NEXTDC FPO [NXT] NXT.AX

Business Summary

NEXTDC owns and operates data centres in Australia and, increasingly, South East Asia, selling colocation, power, interconnection and wholesale capacity to hyperscale cloud providers, enterprises and government. It earns revenue three ways: retail colocation and interconnection, where customers rent racks and cross connects on shorter contracts at higher rates per megawatt; wholesale and hyperscale capacity, where large cloud and AI customers contract whole halls on 10 to 15 year initial terms with extension options out to 35 years; and power pass through, billed at cost, which inflates reported revenue without adding margin. Scale is now material, with FY26 statutory revenue of $496.5m, net revenue excluding power pass through of $405m and underlying EBITDA of $248.8m, against a market capitalisation of roughly $7.3b at the current $6.53 share price.

The competitive position rests on scarcity and certification. NEXTDC operates Australia's only Uptime Institute Tier IV certified network, which matters because hyperscale and government workloads increasingly require that standard, and it has assembled land and power banks that are difficult to replicate. The customer mix has shifted decisively toward hyperscale, with more than 70% of billing capacity now hyperscale workloads, which lowers the blended rate per megawatt but raises volume, lengthens contract duration and, using liquid to chip cooling, lowers the cost per megawatt to deliver. The investment consequence is that NEXTDC is now a capitalised infrastructure developer with a contracted revenue backlog rather than a high margin colocation operator, and it should be valued on contracted capacity and stabilised yield rather than near term earnings.

Investment Summary

We rate NEXTDC a Buy with a 12 month target of $8.00, implying 23% upside from $6.53. The shares have fallen 54% in three months and 58.6% over twelve months to a near nine year low, including a 42% single day fall on 2 November 2026 after the outlook cut, and the de rating has run well past the change in fundamentals. Three things drive the call: a contracted backlog of 740MW against 175MW currently billing, which is 4.2x revenue visibility on binding contracts; FY27 EBITDA guidance of approximately $400m, up 60% on FY26's $248.8m, with management calling FY28 EBITDA of $700m to $800m directionally reasonable; and a price to book of 0.8x against a two year average of 2.3x, or 2.4 standard deviations below, on a net asset base of $6.1b.

What the market is pricing is a broken growth story rather than a delayed one. On our numbers the shares trade on 20.5x FY27 EV/EBITDA falling to roughly 11x FY28 and 8 to 9x FY30, against Equinix and Digital Realty on 20 to 25x with materially slower growth, and the funding is already in place: $8.7b of liquidity, no debt maturities until FY30, a weighted average maturity of about five years and no leverage ratio covenants. The market is treating the M3 Melbourne power delay as a demand problem when it is a timing problem, and it is discounting the fact that 495MW was contracted in FY26, the largest year in the company's history. The call would be wrong if power procurement slips beyond the February 2027 Jemena restart date, or if the S4 and S7 build forces another equity raise at these levels.

Financial Performance

FY26 revenue rose 16% to $496.5m, net revenue rose 16% to $405m and underlying EBITDA rose 15% to $248.8m, about 4% above the top end of guidance. Statutory net profit after tax of $82.1m looks healthy but is an accounting artefact: it includes a $129m fair value gain on investment properties at M3 and S4 plus recognition of previously unrecognised tax losses, and the operating business remains loss making at the EBIT line with an operating margin of -7.8%. Gross margin of 65.9% and an underlying EBITDA margin of 61% on net revenue show the unit economics are sound; the loss sits below that, in depreciation on a rapidly growing asset base and in facility and corporate costs that each rose 17% to build ahead of capacity. We flag a data quality problem once here: the vendor files conflict on price ($6.53 current against a stale $13.36), on consensus EBIT ($12.8m against -$36.9m), on NTM EPS (-$0.10 against -$0.32) and on share count (760m against an implied 1.11b), so we anchor on the current price and treat the market capitalisation as approximately $7.3b.

The balance sheet is the story. Total assets reached $10.2b, comprising $5.8b of property, plant and equipment and $3.2b of investment properties, against net assets of $6.1b and available liquidity of $8.7b. Net debt is only $926m, or 0.07x EV, and all drawn senior debt and subordinated capital is fully hedged. The quality read is that NEXTDC has converted an equity story into a funded infrastructure platform, but returns are still thin: ROE of 1.6% and a trailing EV/EBIT of 865x are not meaningful measures of the business today, because the asset base has been built years ahead of the revenue it will earn.

Outlook and Estimates

The forward trajectory is unusually visible for a company this size. Management has guided to 197MW of new billing activation in FY27, which would more than double the 175MW billing today, and to approximately $400m of EBITDA, and it expects billing utilisation to reach 740MW by FY30, more than four times current levels, with over 400MW of build capacity added across FY27 and FY28. The forward order book stands at 565MW, every megawatt of which is a binding customer contract, and new hyperscale and AI deployments are targeting better than 10% yield on cost against data centre cap rates of 6 to 7%, implying a development spread of 1 to 1.5x on capital deployed. On the sell side, consensus revenue of $490.6m looks stale and inconsistent with that activation schedule, and NTM EPS has been revised down 36% in the past month, so estimates are still catching down to the M3 delay rather than reflecting the FY27 ramp.

