GDS Holding: A Sum-Of-The-Parts Opportunity With AI As The Growth Engine

GDS Holdings is framed as a catalyst-driven opportunity, with record Q1 2026 bookings and AI-driven demand expected to lift EBITDA from about $850 million today to $1.5–2.0 billion by 2029. The article also highlights that, after accounting for the DayOne stake, the market values GDS's core China operations at only about 7.4x EV/EBITDA, implying potential upside if the AI growth thesis plays out.

Analysis

This is less a pure “data center beta” trade than an emerging platform re-rating: the market is still pricing GDS like a domestic colo operator, while optionality from AI workloads and the DayOne stake creates a hidden asset-value floor. The second-order effect is that incremental capacity adds can be more valuable than the current EBITDA run-rate implies if AI tenants lock in scarce power/interconnect for multi-year terms, lifting contract quality and reducing cash flow volatility. That makes the equity more levered to booking momentum than to near-term reported earnings.

The main winners are likely GDS equity holders and adjacent suppliers with exposure to power, cooling, and network gear; the losers are slower-moving regional peers whose valuation multiples could be compressed if investors start underwriting an “AI leader” scarcity premium to only one or two operators. A subtle competitive dynamic is that strong bookings can force rivals into more aggressive pricing or higher capex just to defend relevance, which may protect GDS’s occupancy and pricing power over the next 2-4 quarters.

The biggest risk is timing mismatch: the stock can rerate immediately on bookings and guidance, but the EBITDA bridge depends on multi-year execution, financing, and uninterrupted AI demand. Any slowdown in hyperscaler commit rates, regulatory friction, or capital market stress could break the thesis before the 2029 earnings power becomes visible. In that sense, the trade is most attractive over weeks-to-months into catalysts, but the fundamental underwriting should be measured in years.

Consensus appears to be underestimating the embedded DayOne value and over-discounting the China-core multiple because it is anchoring on legacy colo comps rather than AI infrastructure scarcity. The move may still be underdone if the market begins to treat GDS as a hybrid operating company plus strategic asset rather than a simple EBITDA multiple story. The risk is that investors overpay for the growth narrative before seeing conversion into free cash flow, so upside is strongest if management keeps pre-leasing and funding discipline tight.

Key Decisions for Investors

• Initiate a tactical long in GDS into the next booking/guidance catalyst, targeting a 3-6 month horizon; use a 15-20% stop if AI bookings fail to re-accelerate or if guidance implies slower EBITDA conversion than expected.
• Buy GDS Jan-2027 call spreads to express the multi-year EBITDA re-rating while limiting downside to premium paid; this is the cleaner way to capture a potential move from sub-8x to a premium AI-infrastructure multiple.
• Pair trade: long GDS / short a slower-growth regional data center peer or basket, to isolate the market-share and scarcity premium rather than take broad sector beta.
• If the stock gaps higher on catalyst headlines, fade part of the move only if bookings are not accompanied by improved visibility on contracted revenue and capex discipline; otherwise stay with the trend because the rerating can extend for several quarters.