NEXTDC builds and operates data centres across Australia, leasing rack space and connectivity to hyperscalers like Microsoft, Google, and Amazon. At A$14.34 versus our fair value of A$10.40, the stock trades 38% above what we think it is worth. The central question is not whether the business is good — it is — but whether the AI infrastructure boom is already priced in at a level that leaves little room for error.

NEXTDC is best suited to the growth or thematic investor with a three-to-five year horizon and high risk tolerance. The contracted pipeline makes the revenue trajectory unusually visible for a growth stock, but the investment requires patience through years of negative free cash flow, rising debt, and no dividends before the earnings inflection arrives. Investors who need current income, margin of safety, or near-term returns will find this stock frustrating. Those who can hold through the build phase and believe the AI infrastructure cycle has duration may find the wait worthwhile — but not at today's entry price.

NEXTDC operates 17 data centres across Australia, leasing space, power, and connectivity to hyperscalers, enterprise clients, and government agencies. It earns recurring revenue through long-term colocation contracts, with power costs passed directly to customers, insulating EBITDA margins from energy price swings.

The FY25 result delivered A$350m in net revenue and A$217m in underlying EBITDA — both in line with guidance. The more significant disclosure was the contracted pipeline: 416.6MW committed against only 119.8MW currently billing, a ratio of 3.5 times. The gap represents locked revenue that converts mechanically as new facilities open over the next three years.

The investment case rests on that conversion. When 296.8MW of pre-sold capacity begins billing, EBITDA is forecast to reach A$387m by FY28 — nearly double today's level — on largely fixed infrastructure. EBITDA margins should expand from 60% toward 65% as revenue scales over a cost base that barely moves. The risk is not demand; it is the A$1.9bn in annual capital expenditure required to build what has already been sold, and the A$5–6bn in peak net debt that accumulates before free cash flow turns positive around FY30.

At A$14.34 versus our fair value of A$10.40, the stock is 38% overvalued.

What happened?

FY25 revenue grew 14% to A$350m, supported by 9.5MW of new billing capacity added during the year. EBITDA margins compressed from 66% to 62% as the company deliberately expanded corporate headcount by 28% — building the organisation ahead of the next wave of capacity additions. The result was not a disappointment; it was a controlled investment in future delivery capability. Contracted capacity reached 416.6MW, up from 176MW a year prior, with the entire 264% increase in Victoria reflecting hyperscaler AI infrastructure commitments.

What's next?

FY26 guidance of A$390–400m in net revenue and A$230–240m in EBITDA is management's immediate test. Delivery requires converting roughly 35MW of new billing capacity during the year, weighted toward the second half as new facilities come online.

The real inflection arrives in FY27. New capacity additions from FY26 will generate their first full year of billing revenue, producing a 30% revenue step-change on a largely fixed cost base. EBITDA margins should recover sharply toward 63% and beyond as that operating leverage flows through.

Two events will determine the FY27–28 trajectory more than any financial metric. The first is whether NEXTDC announces a joint venture structure — known internally as JVCo — to co-fund the S4 and S7 campus builds in Sydney. A successful JVCo would remove up to A$2bn from the capital expenditure pipeline and materially reduce peak net debt. The second is the billing count at the June 2026 half-year result: fewer than 140MW billing would signal delays in the contracted pipeline conversion and warrant a reassessment of the FY27 outlook.

The current price of A$14.34 implies roughly 30 times FY28 EBITDA — equivalent to the price Blackstone paid for AirTrunk in September 2024, a private-market control transaction at peak AI demand sentiment. Our base case applies 25 times, which sits at the top of the 22–25 times range where Equinix and Digital Realty trade as mature, positive free cash flow businesses. The 38% overvaluation is therefore almost entirely a question of multiple: the market is pricing in a scarcity premium for Australia's only listed pure-play data centre operator, combined with an assumption that the JVCo capital recycling structure closes within 12–18 months and the RBA begins cutting rates by late 2027.

The biggest single risk is the joint venture. If JVCo fails to materialise, NEXTDC must fund the S4 and S7 campuses entirely from its own balance sheet. Net debt would approach A$6bn before free cash flow turns positive — and with no JVCo offset, an equity raise of A$600–800m at a discount becomes the most likely funding solution. The share count would grow, per-share value would fall, and the premium multiple would compress simultaneously. That scenario — which we assign a 30% probability — would likely see the stock re-rate toward our bear case of A$6.42.