Why Amazon is raising AI spending to $220 billion despite soaring costs
AWS cloud sales beat forecasts, but memory expenses force bigger investment bet

Amazon's decision to raise 2026 capital spending to $220 billion reveals how aggressively the company is doubling down on AI infrastructure, even as the cost of that bet climbs sharply. The hike explicitly reflects higher memory expenses, a direct acknowledgment that building the chips and systems for generative AI has become more expensive than prior forecasts suggested. Yet Amazon is willing to absorb those cost pressures because its core cloud business is delivering results that justify the outlay.
AWS revenue grew 37 percent year over year, according to Bloomberg reports, outpacing analyst expectations of 31 percent growth. That performance marks the fifth consecutive quarter of accelerating cloud sales, a streak that suggests corporate demand for AI-capable infrastructure remains robust. The stock market rewarded the narrative: Amazon shares jumped roughly 10 percent to $258 in after-hours trading, their highest level since early June.
CEO Andy Jassy framed the spending surge to investors as the cost of winning in AI. By committing to $220 billion in capex, Amazon is betting that the memory chips and data-center capacity it builds today will remain in shortage tomorrow, allowing AWS to capture pricing power and market share from rivals still ramping their own cloud and AI offerings. The scale of the commitment signals confidence that cloud demand will not crater, even as inflation in semiconductor and energy costs persists.
For readers watching big-tech strategy, this matters. Amazon's willingness to spend its way into AI leadership, despite rising component costs, suggests the company expects years of sustained demand for compute rather than a near-term glut. It also illustrates how cloud-infrastructure plays have become inseparable from AI bets, companies no longer spend on data centers; they spend on AI data centers. The margin question is no longer whether AWS grows, but whether growth at these scales can absorb billion-dollar cost increases and still deliver shareholder returns.


