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

Applied Digital (APLD) is pursuing aggressive AI/data-center buildout with plans to borrow roughly $2.15B for its Polaris Forge 2 campus; the company had a market value of about $7.5B and carries $2.6B in long-term debt. Latest reported Q2 revenue surged 250% to $126.6M while it still posted a net loss of $31.2M; shares are up ~286% over the past year. Management is spinning off its cloud unit to merge with EKSO Bionics into ChronoScale, which may slow near-term revenue growth but could stabilize margins; the author remains cautious on valuation and ongoing losses.

Analysis

Applied Digital’s business model acts as an accelerant for hardware demand while concentrating execution risk: whoever aggregates rack-level demand (colo + build/operate) becomes the marginal buyer of GPUs, networking and power gear and can swing OEM pricing and lead times. That creates a two-way exposure — GPU vendors (Nvidia/Intel) see durable unit demand and pricing power, while infra operators absorb project, grid and financing risk that can amplify earnings volatility on lumpy capex cycles.

A near-term bifurcation is likely: equity upside is concentrated in execution milestones (financing, permitting, spin-off close) that de-risk future FCF, while downside is compressed into financing-market, permitting or GPU-price reversals that are binary and fast. Credit-sensitive owners (banks and bond markets) will reprice quickly if a planned raise comes with covenant-heavy terms or if used hardware becomes impaired in a downturn, turning an operational story into a solvency story within months.

Second-order winners include GPU OEMs (benefit from captive large orders) and grid/energy service providers (PPAs, transformers, storage) who will capture durable high-margin retrofit spend; losers are small colo players and legacy cloud tenants who can't match scale-derived procurement. From a positioning lens, the current move looks to have priced optionality on successful build-outs rather than execution certainty — that asymmetry favors structured downside exposure now and selective long exposure to GPU winners.

Key Decisions for Investors

• Short APLD via put spread (buy 3–6 month APLD puts, sell lower-strike puts) sized 1–2% NAV: targets asymmetric payoff if financing terms or permitting updates are negative. Risk: max loss = premium paid; Reward: 30–60% in stressed deleveraging scenarios. Monitor: financing announcement, spin-off timeline, and any credit-line amendments.
• Pair trade — long NVDA (1.5–2% NAV) vs short APLD (1.5–2% NAV) for 6–18 months: owns secular GPU upside while hedging infra execution risk. Risk/Reward: keep position small vs portfolio volatility; if GPU cycle accelerates, NVDA gains should outpace APLD re-rating, but if macro shock hits capex both could fall.
• Event-driven long EKSO (to capture positive re-rating of ChronoScale post-close) sized 0.5–1% NAV with strict close-date stop: asymmetric payoff if market awards standalone valuation premium after deconsolidation. Timeframe: 3–9 months around merger close. Key risks: integration failure or revenue cliff realization from divestiture.
• Tactical hedge: buy short-dated NVDA or INTC puts (calendar) as a macro hedge for a 1–3 month window around financing/spin-off milestones; inexpensive insurance if market-wide risk premia spike due to higher rates or a liquidity squeeze.