Macquarie raises GDS Holdings stock price target on capex outlook By Investing.com

GDS reported Q4 2025 EPS of $0.40 versus an expected loss of $0.26, a 253.85% surprise beat, while revenue missed at $2.88B vs $2.94B (-2.04%). Macquarie raised its price target to $67.50 (from $61.80) and kept an Outperform/Marquee Buy, and Morgan Stanley reiterated Overweight with a $64 PT; the stock trades at $45.09 (up ~27% YTD) and is ~11x 2026 China EV/EBITDA. Company fundamentals cited include ~15 billion yuan cash and a current ratio of 2.6, plus a 500MW sales target mostly to be delivered by 2027 amid an AI-driven data-center capex cycle in China.

Analysis

The AI-driven capex catch-up by Chinese hyperscalers creates a multi-year demand wave for data-center real estate and high-performance compute gear, but the bottlenecks will be narrowly concentrated: high-voltage grid upgrades, precision cooling systems, and AI accelerator supply (in particular the power/heat integration packages). That concentration means margin capture will favor operators with deep land reserves, fast permitting relationships with provincial utilities, and direct OEM partnerships rather than pure-play colo landlords that compete on price.

Key near-term catalysts are booking cadence and utilization signals over the next 6–12 months; meaningful upside requires visible, recurring revenue contracts and sustained utilization above mid-70% levels to justify current multiple cushions. Tail risks cluster around three vectors over 12–36 months: accelerated model-efficiency gains that reduce training compute intensity; export-control or cross-border capital frictions that delay foreign vendor deliveries; and a local power-permitting squeeze that forces staggered rollouts, all of which can turn an expected “supercycle” into a drawn-out capacity build with pricing erosion.

Consensus frames this as a simple capacity race, but it underestimates project-level execution risk and the capital intensity of integrating substations and local grid upgrades — these are 6–18 month lead-time items that can bottleneck revenue recognition even when servers are available. Structurally, that makes asymmetric option structures preferable to naked equity exposure: you get upside to a faster-than-expected ramp while limiting losses if bookings slip or utilization falls below the break-even threshold.

Key Decisions for Investors

• Long GDS (GDS) equity with a protective collar: establish a position size of 2–3% NAV, horizon 9–12 months; finance by selling a 12-month 20–25% OTM call and buying a 12-month 10% OTM put. Risk/reward: limits downside to ~10–12% (net of premium) while retaining most upside if hyperscaler bookings convert (upside scenario 30–60%).
• Directional server play via SMCI: buy a 9–12 month call spread (buy ATM call, sell ~30% OTM call) sized at 1–2% NAV to capture backlog-driven re-rating. Risk/reward: capped cost (loss = premium) with 2–3x upside if OEM order flow and ASPs remain firm over the next 6–12 months.
• Macro/tail hedge for extended conflict scenario: allocate 1% NAV to 3-month 10% OTM SPX puts or equivalent long-VIX call exposure to protect against a sudden >20% equity drawdown tied to geopolitics. Risk/reward: modest insurance cost that pays >3x notional in severe drawdown scenarios highlighted by UBS.
• Event-driven alert: set trading triggers to take profits or add on GDS when quarterly utilization >75% and sequential booking growth >20%; trim or hedge if sequential ASPs decline or if provincial-level power approvals pause (monitor local permitting notices).