This Tech Stock Is 1 of the Most Shorted Stocks of 2026: Is Now the Time to Buy?

Oracle is down ~53% from its Sept. 2025 all-time high to about $153/share, after $21B capex last year and an elevated $55B capex plan for fiscal 2026 drove negative free cash flow of $23.7B (fiscal 2026). Management also plans to raise ~$40B via debt/equity in fiscal 2027, with debt at $167B and a 388% debt-to-equity ratio raising solvency/earnings risk. Despite a perceived valuation discount (forward P/E 18; PEG 0.85) and an analyst median price target of $241 (+57%), investors remain concerned that AI infrastructure spending and OpenAI contract execution could pressure returns.

Analysis

Oracle is increasingly being treated less like a software compounder and more like a leveraged financing vehicle for AI buildout. That matters because equity value in this setup is driven by the spread between promised future utilization and the carrying cost of the infrastructure today; when debt-funded capex outruns visible cash conversion, the market typically compresses the multiple long before any revenue miss shows up. The immediate loser is ORCL, but the bigger second-order risk is to any adjacent “AI infrastructure” name whose valuation assumes a clean, rapid return on capital rather than a multi-quarter payback curve.

Near term, the stock can still bounce because oversold/high short-interest names often rally on less-bad guidance or on any sign that financing is manageable. The real catalyst path is 1-3 months: watch for debt issuance terms, credit spread behavior, and whether the company is forced to defend growth with even more external capital. If bond markets start to demand a higher risk premium, the equity thesis weakens quickly; if spreads stabilize and capex moderates, the bear case loses force. Over 6-18 months, the key falsifier is evidence that backlog monetization improves enough to turn free cash flow positive without further balance-sheet stress.

Contrarian view: the market may be extrapolating financing optics into a permanent impairment of franchise quality. If the backlog is genuinely long-duration and counterparty risk is overstated, ORCL could rerate sharply because the stock now trades more like an earnings multiple reset than a broken business. Still, the cleaner expression is relative value: long higher-quality AI spenders with stronger balance sheets against ORCL, because the market is likely to continue punishing levered capex stories before it rewards eventual revenue recognition.

Key Decisions for Investors

• Initiate ORCL / MSFT pair trade: short ORCL, long MSFT over the next 1-3 months. Thesis: both participate in AI demand, but MSFT has better FCF conversion and far lower balance-sheet fragility. Falsify if ORCL shows improving FCF and tighter financing spreads while MSFT guidance softens.
• Use ORCL put spreads into the next financing or earnings catalyst rather than outright short stock. Best risk/reward is a 1-3 month structure that profits from another leg down if credit concerns reprice, while capping theta bleed if the name bounces on short covering.
• Watch ORCL credit aggressively: if spreads widen materially over the next 4-8 weeks, treat that as a stronger bearish signal than the equity alone. Add to short only if bond market confirms the capital structure is becoming the story.
• Avoid chasing NVDA/AMD on the basis of Oracle’s backlog alone. Near-term hardware demand may benefit, but if the market shifts from 'AI spend is good' to 'AI spend must earn a return,' high-multiple suppliers can de-rate together; use pullbacks selectively, not as a momentum bid.
• If ORCL reclaims and holds the pre-breakdown valuation band after the next update, cover part of any short exposure. That would indicate the market is willing to underwrite the capex cycle for another quarter, reducing the edge on the bearish side.