GDS Holdings (GDS.O; 9698.HK) – Leading operator of data centers in China and Southeast Asia
GDS Holdings (GDS.O; 9698.HK) – Leading operator of data centers in China and Southeast Asia
GDS’ stock has been dragged down by weak growth in China and an overextended balance sheet, but management is on track to achieve positive cash flow and a rebound in share price
This brief is co-authored with @valuesalon, a former PE analyst with deep China internet and data center experience.
GDS Holdings (GDS.O; 9698.HK) (“GDS”) is an owner and operator of data centers in China and Southeast Asia, with a market cap of about USD 1.75 billion (USD 9.50 per ADS as of 17 May 2024 market close). In this brief, we explain how GDS’ share price could improve to above USD 14 if management can improve data center utilization, generate free cash flow and de-leverage debt. Favorable macro tailwinds (e.g. easing of geopolitical tensions, generative AI-driven cloud demand, and recovery of onshore financial markets) could further drive a valuation re-rating.
GDS’ data centers outside of mainland China are owned by its Singapore-based subsidiary GDS International (“GDSI”), which was established in 2022. In March 2024, GDS announced GDSI raised USD 587 million from private equity funds including Hillhouse, Boyu and Princeville. Post fundraising, GDS’ stake of GDSI decreases from 100% to 56.1%. GDSI’s pre-money equity valuation is USD 750 million, or about USD 3.92 per ADS of GDS. GDSI will have its own management team, led by Ms. Jamie Khoo, who was previously COO at GDS and worked at ST Telemedia, a Singapore data center investment company.
GDS’s current market cap implies that GDSI is worth about 75% of GDS’ data centers in mainland China (“GDS China”). Yet GDS China’s data center capacity is 5 times larger than that of GDSI, and the China business generated 99.6% of total 2023 EBITDA, implying a significant discount on its domestic operations. Critically, GDS China’s data centers lead the market in the heavily industrialized eastern corridor (tier 1 cities). GDSI is valued at 24x FY25e EV/EBITDA; GDS as a whole is valued at ~7.5x FY25e EV/EBITDA (10.1x FY23A EV/EBITDA).
Private investors were willing to pay a premium for GDSI, because GDSI has a clear growth trajectory in the attractive and rapidly growing Southeast Asia and Hong Kong data center markets. In terms of operating environment (regulatory stability, liquidity, supply-demand dynamics) and strategy (focus on global public cloud customers), GDSI is similar to developed market peers. On the other hand, even after the recent rise in share price, public perceptions about GDS China’s growth trajectory, regulatory uncertainty, and overextended balance sheet are more pessimistic.
The story of GDS from US IPO to 2023
Since its US IPO in 2016, GDS has heavily relied on debt to fund its expansion:
In 2020, China’s digital economy was booming as society adapted to COVID-19. GDS then accelerated its expansion, raising funds with a Hong Kong secondary listing in late 2020 and issuing almost CNY 10 billion of net debt in 2021. Unfortunately, by mid-2021, China’s data center market had peaked, and the operating environment was deteriorating due to:
Aggressive expansion of private equity-backed players and the big 3 telecom operators (China Mobile (0941.HK), China Telecom (0728.HK), and China Unicom (0762.HK)), leading to an oversupply of data centers;
Regulatory scrutiny of the fintech, internet and private education sectors hitting the growth of GDS’ customers, especially its top two customers Alibaba Cloud and Tencent Cloud; and
The government introduced new power and renewable energy regulations for data centers, lowering data center operators’ gross margin.
In 2023, although regulatory pressure on the internet sector eased, growth at GDS’ key customers was still sluggish due to internal restructuring and a weak macro environment. GDS’ data centers remained under-utilized. As of Q423A, GDS China’s utilization rate was 73%, below the target utilization of 85-90%. Given the weak demand, GDS delayed the completion of several new data centers. Pricing (disclosed in terms of monthly service revenue, or “MSR”) also declined steadily. Revenue and EBITDA growth sharply decelerated in FY23A.
As GDS’ massive capex in 2020-2022 did not lead to an acceleration in EBITDA growth, GDS’ net leverage ratio (net debt divided by EBITDA) grew from 5.7x in 2019 to 7.5x by the end of 2023.
China GDS: FY24e and beyond
GDS China is currently navigating a slower growth trajectory due to a broader macro slowdown and changes in China’s internet sector dynamics.
Management has indicated that GDS China will incur capex only as needed for customer move-in; focus on increasing utilization rates; and monetize assets for capital recycling when possible.
Management has said they assume no short-term changes to the current weak market conditions, i.e. no imminent rebound for their major customers. While they see signs that customers’ AI-related demand is increasing, they believe this will only meaningfully drive demand for GDS data centers starting in 2025, when high-end chip supply issues are hopefully resolved.
