VNET Group, Inc. (VNET): what the price assumes

In the published model solve dated 2026-Q2, anchored at $7.21, VNET Group, Inc. (VNET) is priced for +18.1% growth. boothcheck publishes no house fair value, target price, or buy/sell rating; individual model outputs and user-controlled scenarios are analytical inputs, not Boothcheck targets. Narrative composed 2026-06-28.

Generated: 2026-08-10 · Exported: 2026-08-12 · Source: https://boothcheck.com/report/VNET

Headline

FieldValue
TickerVNET
CompanyVNET Group, Inc.
Current price$7.21/sh

What The Price Assumes (Inversion)

The assumption today's price embeds, recovered by inverting the valuation.

FieldValue
Inversion basiswhole-company
Implied growth18.1%
Multiple paid27x operating income

Solve inputs: computed at a 8.2% cost of capital with 4% terminal growth over a 5-year stage; each 1pp of cost of capital moves the implied operating-profit growth ~8pp.

Reconcile: at the x-ray's 9.3% required return this reads ~5.3 years; the models below use their own rates.

How unusual the bet is: within-range

ReferenceValue
vs own history+0.16σ
sustained it ~5 years at this level46%
implied end-window share0%

Valuation X-Ray

The price is justified by relative-multiple and growth-DCF; asset-based land below the price.

How the valuation models price the stock relative to the market price. Price/FV above 1.0 means the market pays more than that lens defends (expensive); at or below 1.0 the lens can defend the price.

FamilyMedian price/FVModelsReads
Asset2.42x3expensive
Earnings—0—
Relative0.24x1justifies
Growth0.52x1justifies

Families that justify the price: Relative, Growth Families that call it expensive: Asset

The models below discount at their own flat-beta convention rates (cost of equity 9.3%, WACC 5.0%); the inversion above states its own rate.

Per-Model Detail (n=5)

ModelFamilyFVPrice/FVApplicableMethodology
DCF Perpetual GrowthGrowth$13.940.52xyesReference only (OCF-based, capex excluded): OCF $0.3B
DCF Exit MultipleGrowth$0.00—noNegative/zero FCF or EBITDA — equity value floored at $0
Relative ValuationRelative$0.00—noNegative forward EPS and no revenue for P/S fallback
Simple DDMGrowth——no—
Two-Stage DDMGrowth——no—
Simple Excess ReturnAsset$3.312.18xyesBook value floor: BV/sh $3.31, ROE negative
Two-Stage Excess ReturnAsset$2.982.42xyesBook value with convergence: BV/sh $3.31, ROE converges to ke
Discounted Future Market CapGrowth——no—
Peter Lynch Fair ValueRelative$0.00—noNegative/zero EPS — earnings-based value floored at $0
Margin TrajectoryGrowth——no—
Earnings Power ValueEarnings$0.01721.00xyesNormalized EBIT (5y avg op income, one-time charges added back) $0.03B × (1−21%) / WACC 5.0% → EPV (no growth) (excluded from median)
Residual IncomeAsset——no—
Graham NumberAsset——no—
EV/EBITDA RelativeRelative$30.300.24xyesEBITDA $0.42B × sector EV/EBITDA 25.0x
FCF YieldEarnings——no—
SBC-Adj FCF YieldEarnings——no—
Ben Graham FormulaEarnings——no—
ROIC-Justified P/BAsset$1.873.86xyesBV $3.31 × (ROIC 2.8% / WACC 5.0%)
P/Sales SectorRelative——no—
PEG Fair ValueRelative——no—
Earnings YieldEarnings——no—
Funds From Operations MultipleRelative——no—
Clinical Phase NPVGrowth——no—
MertonAsset——no—
V5 Mechanical———no—

Solvency

FieldValue
Net debt$344.6m
Net debt / NOPAT (after-tax)3.91x
Net debt / operating income (pre-tax)3.09x
Interest coverage1.3x
Share count CAGR (buyback)-3.0%
Burning cashno

Bullet Takeaways

Bull Case

The moat for a wholesale data-center operator is land, power and committed capacity in the right places, and VNET has been converting that into demand at speed. Building large-scale data centers in China requires securing power allocations and sites near the demand centers, then signing customers to long-term contracts before the steel goes up. VNET ended Q1 2026 with 907 megawatts of wholesale capacity in service at a 75.7% utilization rate and 869 megawatts of total committed capacity, concentrated in the Greater Beijing area where hyperscale demand is densest. That installed, contracted base is the asset that is hard to replicate, and it is filling up: wholesale net revenue grew 58.1% to RMB1.06 billion and, for the first time, the wholesale business out-earned the legacy retail colocation business.

The AI cycle is the demand the moat is capturing. Year to date VNET has secured 519 megawatts of new AI and hyperscale-driven orders, with 510 megawatts coming from a single leading internet customer for facilities in Greater Beijing. That is the kind of anchor commitment that underwrites a multi-year build. The financials are following: Q1 net revenue rose 19.8% to RMB2.69 billion and adjusted EBITDA grew 30.6% to RMB891.5 million, lifting the EBITDA margin to 33.1%. Management guides to full-year revenue of RMB11.5 to RMB11.8 billion, up 15.6% to 18.6%, and adjusted EBITDA of RMB3.55 to RMB3.75 billion, up as much as 26%. EBITDA growing faster than revenue is the signature of a capacity business gaining operating leverage as utilization rises.

