Microsoft May Have Its Worst Showing in 20 Years. Is It Too Cheap to Ignore?

Microsoft shares are down roughly 25% in Q1 2026, the steepest quarterly decline since Q4 2008. Fiscal 2026 capex is forecast at $146B (a 66% increase from $88B in FY2025), with analysts projecting $170B in 2027 and $191B in 2028 to support Azure and Copilot; Azure growth has shown a recent deceleration and Copilot adoption and retention lag expectations. Valuation has compressed below 20x forward earnings (lowest since June 2016) and briefly traded below the S&P 500, presenting a potential multi-year entry for investors who view the capex as convertible into durable Azure/Copilot revenue growth.

Analysis

Microsoft’s AI capex cycle is creating an ecosystem bifurcation: capital-intensive cloud scale benefits hardware, power, and colocation suppliers even if software monetization lags. Expect outsized revenue flows to GPU vendors, PSU/chiller makers, and hyperscaler-adjacent REITs as capacity is contracted and turnkey deployment wins the next 12–36 months; that funneling of cash can sustain portions of the AI supply chain even while Microsoft’s ASPs and seat-based upsell hit friction.

A key second-order risk is operational cadence — permitting, grid upgrades, and long lead-times for exascale interconnects create a multi-quarter rollout cliff where demand outstrips available capacity. That timing mismatch amplifies near-term churn metrics for Copilot-style products (trial-to-paid conversion) and gives competitors a window to lock-in customers with lighter-weight, API-first offerings; reversals will come from demonstrable retention improvements or material FCF inflection, not marketing spend alone.

The market is pricing optionality into Microsoft: a heavy discount to historical growth implies investors expect either margin compression or permanent displacement by pure-play AI providers. That makes a structured, time-limited asymmetric exposure optimal — capture the embedded optionality of Azure/Copilot re-acceleration while funding downside protection via hedges in high-beta AI/software names and owning the direct beneficiaries of infrastructure scarcity.

Key Decisions for Investors

• Long MSFT via calendar/LEAP call spread: buy 18–24 month LEAP call and sell nearer-term call to finance (target 2:1 to 4:1 payoff). Timeframe: 12–24 months. R/R: limited premium with 2–4x upside if Azure/Copilot re-accelerate; tail risk if AI monetization structurally weak.
• Long infrastructure winners: EQIX and DLR (6–12 month horizon). Rationale: increased colocation demand and sticky contracts from hyperscaler capacity expansion. Risk: oversupply or macro slowdown; target 15–25% price appreciation plus dividends.
• Long NVDA (or call spread) vs short high-multiple software/software-adjacent name like SNOW as a pair (6–12 months). Mechanism: capture GPU pricing power while hedging headline AI sentiment. Risk/Reward: asymmetric — NVDA upside if capex sustains; short cushions cost of carry but monitor correlation drawdowns.
• Hedge tactical downside in MSFT with put-buyers or credit structures: buy 3–6 month puts sized to cover position shock around earnings/Q announcements, financed by selling out-of-the-money calls on overbought names. Timeframe: tactical around next two earnings beats/misses. Rationale: protects against short-term sentiment-driven gaps while keeping long-term exposure.