Why Big Tech's $700 billion AI splurge is misleading

Big Tech's expenditures on artificial intelligence are on the rise, but the reasons behind this increase may not be as straightforward as they seem. Industry leaders have pointed to escalating component costs, particularly for memory chips, as a significant factor driving up capital investments in data centers and other infrastructure. Meta CEO Mark Zuckerberg addressed this issue during a recent analyst call, indicating that their forecast for infrastructure capital expenditures this year has been adjusted primarily due to soaring component prices. Similarly, Microsoft CFO Amy Hood revealed that a substantial portion of the tech giant's anticipated $190 billion in 2026 capital expenditures—approximately $25 billion—can be attributed to these rising costs. Amazon's CEO Andy Jassy also highlighted the challenge, stating that while the company did not revise its capex forecast, it is actively working to manage skyrocketing memory costs. As demand for AI technologies intensifies, the supply of essential components is being stretched thin, leading to significant price increases. Research firm TrendForce predicts that DRAM prices could rise by as much as 63% in the second quarter of 2026, while NAND flash prices may see an increase of up to 75%. These two types of memory chips are critical for data storage, and their rising costs are reshaping the narrative surrounding Big Tech's investment strategies. To illustrate, consider a scenario where a company purchases 100 AI components at $1,000 each, totaling $100,000 in expenditures. If the price per component increases by 25%, the total cost would rise to $125,000 without any increase in capacity. This scenario underscores how inflated pricing can significantly distort the perception of growth in AI investment. An analysis from BNP Paribas reveals that Microsoft's capital expenditure plans exceeded Wall Street estimates by $32 billion, yet $25 billion of that increase is linked to higher component prices rather than a more aggressive expansion strategy. Similarly, Meta's $10 billion increase in capex is likely driven by higher memory costs. This trend was previously noted by RBC Capital analysts, who estimated that rising memory prices could account for approximately 45% of the total capex increase among leading cloud providers this year. As a result, the ongoing AI investment boom may not reflect an actual acceleration in capacity building, but rather an industry grappling with the implications of escalating component prices, highlighting memory costs as a crucial challenge for future growth.