The data center operator recorded faster bookings and greater utilization on surging demand for AI computing, though asset impairments and heavy debt are clouding its profit picture

image credit: Bamboo Works

Key Takeaways:

  • GDS swung to a net profit of 960 million yuan last year, reversing years of losses, as its revenue rose 10.8% to 11.43 billion yuan
  • The data center operator recorded a 2.36 billion yuan one-time gain during the year from deconsolidation of a subsidiary

AI's rapid rise has created a large array of winners, including data center operators that sit squarely in the sweet spot of this transformative wave. As demand for computing power skyrockets, such operators have become the critical backbone of the AI supply chain. For leading Chinese operator GDS Holdings Ltd. (NASDAQ:GDS) (9698.HK), the most tangible evidence of that shift was a marked acceleration in new orders and customer demand last year, which was disclosed in its latest financial report released last week.

The report showed GDS achieved five-year highs in both new bookings and gross move-ins. Its utilized area climbed to 504,800 square meters, driving its utilization rate up to 75.5%. Its revenue for the full year hit 11.43 billion yuan ($1.63 billion), up 10.8% year-on-year. Its adjusted earnings before interest, taxes, depreciation and amortization (EBITDA) also posted double-digit growth, rising to 5.4 billion yuan. Those factors helped the company to end its years of losses, reporting a net profit of 960 million yuan for the year.

Paper profits

But a deeper dive into the report reveals a less rosy picture, at least in terms of the newfound profits. The company's core operations remained in the red, with an operating loss of about 55.8 million yuan. At the same time, GDS booked a hefty 1.56 billion yuan loss on asset impairments, signaling underperformance of some assets.

The path to profitability was paved largely by non-recurring items, including 716 million yuan from its share of results of associates and joint ventures, and a massive 2.36 billion yuan one-time gain on deconsolidation of a subsidiary. That gain stemmed primarily from bringing external investors into its global data center operation, DayOne, through a partial stake sale.

To balance its breakneck expansion while maintaining financial stability, GDS, like many data center operators, has needed to constantly raise money through measures like issuing asset-backed securities (ABS) and convertible preferred shares, and setting up a real estate investment trust. Such steps helped to raise its cash balance to 14.3 billion yuan by the end of last year. But all the financing has also left GDS with more than 46 billion yuan in short- and long-term debt. The industry's need for constant high capital spending means that profitability remains significantly reliant on capital maneuvers and accounting treatments for many companies.

That said, the AI era is driving a change to GDS' client mix. Unlike the past, when its enterprise client base was typically quite fragmented, the AI boom is bringing a new group of larger hyperscalers and major AI computing customers. Such clients typically bring larger orders and faster deployment cycles, helping data center operators to boost their utilization and occupancy rates, driving steadier revenue growth. But the shift also means rising customer concentration, and also potentially having to grant these large buyers discounts that further erode a data center operator's margins.

That trend is already affecting GDS' profitability. In the fourth quarter of last year, the company's adjusted gross margin dipped slightly to 50.6% from 51.9% a year earlier, while its overall gross margin slipped to 21.0% from 21.5%. Management cited rising utility expenses as a percentage of revenue as a key factor. Structurally, however, it also correlates with the shift towards large-scale computing clients, which typically consume more resources per unit.

The AI boom, with its high-performance computing requirements, is also keeping capital expenditures high for data center operators. GDS' cash flow statements show it continues to invest heavily in new data center construction, with projected 2026 capex around 9 billion yuan. Notably, the company was back in the red during the fourth quarter with a net loss of 460 million yuan, wider than the year-ago period, primarily due to asset impairments. This highlights the potential for new investments to underperform as companies are forced to spend rapidly on new demand that they can only try to anticipate.

Amassing computing resources

Despite the uncertainties, GDS continues to double down on computing infrastructure. It has amassed approximately 900 MW of resource reserves in China's tier-one cities, laying the groundwork for future AI demand. Its overseas business through DayOne is also accelerating, with Southeast Asia commitments reaching 431 MW. The launch of a 120 MW project in Bangkok signals its younger international business is gradually transitioning from investment phase to harvest mode.

GDS's U.S.-listed shares initially jumped about 5% the day it announced its results before paring the gains to close up by a much smaller 0.52% at $44.49, putting the stock up roughly 28.5% this year. Its Hong Kong shares fell over 5% after the announcement, but are also still ahead this year by a similar 28.6%. Morgan Stanley reaffirmed its "overweight" rating on the company's U.S.-listed shares with a $64 target price, citing sustained AI-driven demand for bookings and occupancy.

Valuation-wise, GDS trades at a price-to-sales (P/S) ratio of about 6.5 times. That trails global leader Equinix (EQIX.US) at 10.4 times but is more than triple Chinese peer VNET (VNET.US) at 1.8 times. Unlike the more mature Equinix, GDS still remains in a heavy investment phase, with spottier earnings quality and cash flow stability. Investors may prefer GDS to VNET for the former's positioning in AI demand and operational efficiency.

Some may see GDS as transitioning to a new phase from pure expansion toward trying to improve its profitability. AI demand is offering an unprecedented growth opportunity, while also placing greater demands on its cost structures and capital spending. Still, more premium clients as a result of growing AI demand should bring better visibility of its future spending needs, positioning the company for improving performance over the longer term. But in the shorter-term, its profits and cash flow are likely to remain more volatile.

To subscribe to Bamboo Works weekly free newsletter, click here

Benzinga Disclaimer: This article is from an unpaid external contributor. It does not represent Benzinga’s reporting and has not been edited for content or accuracy.