Hyperscaler AI Capex Hits $1.1 Trillion, Fueling Chip Stock Rally

Google, Amazon, Microsoft, and Meta have poured a combined $1.1 trillion into capital expenditures since the start of the AI boom in 2023 through June 2026, according to a Financial Times analysis. The four hyperscalers plan to spend $745 billion this year alone, underscoring an unprecedented buildout of data centers and AI compute capacity.

The spending spree has directly boosted semiconductor stocks. On Thursday, chip shares rallied sharply after strong earnings from Microsoft and Lam Research. Lam Research closed up 18 percent, Micron gained 18 percent, Sandisk rose 26 percent, Arm added 7 percent, AMD climbed 13 percent, and Intel jumped 11 percent, per CNBC. In Seoul, SK Hynix and Samsung Electronics surged more than 20 percent on Friday, reversing a sell-off earlier in the week.

Cloud Hosts Reap Rewards While Chipmakers Face Scrutiny

Investors have rewarded cloud providers for their AI spending, but the calculus for chipmakers is more complex. Amazon, for instance, is not slowing its data center investments, and the market has largely shrugged off concerns about overbuild, according to TechCrunch. The Next Platform questioned whether AI infrastructure spending forecasts are too rosy, noting that the pace of investment may outstrip near-term demand from enterprise customers.

• Hyperscaler capex from 2023 through June 2026 totaled $1.1 trillion, with $745 billion earmarked for 2026.
• Lam Research, Micron, and Sandisk each saw double-digit percentage stock gains on Thursday.
• SK Hynix and Samsung shares surged over 20 percent in Seoul on Friday.
• Microsoft and Lam Research earnings triggered the rally, with Microsoft reporting strong cloud and AI revenue.

Analysts Warn of Potential Oversupply as Spending Accelerates

The scale of investment raises questions about return on capital. While cloud providers can monetize AI through services like Azure OpenAI and AWS Bedrock, chipmakers face a more volatile demand cycle. The Next Platform analysis suggests that if AI workloads fail to grow as quickly as expected, the glut of GPU and memory capacity could pressure margins. However, the immediate market reaction has been bullish, with investors betting that hyperscaler spending will continue to drive chip demand through 2027.

What comes next hinges on enterprise adoption. If businesses broadly deploy AI applications, the current spending trajectory may prove justified. If not, the chip rally could reverse as quickly as it began. For now, the hyperscalers show no signs of easing their buildout, and semiconductor suppliers are riding the wave.