21Vianet

VNET reported a strong first-quarter 2026 result with net revenue growth of 19.8% and adjusted EBITDA growth of 30.6%. The first quarter was ahead of management's unchanged 2026 guidance of net revenue growth of 15.6%-18.6% and adjusted EBITDA growth of 19.2%-25.9%.

Why it matters: Internet data center demand was weak over 2022 and 2023, but artificial intelligence has driven demand up since 2024 and seems to be accelerating. This supports VNET's long-term target of growing data center assets under management from 1 gigawatt currently to 10 GW by 2036. VNET won 519 megawatts of wholesale orders in first-quarter 2026 and increased its wholesale capacity in service by 34% to 687 MW, and it has a further 220 MW unutilized and a further 516 MW under construction. Wholesale revenue surpassed retail revenue for the first time this quarter.

The bottom line: We make only minor changes to our forecasts and retain our fair value estimate of USD 7.50. We see the stock as overvalued. The price/book ratio is now around 3.2 times, well above GDS' 1.7 times. VNET does not revalue its data centers each year like a property REIT does, and we estimate that its early data centers in central business districts of major cities have substantially increased in value.

Key stats: Monthly recurring revenue per retail cabinet has risen from CNY 8,794 in fourth-quarter 2024 to CNY 9,448 in first-quarter 2026 as AI demands higher power density, for which VNET can charge extra. Ninety percent of retail IDCs are mature, while only 63% of the faster-growing wholesale capacity is mature.

Long view: With VNET leveraged to long-term growth themes such as the shift of IT to the cloud and AI, we see it as a decent long-term investment. However, high financial leverage gives it a Very High Morningstar Uncertainty Rating and has driven a volatile share price over the past eight years.

Our fair value estimate for VNET is $7.50 per share. We assume VNET grows its revenue at an average 12% per year over the next five years, with operating margin improving from 10% in 2025 to 38% in 2030. The margin improvement is because we assume a lower growth year in the terminal year of our forecasts, with a greater percentage of mature data centers compared with lightly loaded or under-construction data centers. Our forecasts assume 59% consolidated EBITDA margin by the end of 2029. Capacity utilization, one of the main drivers of margins and returns, peaked at 75%-76% in 2016-17, then fell to around 59% by 2023 but has started to rebound and was 70.1% for wholesale and 64.0% for retail by end-2025. We assume the utilization recovers to 75% by 2029. We forecast capital expenditure of CNY 11 billion in 2026, in line with management guidance. This means VNET remains free cash flow-negative from 2025 to 2027, but VNET will likely use asset sales to help fund this. However, we assume a reduction in capital expenditure from 2028 onward, which will make VNET free cash flow-positive from 2029.

We give VNET Group a no-moat Morningstar Economic Moat Rating as we believe the bargaining power that its large hyperscale customers will have will offset the switching costs normally associated with the data center industry. We also believe that the new capacity that is being deployed in outer suburbs will not enjoy the scarcity of data center space that its data centers near CBDs in large Chinese cities do. Overall, we see little differentiation between data centers except for location. Moats in data center businesses under Morningstar’s coverage can be derived from network effects, customer switching costs, and cost advantages. We generally rate hyperscale data center businesses (those that lease large spaces to hyperscale tenants on lengthier contracts) as less moaty than retail co-location businesses. This is mainly because wholesale customers have more bargaining power, given their size and higher potential to build their own data centers.

Most data center companies do both hyperscale and retail, with VNET having started as a pure retail business and adding wholesale from 2019. Most of its growth has since come from wholesale with wholesale now slightly larger than retail in terms of revenue. The other China-focused data center operator under our coverage, GDS Holdings, is rated no moat, but it generates around 85% of its revenue from hyperscale customers, with the remainder from retail. Like GDS, VNET's current capacity is focused on or near the CBDs of large cities such as Shanghai and Beijing. It is very hard to get suitable new data center sites in these locations, so it will be hard for customers in data centers in these areas, in particular, to switch to other providers. However, our moat ratings are forward-looking, and we see the bulk of VNET’s new data center capacity being deployed in areas at the fringes of the major cities where large customers will likely have more options to switch to other data center suppliers when their contracts expire, or use the threat of moving to keep a lid on returns for VNET.

We typically see switching costs as a competitive advantage for data centers. If data center tenants leave, they need to take significant risks to move critical servers and networking equipment and incur expenses, which, according to Align Communications, could cost tens of millions of dollars for multimegawatt customers (generally the hyperscalers that use massive amounts of power) or at least $10,000 per rack for smaller customers. Moving also generally takes more than six months and can take more than one year, including planning and execution for larger moves. VNET reports average yearly churn of 4.8%, 2.4%, and 1.2% for 2023, 2024, and 2025 for its retail managed service product, respectively, which implies that the average customer stays with VNET for at least 20 years.

The data center market has been growing strongly, and we believe most of the recontracted data centers have been in inner city areas where, as previously discussed, demand has been high, and supply has been low. We expect that customers will likely hold better bargaining power in future contracts that expire in outer city data centers, and that churn could therefore increase. VNET is transparent with its large hyperscale customers with its costs and pricing on a “cost plus reasonable return” basis. Its returns for the hyperscale customers are therefore slightly lower than for the retail component of its business, but still above the cost of capital, according to the company. The reason we believe hyperscale data centers are less moaty than retail is that, given the longer contracts wholesale customers sign, we believe they are far more discerning and more likely to justify the costs of moving. We are also aware that hyperscale customers have the ability to replace a presence at VNET with one in a self-owned data center.

We also believe VNET is helped in the tenders by its experience and reputation for building quality data centers and its growing network of data centers, but do not see this as strong enough to warrant an intangible asset moat. The network effect for data centers is generated in major internet exchanges where key internet service providers are connected, particularly for applications where latency is important. In this situation, service providers attract customers to a data center, who then attract more service providers. In China, the key areas where network effect is likely present are in the CBD areas of Tier 1 cities, where new data center capacity is difficult to build, such as in Shanghai and Beijing. VNET has a significant capacity in service in these areas. It has 25,000 cabinets in service in downtown Beijing and a further 155 MW of wholesale capacity. However, most of its area under construction and land held for future expansion is at the edge of cities where suitable sites are more plentiful, so we would expect the amount of capacity in prime downtown areas to fall below 50% within the next two to three years.

VNET has around half as much wholesale capacity as GDS and both are larger than smaller independent operators such as Chindata and Sinnet. However, the three Chinese telecom companies have bigger data center networks than VNET. We note that the Chinese telecom companies have a much greater percentage of their data centers located outside the main Tier 1 markets and, in many cases, are the main data center providers in Tier 2 and Tier 3 cities. To the extent that any scale-based cost advantages are available in this industry, we would expect the Chinese telecom operators to be the main beneficiaries. Note, in addition to the listed data center operators, there are numerous smaller data center owners who own only one or two data centers.

VNET’s rapid expansion leaves it well positioned to capitalize on the huge demand for data centers brought on by cloud usage and artificial intelligence.

The Internet of Things, artificial intelligence, and other innovations that increase the demand for data and connectivity leave us in the early innings of a data center renaissance.

We believe VNET's large retail customer base will be less likely to churn and therefore allow higher margins than its wholesale base.

VNET's data centers are not significantly differentiated from those that others, including its large customers, could supply.

VNET has spent too much and expanded too rapidly, and it will not hold up well if the industry cools and it can't cover its borrowings.

VNET has very large and financially strong competitors in the three large Chinese telcos whose data centers already have four to seven times the total capacity of VNET’s, and who have aggressive expansion plans that can be funded by telecom profits.