Jefferies Projects 30% Upside for IREN's AI Cloud Pivot

Energy infrastructure company IREN is sitting on a massive untapped resource: roughly 6 gigawatts of global power capacity, of which only 10% is currently being utilized. Jefferies analysts project the company's strategic shift from traditional data center leasing to AI cloud services could unlock 30% upside potential. But the transition faces substantial headwinds, from competitive pressure to the capital and technical expertise required to operate in a market increasingly dominated by specialized players.

The numbers tell a stark story about IREN's current position. A 90% unutilized power portfolio represents both an enormous opportunity and a glaring operational failure under the existing business model. Data center leasing, the company's traditional revenue driver, has failed to absorb the infrastructure the company built. That gap is precisely why IREN is pivoting. AI compute demand has exploded over the past 18 months, with companies desperate for GPU capacity to train and run large language models. AI cloud services, where customers pay premium rates for scarce compute resources, far exceed the thin margins of traditional data center leasing.

Jefferies estimates that AI cloud services could generate substantially higher revenue per megawatt than legacy data center contracts. For a company with 6 gigawatts of idle capacity, even partial monetization represents a multi-billion-dollar opportunity. If IREN can increase utilization from 10% to 30% or 40% while charging AI cloud rates instead of data center rates, the profitability uplift would be significant.

Yet the path from projection to reality is treacherous. IREN is entering a market where CoreWeave, Lambda Labs, and other specialized AI infrastructure providers have already established operational playbooks, customer relationships, and technical expertise. These competitors have been building AI cloud infrastructure from the ground up, optimizing for the unique demands of GPU workloads, distributed training, and real-time inference. IREN, by contrast, is a legacy energy infrastructure company attempting to transform itself mid-career. That transition requires not just capital deployment but a complete operational and cultural shift.

The capital requirements alone are substantial. Operating an AI cloud service means building and maintaining specialized software platforms, hiring deep technical talent in a brutally competitive market for AI engineers, and investing in customer support and sales infrastructure. IREN must also navigate the risk that AI compute demand could shift rapidly toward different chip architectures, more efficient training methods, or on-device inference, potentially stranding expensive infrastructure. The company's previous inability to fully utilize its power capacity under a simpler business model raises questions about its operational execution capability.

The AI cloud market is crowded and well-funded. Hyperscalers like AWS, Google Cloud, and Azure have begun offering GPU rental services. Specialized providers have venture backing and founder-led teams obsessed with the problem. IREN enters as a latecomer with a different organizational DNA. Pricing power in a saturated market is limited, and IREN's 30% upside assumes it can defend margins against both new entrants and established players willing to compete aggressively for market share.

Jefferies' projection is conditional. The 30% upside assumes flawless execution, successful talent acquisition, rapid customer adoption, and stable AI compute demand. Those are not guaranteed. IREN's unutilized capacity is a liability under the old model, but it could become a strategic asset under the new one if the company can build the operational muscle to serve this market. For investors, the pivot represents a high-risk, high-reward bet on management's ability to execute a fundamental business transformation while competing against entrenched and well-capitalized rivals.