Given the above, management expects the following for GDS China in FY24e:
Revenue and EBITDA to grow mid-single digits;
MSR to decrease by 3% and reach a bottom by the end of the year, as backlog is digested; and
Capex to decrease to CNY 2.5 billion (down from CNY 3.5 billion in FY23A), as GDS China has already incurred “a substantial part of the development cost” of its data centers, leaving only completion costs
To assess what the Company could achieve up to FY30e, we take the midpoint of management’s forecast for FY24e as a baseline, and add the following assumptions:
MSR: Supply and demand dynamics normalize in FY27e, driving annual 2% increases;
Area utilized: No new data centers built; existing data centers completed by FY26e; utilization approaches the industry standard of ~85% for mature data centers by FY30e;
EBITDA margin: Steadily increase to 54.5% by FY30e, driven by higher utilization rates; for reference, Chindata’s Q123A margin was 56.4% at a utilization rate of 84%;
Capex and depreciation: Annual maintenance capex for fully built data centers to be about CNY 2.5 billion (relatively conservative compared to sell-side estimates of CNY 2.0 billion). Annual depreciation tracks slightly higher than maintenance capex due to historical overpayment for data centers; and
Interest expense: Current debt consists primarily of CNY-denominated project-level debt and fixed-rate convertible bonds. In the past two years, GDS’ effective interest rate declined from 4.7% to 4.3%, thanks to lower rates on CNY loans.
Critically, given the above assumptions, GDS China is expected to generate positive free cash flow (FCF) starting this year as capex starts to moderate, in line with management commentary. We assume GDS to use 100% of FCF to pay down its debt and reach its target 5x net leverage ratio by FY26e.
A contrarian play poised for a rebound amid shifting China narratives
Based on a simplified discounted cash flow model, our estimate of USD 11.05 per ADS is equal to 10x FY24e EV/EBITDA, compared to 20-23x for mature peers (Equinix (EQIX.OQ), Digital Realty (DLR.N), Keppel (KPDCF.PK)), 13x for HK-listed peer SUNeVision (1686.HK), and 11x for A-share listed peer Sinnet (300383.SZ). The implied FY25e free cash flow yield would be 9.6%.
GDS’ major US-listed peer VNET (VNET.O) has traded at a meaningful discount even to other China-based peers, due to concerns about repaying outstanding convertible loans, frequent management turnover and founder-related issues, and a business mix more tilted towards smaller retail customers which are more affected by the macro slowdown.
GDSI’s recent fundraising round appears to be aimed at an eventual full spin-off in 3-5 years when it achieves larger scale. For now, it would be appropriate to apply a “HoldCo discount” to GDSI. Applying a 25% discount to GDSI’s latest private market valuation, the total valuation for GDS would be USD 13.99 per ADS. This valuation would be almost 50% higher than the current stock price, even after the latest rally, but still far below the stock price from 1-2 years ago. Due to GDS’ high net debt, the stock price is highly sensitive to the EV/EBITDA multiple the market is willing to give.
The downside risks for GDS include:
Slower-than-expected improvement in utilization and/or lower-than-expected pricing trend in China, if a persistently weak economy further damages growth of GDS China’s public cloud and internet customers;
Execution issues at GDSI lead to slower overseas revenue and profitability ramp-up; and
Sustained higher U.S. interest rates affect future financing at GDSI.
On the other hand, the GDS stock could conceivably rebound to 2022 levels (~USD 30 per ADS), if macro circumstances augur in its favor:
Geopolitical concerns fade and the “China discount” narrows, such that GDS China can trade at multiples closer to that of developed-market peers and historical levels;
Generative AI-driven cloud demand materializes in 2025-26, which GDS is well-positioned to capture;
Onshore financial markets recover, and GDS can accelerate the deleveraging process via onshore asset disposals; and
Rapid growth at GDSI leads to higher eventual spin-off value.
Currently, we have established a trading position in GDS. Valesalon also holds a position in the stock. In the next few quarters, if management can successfully increase utilization, consistently generate positive free cash flow, and show meaningful progress with the de-leveraging progress, we believe the market should be willing to re-rate GDS closer to its historical and developed-market peer multiples.
Disclaimer: This research piece above is for informational, entertainment, educational, and/or study or research purposes only. The information contained herein or discussed does not, should not, and cannot be construed as or relied upon and, for all intents and purposes, does not constitute or provide professional financial, investment, or any other form of advice. This research does not and should not be construed as an offer to sell or the solicitation of an offer to buy any securities or any other financial instruments in any jurisdiction, including where such actions are illegal. This research is not intended for publication in jurisdictions where it would violate laws. The research does not consider individual investment objectives or financial positions and merely expresses the opinions of its authors. Any investment involves taking substantial risks, including (but not limited to) the complete loss of capital. Every investor has different strategies, risk tolerances, and time frames. You are advised to perform your own independent checks, research, or study, and you should consult a licensed professional before making any investment decisions. The assumptions and parameters discussed or used are not the only reasonable ones, and no guarantee is given for their accuracy, completeness, or reasonableness. No promise is made that any indicative performance return will be achieved. The research is derived from public information sourced by Pyramids and Pagodas. No representation or warranty is given for the reliability, completeness, timeliness, accuracy, or fitness of this research, nor is any responsibility or liability accepted for any loss or damage. The authors (Pyramids and Pagodas) shall in no event be held liable to any party for any direct, indirect, punitive, special, incidental, or consequential damages arising directly or indirectly from the use of any of this material.