The validation came from a strategic buyer. Investors affiliated with the battery giant CATL agreed to acquire up to a 38% stake, a deal valued at as much as $942 million, recasting VNET from a former data-center laggard into a credible China AI-infrastructure player. Notably, the structure leaves the founder in operational control while bringing in a deep-pocketed strategic partner. That endorsement, the AI order book, and the EBITDA inflection are why the small group of analysts covering the name carry a strong-buy consensus with targets well above the current price. The bull case is that VNET owns scarce, contracted capacity in front of a Chinese AI build that is only beginning.

Bear Case

The structural problem sits on the balance sheet, because a data-center build at this pace is a debt machine. VNET guides to roughly RMB10 to RMB12 billion, about $1.4 billion, of capex this year to deliver another 450 to 500 megawatts, and it is funding that with borrowing: a $430 million convertible note due 2030 priced at 2.50%, $250 million of convertible debt from Blackstone, and gross debt already near $1.2 billion. Capacity-led growth only works if the spread between the return on new megawatts and the cost of the capital funding them stays positive, and that spread narrows fast if utilization on new builds lags or financing costs rise. The company is, in effect, pre-spending years of cash flow on the bet that the AI orders keep coming and that customers honor long-dated contracts.

The earnings picture underneath is thinner than the EBITDA headline suggests, and that is the honest read on profitability. Trailing operating profit is around break-even, with interest coverage slightly negative on one basis, and the two measurement bases for operating income diverge by nearly 300%, a sign of how noisy the bottom line is in a heavy-depreciation, heavy-interest business. Adjusted EBITDA is real and growing, but it sits well above the cash that actually reaches equity holders after the depreciation on a multi-billion-renminbi asset base and the interest on the debt funding it. A reader who anchors on the 33% EBITDA margin without subtracting those costs will overstate how profitable VNET actually is today.

Then there are the risks specific to a Chinese ADR in a strategic sector. The customer concentration is stark: a single internet customer accounts for the overwhelming majority of recent AI bookings, so the order book is only as safe as that one relationship and that customer's own capex plans. The CATL-linked stake brings capital and credibility but also ties the company more tightly to China's industrial policy and to the geopolitics around data centers and AI, and US-listed Chinese shares carry their own delisting and disclosure overhang. The price already reflects a great deal of optimism about how this build resolves; if AI demand cools, financing tightens, or policy shifts, the leverage that powers the upside works just as hard in reverse.

Valuation

The standard operating-income inversion does not work for VNET, and it is more honest to say so than to dress it up. The model reads about 136 times company-wide operating income and an implied 23-year growth horizon, but that is the artifact of a near-zero, even slightly negative, trailing operating profit, and the reliability on the solve is explicitly low. The two operating-income bases here diverge by almost 300%, which is the clearest signal that this is not a business to value off an operating multiple. The right anchors are EBITDA, capacity and the capital plan.

On those terms the picture is coherent if demanding. Adjusted EBITDA of RMB891.5 million in the quarter, guided to RMB3.55 to RMB3.75 billion for the year, sits on a 907-megawatt installed base growing toward 1.4 gigawatts, funded by roughly $1.4 billion of annual capex. The valuation question is whether the returns on that new capacity, once utilized, exceed the cost of the debt and equity funding it. The X-ray reinforces the caution: with so few applicable methods and asset-based readings far above the price, no family supports the quote on current fundamentals; the price is a forward bet on the capacity filling.

The external anchors lean bullish but with a wide spread. The handful of covering analysts carry a strong-buy consensus and an average target well above the current price, the CATL-linked group agreed to buy in at about $1.4486 per Class A share, and the 2030 convertible was struck at a $13.75 ADS conversion price, all reference points above $8.75. The honest synthesis is that VNET is an EBITDA-and-capacity growth story whose equity value depends on a high-leverage build executing into durable AI demand. It is not value-supported on today's earnings; it is a wager that the contracted megawatts and the AI order book turn into cash before the debt comes due.

Catalysts

The strategic catalyst is the CATL-affiliated investment, a deal valued at up to $942 million for a stake of about 38%, expected to close in the fourth quarter of 2026, with the founder retaining operational control. Closing, and the capital and credibility it brings, is a key event, as is any change to the terms or timeline.

The operating catalyst is the AI order book converting to revenue. VNET has booked 519 megawatts of new AI and hyperscale orders year to date, mostly from one leading internet customer in Greater Beijing, against 907 megawatts in service at about 76% utilization. The pace of new bookings, the utilization ramp on delivered capacity, and the plan to add another 450 to 500 megawatts this year are the checkpoints that confirm whether the build is filling as fast as it is being funded. Q1 2026 set the trajectory: revenue up 19.8% to RMB2.69 billion and adjusted EBITDA up 30.6% to RMB891.5 million, with full-year guidance of 15.6% to 18.6% revenue growth.

Financing and external risk frame the rest. The roughly RMB10 to RMB12 billion capex plan is funded with convertible debt, including a $430 million 2030 note and Blackstone-provided convertibles, so the cost and availability of financing is a recurring catalyst. Watch customer concentration, China data-center and AI policy, and the general overhang on US-listed Chinese shares, all of which can move the stock independently of operating results.

Peer Cohorts (Per Segment, With Filing Citations)

Core business (reported)

Methodology Note

Fundamentals sourced from SEC EDGAR filings. Current price from Databento. The priced-in inversion and valuation x-ray are computed by the boothcheck engine; narrative composed by AI from the structured data.

View the full interactive VNET report on boothcheck