Four hyperscalers plan to pour $720 billion to $745 billion into capital projects in 2026, a spending wave that ripples through every layer of the AI chip supply chain.
Back
Hyperscalers' $745B AI capex fuels semiconductor supply chain demand
Four hyperscalers plan to pour $720 billion to $745 billion into capital projects in 2026, a spending wave that ripples through every layer of the AI chip supply chain.
Amazon, Alphabet, Microsoft and Meta plan to spend $720 billion to $745 billion on capital expenditures in 2026, directing a large share toward AI infrastructure that sustains demand across the semiconductor supply chain. Gartner projects worldwide AI-related spending will rise 47 percent to $2.59 trillion in 2026, with the expansion reaching beyond cloud providers into enterprise deployments across healthcare, finance and manufacturing.
"This is really the first time that technology chips have become an investable asset class," Nvidia Chief Executive Jensen Huang said in an interview with CNBC. The four hyperscalers' combined outlays follow more than $200 billion in capital expenditures during 2025, with Nvidia controlling roughly 80 percent of the data center GPU market and gross margins near 78 percent.
The spending cycle benefits chipmakers across the stack — from Nvidia's accelerators and Micron's high-bandwidth memory to Texas Instruments' power-management chips and FormFactor's test equipment — with consensus estimates pointing to triple-digit earnings growth at several suppliers. Nvidia has moved beyond selling chips to financing the infrastructure that runs them, partnering with Goldman Sachs, Apollo Global Management, BlackRock, Blackstone, Brookfield and KKR to establish independent financing platforms targeting more than $500 billion in third-party capital for data centers, power and computing capacity.
Each gigawatt of AI infrastructure could cost $50 billion to $60 billion to build, including energy, land, power, data-center shells and computing equipment, Huang said. Nvidia estimates the United States alone could require more than 70 gigawatts of power to support AI infrastructure. BlackRock Chairman and Chief Executive Larry Fink called the $500 billion target an initial step, saying the industry will need to "raise trillions of dollars over the coming years." Goldman Sachs Chief Executive David Solomon said the United States holds about $9 trillion in money-market funds and more than $100 trillion in equities, giving capital markets the capacity to finance a significant portion of the buildout.
Memory, Testing and Power Ride the Capex Wave
Micron Technology is among the clearest beneficiaries because AI data centers require substantially more memory than traditional systems. Demand for HBM3E and next-generation HBM4 memory, essential for high-end AI processors, is rising as models grow larger. The Zacks Consensus Estimate for Micron's fiscal 2026 revenue and earnings per share points to year-over-year increases of roughly 247 percent and 791 percent, respectively.
FormFactor supplies semiconductor test and measurement solutions used throughout chip manufacturing. As AI accelerators and advanced memory become more complex, testing requirements multiply. Consensus estimates for FormFactor's 2026 revenue and EPS indicate year-over-year growth of about 29 percent and 135 percent.
Texas Instruments, less obvious as an AI play, supplies analog and power-management chips that regulate electricity in data-center infrastructure. As computing workloads rise, efficient power management becomes more critical. Consensus estimates for Texas Instruments' 2026 revenue and EPS suggest year-over-year growth of approximately 23 percent and 55 percent.
Nvidia's Financing Push Expands the Addressable Market
Nvidia's Blackwell and Vera Rubin platforms are designed to deliver substantially greater performance for increasingly demanding AI workloads. Consensus estimates for Nvidia's fiscal 2027 revenue and EPS indicate year-over-year increases of about 80 percent and 91 percent. The company trades at more than 45 times forward earnings, leaving limited margin for disappointment if infrastructure spending slows.
The financing model carries risks. GPUs can become technologically outdated faster than power plants or data-center buildings, though older chip generations can continue generating revenue, allowing computing assets to be financed and potentially securitized. KKR's global head of digital infrastructure, Waldemar Szlezak, described the AI infrastructure opportunity as a "generational investment opportunity" spanning power, data centers and the compute layer. Taiwan Semiconductor Manufacturing, the sole foundry for Nvidia's most advanced accelerators, remains a single-source risk if geopolitical tensions disrupt production.
For investors, the capex wave supports a broad set of semiconductor names beyond the GPU leaders. Micron, FormFactor, Texas Instruments and Nvidia each hold strong technology positions with long-term growth drivers tied to hyperscaler spending. The key question is whether current valuations already discount the multi-year buildout — Nvidia at 45 times forward earnings versus AMD at roughly 28 times — or whether the infrastructure cycle still has room to run.
This article is for informational purposes only and does not constitute investment advice.
