech Memo newsletter here. Reach out to me via email at abarr@businessinsider.com.
