Key Metrics
| Metric | Value | Context |
|---|---|---|
| Hyperscaler CapEx (2026) | $725 Billion | Up 77% from $410B in 2025 |
| Amazon CapEx | $200 Billion | Up 56% YoY |
| Cloud Contract Backlogs | $460 Billion | Up from $240B end of 2024 |
| SanDisk YTD Gain | +464.5% | NAND flash demand surge |
| AMD YTD Gain | +114% | Outpacing Nvidia's +18% |
| Nvidia AI Chip Share | 81% | Still dominant in AI GPUs |
| Fervo IPO Pop | +35% | Geothermal for data centers |
| Power Capacity in Dev | 245 GW (U.S.) | Wood Mackenzie estimate |
May 17, 2026 — Wall Street has spent years talking about the AI boom. In 2026, the money is finally making the story impossible to ignore.
The biggest cloud companies are pouring staggering sums into the servers, chips, memory, and data centers needed to support artificial intelligence. Based on first-quarter earnings reports, Amazon, Microsoft, Alphabet, and Meta are on track to spend roughly $725 billion on capital expenditures this year, up 77% from 2025’s already record $410 billion. For investors, that spending wave is increasingly shaping one of the market’s clearest themes: AI infrastructure stocks may have far more room to run than the broader S&P 500.
Unprecedented Capital Deployment Signals a Multi-Year Growth Cycle
This is not a typical burst of tech spending. It looks more like a foundational rebuild of the digital economy.
Amazon alone plans to spend $200 billion in 2026, up 56% from last year, as demand for AI computing continues to outstrip supply. Across the hyperscalers, the message is consistent: this money is being committed to meet real customer demand, not vague future expectations. Cloud contract backlogs have climbed to about $460 billion, up from $240 billion at the end of 2024, giving suppliers unusually strong visibility into future revenue.
That matters because AI infrastructure is no longer a niche corner of the market. The buildout now reaches across semiconductors, storage, networking, power, and cooling. It is also becoming large enough to influence the wider economy, with some analysts arguing AI-related investment has been a key support for U.S. growth amid broader headwinds.
Key Insights
The trade is broadening: SanDisk (+464%) and AMD (+114%) are outperforming Nvidia (+18%), signaling investors are rewarding the wider ecosystem beyond GPUs.
Power is the new bottleneck: Data center energy demand is set to double by 2030, driving fresh capital into geothermal, nuclear and grid infrastructure.
Memory is cyclical: NAND flash demand is surging now, but memory markets have a history of boom-bust cycles that could reverse sharply.
Valuation risk is real: At 56x sales multiples and expanding multiples across the sector, any spending slowdown could trigger sharp corrections.
Semiconductor and Memory Stocks Are Leading the Market
The clearest winners so far have been the companies supplying the core hardware behind the buildout.
SanDisk has become one of 2026’s standout performers, surging 464.5% as demand and pricing for NAND flash memory accelerated. The company, which became a standalone business after its separation from Western Digital in early 2025, reported second-quarter revenue of $3.03 billion, up 31% sequentially, with year-over-year growth of 61% in its latest quarter. Its rise underscores a broader shift in the AI trade: investors are looking beyond GPUs alone and rewarding other critical parts of the stack.
That shift is also visible in semiconductors. AMD has gained 114% year to date, well ahead of Nvidia’s 18% advance, even though Nvidia still holds an estimated 81% share of the AI chip market. The divergence suggests investors are broadening their bets, favoring companies that can offer cost-effective alternatives as data center operators juggle both explosive demand and margin pressure. Nvidia remains central to the story, but the market is increasingly rewarding the wider ecosystem around it.
The Broader AI Infrastructure Ecosystem Is Gaining Momentum
The boom does not stop at chips and memory. Energy suppliers, cooling specialists, and networking companies are also attracting fresh investor attention as AI data centers push against physical limits, especially power availability.
That pressure helps explain the enthusiasm around companies like geothermal startup Fervo, whose IPO jumped 35% after it raised $1.89 billion in an upsized offering. Data center energy demand is projected to double by 2030, according to Gartner, while Wood Mackenzie estimates roughly 245 gigawatts of U.S. power capacity is already in development or planning.
The market has responded accordingly. Semiconductor shares have helped drive major indexes to fresh highs, reinforcing the idea that AI infrastructure is not just a technology story, but a market leadership story.
Forward-Looking Implications for Investors
The bull case is straightforward: hyperscaler spending is enormous, customer demand appears real, and the need for AI infrastructure is broadening well beyond a handful of chipmakers.
Still, investors should not confuse a powerful trend with a risk-free one. Memory markets remain cyclical, valuations have expanded sharply, and the sheer scale of the current buildout raises the possibility of oversupply in some segments down the line.
Even so, the larger shift looks structural. Unlike past tech cycles driven mainly by consumer adoption, this one is rooted in the physical backbone of computing itself. For retail investors, that may be the most important takeaway. The AI race is no longer just about which app wins. It is about who supplies the roads, rails, and power lines—and in 2026, that is where the market’s momentum is strongest.