Applied Digital stock surges after $7.5B hyperscaler deal By Investing.com

Applied Digital signed a 15-year take-or-pay lease for its Polaris Forge 3 AI data center campus worth about $7.5 billion in base-term contracted revenue, or $18.2 billion including renewal options. The deal adds 300 MW of critical IT load and lifts total contracted lease revenue to $31 billion, or $73 billion with renewals, across four AI Factory campuses. Shares rose 7% after hours as the company highlighted continued hyperscaler demand and 1,200 MW of contracted load across its portfolio.

Analysis

This is less a one-day sympathy pop and more a signal that hyperscaler demand is becoming sticky enough to pre-commit capacity years ahead, which should re-rate the whole private AI-infrastructure complex. The key second-order effect is financing: take-or-pay, investment-grade-backed cash flows materially lower project-risk premia, so names with secured power and shovel-ready land should be able to raise at better spreads and accelerate their pipeline conversion. That’s bullish not just for APLD, but for the broader ecosystem of power, cooling, and electrical equipment vendors that monetize every incremental MW.

The market may be underestimating the bottleneck shift from GPUs to power delivery and interconnect infrastructure. If more hyperscaler demand gets locked in through 2027-2029, the scarce assets become grid access, substation gear, transformers, and liquid-cooling systems, creating a multi-quarter tailwind for the picks-and-shovels layer while leaving pure compute operators more exposed to capex intensity. APLD’s ability to monetize >1.7 GW of connected power also suggests optionality value on its land bank is rising faster than consensus models typically capture.

The main risk is execution timing: these contracts are long-dated, so the equity can outrun near-term fundamentals and then correct if investors focus on cash burn, project delays, or power permitting slippage over the next 6-12 months. Another failure mode is customer concentration — one hyperscaler can become a pricing anchor if future renewals are repriced lower or if demand shifts to vertically integrated capex. The move is likely directionally right, but the stock may be overextended on headline revenue optics versus present value of cash flows.

Contrarianly, the real trade may be to fade the more crowded AI compute names and rotate into infrastructure beneficiaries with earlier cash conversion and lower customer concentration. If AI buildout broadens, the market should reward de-risked contracted power more than speculative model growth; if the AI capex cycle slows, APLD still has downside because the market is paying today for 2027+ realization. That makes this a good relative-value event rather than a clean outright momentum chase.

Key Decisions for Investors

• Long APLD on a 2-6 week pullback only; use the post-print squeeze to avoid chasing, with a target tied to re-rating of contracted revenue value and a stop if the stock loses the breakout level and the market starts pricing in execution delays.
• Pair trade: long APLD / short a basket of high-multiple AI compute beneficiaries with weaker contracted visibility over 1-3 months; the spread should benefit if investors shift from narrative to de-risked infrastructure cash flows.
• Add to infrastructure suppliers with power-grid exposure over 3-12 months (e.g., ETN, PWR, ABB) on the thesis that AI demand is now constrained by electrical bottlenecks, not semiconductors; risk/reward is favorable if MW backlog keeps compounding.
• Use call spreads rather than outright longs in APLD for 3-6 months; upside remains substantial if the market capitalizes the long-dated lease stream, but options cap downside if the stock mean-reverts on execution skepticism.
• Watch for a financing/rating catalyst over the next 1-2 quarters; if APLD secures cheaper project debt or additional hyperscaler leases, it becomes a higher-conviction long, but absent that, size should stay tactical.