AI Capex Surge: US Cloud Firms Plan $660B-$690B Spend in 2026 Driven by Supply Constraints

Major US cloud providers are committing between $660 billion and $690 billion in capital expenditure for 2026. This figure nearly doubles their spending from the previous year, according to InfoWorld.

What happened

The five largest US cloud and AI infrastructure providers have collectively committed to spending between $660 billion and $690 billion on capital expenditure in 2026, according to InfoWorld. Amazon leads this investment with a projected $200 billion, primarily directed at data centres and AI compute. Alphabet has committed between $175 billion and $185 billion, a significant increase from $91 billion in 2025.

Microsoft is tracking toward $120 billion or more, having already spent $37.5 billion in a single quarter. Meta is targeting between $115 billion and $135 billion, while Oracle aims for roughly $50 billion. Four of these companies alone plan to invest about $630 billion, representing a 62% increase from the $388 billion spent in 2025.

Why it matters

The spending splits across the AI data centre stack, covering GPUs, custom accelerators, physical facilities and power infrastructure, according to InfoWorld. Roughly two-thirds of the money goes to compute, with one-third allocated to facilities, cooling and networking. The power share is rising every quarter as providers secure electricity through purchase agreements or on-site generation.

Hyperscalers report being supply-constrained rather than demand-constrained, with Microsoft holding an $80 billion backlog of Azure orders due to power availability, according to InfoWorld. Enterprises are buying cloud capacity now to avoid being left behind by the AI boom. Hardware vendors and hyperscalers expect sustained growth as long as this demand continues.

Next steps

• Monitor quarterly capital expenditure reports for spending trends
• Assess internal power and cooling capacity for new hardware
• Review supply chain constraints for AI compute components

Detection notes

Teams usually monitor resource utilisation metrics to identify sudden spikes in compute or storage consumption. Check whether unauthorised instances are being launched or if data egress patterns deviate from baseline norms. Look for alerts related to unusual network traffic or configuration changes in cloud environments. Containment moves typically involve isolating affected resources and reviewing access logs for anomalous activity. Verify that security policies are enforced across all new infrastructure deployments.

Background: Supply chain attacks

Prioritise binary verification and least-privilege execution for third-party code. Verify cryptographic signatures for every update, not just the initial install. Isolate vendor access to minimise lateral movement. These steps reduce the attack surface more effectively than perimeter firewalls alone.

Read the full guide: Prevent Supply Chain Attacks: Practical Controls for Third-Party Risk

What to do and how to stay safe: Amazon

• Review cloud spending reports for unauthorised resource creation
• Implement strict budget alerts and usage limits
• Audit access controls for new AI workloads
• Monitor for unusual data transfer patterns

General security guidance from the Packet Tribune newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

How much are US cloud giants spending on AI in 2026?

The five biggest providers committed between $660 billion and $690 billion in capital expenditure for 2026, according to InfoWorld.

Which company is spending the most?

Amazon leads with a projected $200 billion in capital expenditure for 2026, according to InfoWorld.

What are they buying with this money?

They are buying chips, buildings, electricity, GPUs, custom accelerators, and power infrastructure, according to InfoWorld.