Hut 8 signs $9.8 billion lease for Texas data center expansion

The market is likely to overread this as a near-term earnings step-up, but the important variable is not contracted headline value — it is how much equity dilution and construction risk sits between today and first meaningful cash flow. A 15-year rent stream only matters if the asset is delivered on schedule, the tenant stays put through multiple technology cycles, and financing costs do not eat the spread; that makes this more of a long-duration project finance equity than a clean infrastructure compounder.

Second-order, the clearest winner is not HUT but the AI supply chain around the buildout: NVDA gets validation of its ecosystem, yet the incremental revenue impact is far too small to matter this quarter. The real competitive pressure lands on other power-constrained AI infrastructure names and miner-to-data-center pivots: this deal raises the bar for IREN/CORZ-style stories to prove utility-secured power plus creditworthy tenants, not just megawatt announcements. MCO has a tiny secondary benefit from the financing wave, but fee upside is immaterial versus the balance-sheet and execution risk embedded in HUT.

The contrarian point: consensus is treating contracted MW as if it were already NOI, but the cash actually arrives years out, while the market is capitalizing it today. With energization not until 2027 and phase 2 later, the thesis is most vulnerable to higher-for-longer rates, capex creep, and any slowdown in AI power demand as chips get more efficient. The buyback is more signaling than support against the scale of the buildout, so it does little to de-risk the equity if sentiment turns.