Is It Too Late to Buy Applied Digital (APLD) Stock?

Applied Digital shares have surged ~286% over the past year (as of Mar 23) while the company plans to borrow about $2.15B to build its Polaris Forge 2 AI campus and carries ~$2.6B in long-term debt. Q2 revenue jumped 250% to $126.6M but the company reported a net loss of $31.2M; market cap was roughly $7.5B as of Mar 24. Management is pursuing a spin-off/merge of its cloud business into ChronoScale, which is expected to throttle near-term revenue growth but stabilize margins. Analyst/author tone is cautious due to high valuation and ongoing losses despite strong growth metrics.

Analysis

The market is pricing a high-probability land-grab outcome for AI campus operators; that’s justified only if project execution, grid hookups, and long-term off-take contracts all hit within 12–24 months. The key second-order constraint is not codecs or chips but local power and permitting — incremental GW-scale compute sits behind 6–18 month utility and transmission timelines that can turn booked backlog into delayed revenue and rising carrying costs.

On financing and capital structure, cyclical volatility in access to capital and rates creates asymmetric downside: equity re-rates are rapid on successful pre-sales, but failures or covenant squeezes force dilutive equity raises at much lower prices. A realistic mean-reversion path for valuation requires either visible, multi-year contracted revenue streams or clear margin improvement post-restructuring; absent that, downside from funding shocks materializes within quarters rather than years.

Competitively, hyperscalers’ increasing vertical integration compresses the addressable market for third-party campuses, benefiting operators that lock long-term capacity contracts with specialized AI tenants or monetize adjacent services (managed ops, power optimization, interconnect). This bifurcation implies winners will be those with fixed-price power, guaranteed interconnects, and contractual CPI-linked pricing — the rest become high-beta plays onto financing and sentiment rather than underlying demand.

Key Decisions for Investors

• Buy APLD 12–18 month call spread (e.g., buy Jan-2028 LEAP calls, sell a higher strike to fund ~50% of cost); position size 0.5–1.0% NAV. Rationale: asymmetric upside if execution and contract wins occur; capped loss = premium paid. Close or roll if utility/permit milestones miss 6-month checkpoints.
• Pair trade: Long NVDA (core exposure to AI demand) vs short APLD equity (smaller size, 0.5% NAV). Timeframe 3–9 months. Risk/Reward: NVDA captures broad AI silicon adoption while short captures execution/financing risk; hedge reduces pure sentiment drawdowns in a small net-long AI stance.
• Buy EKSO (small, 0.25–0.5% NAV) as merger optionality play with a 6–12 month horizon; use a protective 20% stop. Rationale: potential multiple arbitrage if the new entity stabilizes margins and reduces capital intensity.
• If concerned about funding risk, buy out-of-the-money APLD puts (3–6 month tenor) sized to cover the equity leg of larger positions. These act as insurance against a financing/dilution event and should be cheaper than full hedging but limit tail risk.