Meta's Zuckerberg says compute bought in 2024-25 already trades at a premium. IREN's $2.8B in new contracts and 5GW uncommitted pipeline let it price into that tighter market.
Meta Platforms CEO Mark Zuckerberg said something on his company's second-quarter 2026 earnings call that cuts straight to the heart of the AI infrastructure trade. Compute bought in 2024 and 2025 is already worth more than what it cost.
"We are getting a lot of offers for compute at a significant premium over what we paid for it," Zuckerberg told analysts. He added that Meta expects to "grow a large business serving large customers" by selling capacity in the future.
The implication is direct. If demand runs ahead of supply and pricing is moving up, the provider that still has capacity to price into that market holds the stronger hand. IREN Ltd. appears to be that provider.
IREN announced on July 20 that it had signed $2.8 billion in new multiyear cloud service contracts with multiple leading AI developers. It raised its year-end annualized run-rate revenue target for its AI cloud business to over $4 billion, up from $3.7 billion. That is real commercial traction. It is also only part of the picture.
The company has 810 megawatts of operational capacity, 2,100 megawatts under construction and 1,600 megawatts in development across six North American sites. That adds up to roughly 5 gigawatts of power dedicated to high-performance compute. All sites are renewable-powered and grid-connected, using Nvidia reference architectures with non-blocking InfiniBand networks.
What matters is how much of that pipeline remains unpriced. IREN has not locked up the bulk of its future capacity in a few giant early-stage contracts. In a market where Zuckerberg himself says compute is trading at a premium to its purchase price, that uncommitted pipeline is not idle inventory. It is pricing power.
IREN's strategic partnership with Nvidia, announced in May, adds another layer. The two companies are working to accelerate deployment of up to 5 gigawatts of AI infrastructure. IREN Cloud is built on Nvidia's reference architectures and integrated directly into Nvidia's AI ecosystem, with bare-metal GPU clusters available for training and inference at scale.
Vertical integration gives IREN a structural advantage here. The company owns and operates its sites, lines up long-term renewable power and engineers facilities specifically for AI workloads. When hyperscalers like Meta come back to the market seeking additional capacity at higher prices, a provider that can reconfigure racks and power feeds without a landlord in the middle has more flexibility than a pure leasing platform.
Zuckerberg's comments effectively validated the strategy of treating compute as a long-lived asset rather than a commodity. For IREN, the combination of $2.8 billion in proven contracts, a 5-gigawatt development footprint and a Nvidia partnership points toward a company that built for exactly this moment. The year-end run-rate target of $4 billion in AI cloud revenue alone gives a concrete number to watch.
IREN's Alpha Score sits at 38/100 on AlphaScala's scale, a Mixed label. The stock page is available here.
Drafted by a large language model from the source reporting linked above, then screened by automated publishing checks. It is not read by a journalist before publication. Some articles cite our Alpha Score. Verify prices and figures against the original source. Educational coverage, not personalized advice.