Hyperscalers Unleash $725 Billion AI Capex Wave
💡 Key Takeaway
AI infrastructure spending has gone parabolic, with hyperscalers committing over $700 billion in 2026 capex, creating massive opportunities for picks-and-shovels providers.
The $35 Billion Validation
Meta Platforms just signed a fresh $21 billion contract with CoreWeave, adding to a previous $14.2 billion commitment and locking in $35 billion of compute capacity through 2032. Crucially, this new deal focuses on inference workloads, signaling the industry's pivot from building models to running them at scale. On the same morning, Amazon CEO Andy Jassy disclosed plans for $200 billion in 2026 capex, with AWS's AI revenue run rate hitting $15 billion and the custom chip business generating $20 billion annually.
These announcements confirm JPMorgan CEO Jamie Dimon's forecast that five hyperscalers will collectively raise AI-driven capex from $450 billion in 2025 to $725 billion in 2026—a staggering 61% increase in a single year. The market is still digesting what $35 billion of locked, multi-year capacity from a single hyperscaler means for the AI infrastructure ecosystem.
Winners and Losers in the AI Arms Race
The massive capex commitments create clear winners: pure-play AI cloud operators like CoreWeave with $66 billion backlogs, chip suppliers like Nvidia seeing $1 trillion in cumulative demand, and hyperscalers like Amazon securing multi-decade moats with customer commitments. The bear case that hyperscalers would insource their AI infrastructure just collapsed—Meta wrote the $21 billion check instead of building it themselves.
This spending wave represents a fundamental shift from training to inference, where durable, high-utilization revenue lives. Companies positioned in the picks-and-shovels layer—providing compute, chips, and specialized infrastructure—stand to capture the majority of this $725 billion flood. The risk of dilution for companies financing buildouts is outweighed by unprecedented revenue visibility through multi-year contracts.
Source: Investing.com
Analysis generated by Bobby AI quantitative model, reviewed and edited by our research team. This is not financial advice. Always do your own research before making investment decisions.
Bobby Insight
The AI infrastructure spending cycle has years of runway ahead, with inference workloads creating durable revenue streams.
Meta's $35 billion commitment to CoreWeave and Amazon's $200 billion capex plan demonstrate that hyperscalers are locking in capacity through 2032, not pulling back. The pivot to inference workloads creates higher utilization and more predictable revenue than the training phase. While power constraints and eventual insourcing are long-term risks, the near-term setup of $725 billion in collective hyperscaler spending creates unprecedented tailwinds for infrastructure providers.
What This Means for Me