What has to go right is narrower than the share price implies. First, power must be restored at M3, where distributor Jemena has suspended infrastructure upgrades on community opposition and does not expect to resume before February 2027; a slip beyond that puts a meaningful slice of the 197MW activation at risk. Second, the 400MW plus of build across FY27 and FY28 has to be delivered on time, which is a construction and supply chain question rather than a demand one. Third, the S4 and S7 joint venture has to be formed with third party capital, which would move roughly $15b of capex and $10b of associated debt off the parent balance sheet and crystallise value without further shareholder dilution. The FY27 number carries what management called an extremely high degree of certainty; FY28, which is where the valuation case really sits, is deliberately unguided.

Valuation

On forward multiples the stock is cheap on the numbers that matter and uninvestable on the ones that do not. NTM P/E is negative, at -$0.10 on the ASX spreadsheet and -$0.32 on the latest consensus, so the multiple carries no information, and trailing EV/EBIT of 865x is an artefact of a near zero EBIT base. EV/EBITDA is the right lens: 31.6x on the vendor's forward basis, which is computed on stale consensus EBITDA of roughly $260m, 20.5x on management's FY27 guidance of $400m, roughly 11x on FY28 EBITDA of $700m and 8 to 9x on FY30 EBITDA approaching $1b, all on an enterprise value of about $8.2b. Against its own history, forward EV/EBITDA of 31.6x sits 74% above the two year average of 18.2x and forward EV/revenue of 9.6x sits 33% above its 7.2x average, but price to book of 0.8x is 64% below the 2.3x average, or 2.4 standard deviations. The mechanical mean reversion prices are $2.49 on EV/EBITDA and $18.36 on price to book, which tells you the historical averages are not a useful guide in either direction: the EV/EBITDA premium reflects a depressed forward EBITDA base, and the 0.8x book multiple is the more honest anchor for an asset heavy developer.

We get to our $8.00 target by applying 18x to FY28 EBITDA of $700m, the midpoint of the range management called directionally reasonable, giving an enterprise value of $12.6b, deducting $2.5b of net debt as the S4 and S7 build draws down, and discounting the resulting $10.1b of equity back twelve months at 9.5%. A DCF on a 9 to 10% WACC and 2 to 3% terminal growth, run off stabilised FY30 EBITDA of $900m to $1b and a 15x exit multiple, produces a similar $7.50 to $8.50 range, and the NAV cross check is supportive: the roughly $9.2b of historical land, building and fit out spend is worth $12b to $14b on stabilised yields, which management agreed was directionally right. Against peers, Equinix trades on 59.7x trailing P/E and a $100.4b market cap and Digital Realty on 74.2x and $68.1b, both at roughly 20 to 25x EV/EBITDA on low double digit growth, while MAQ.AX at 39.6x trailing P/E and a $1.3b market cap is a profitable but far smaller domestic peer and not a useful comparable. NEXTDC deserves a discount to Equinix and Digital Realty on earnings quality, given negative EPS and negative free cash flow, but a premium on growth, and on growth adjusted terms it screens cheap at 20.5x EV/EBITDA against 60% EBITDA growth. We would expect it to trade at a modest discount to the global majors on FY27 and at a premium by FY29, which is what our target embeds.

Catalysts

The next six to twelve months turn on four events. The first and largest is M3 power: Jemena's works are suspended until at least February 2027, so any earlier restart, or any firming of the timeline, is the single biggest swing factor for the shares. The second is the FY27 interim result in February 2027, which will be the first hard evidence that the 197MW of activation is converting into net revenue and EBITDA. The third is formation of the S4 and S7 joint venture, where management has confirmed capital partnership discussions are live and where the equity stake retained is undecided, with 20% at one end of the spectrum; a deal would de risk the funding and crystallise the development spread. The fourth is the annual revaluation cycle for qualifying assets, which added $129m to FY26 statutory profit and will move book value, and therefore the 0.8x price to book, each year. We would also flag possible index related flows given the 54% drawdown, and treat a further capital raise as a negative catalyst rather than a positive one.

Key Risks

The most important risk is power procurement, and it is not fully within management's control. The M3 Melbourne delay stems from Jemena suspending infrastructure upgrades on community opposition, with no resumption expected before February 2027, and power availability is the binding constraint on the entire Australian build out. If M3 slips beyond that date, or if similar community or grid constraints appear at S4, S7 or M5, the 197MW FY27 activation target and the approximately $400m EBITDA guidance come under pressure, and we would expect the shares to retest the $5.07 52 week low. A break below $5.07 on a guidance cut to below $350m of FY27 EBITDA would invalidate the thesis.

Funding and dilution are the second risk. The company raised $9.75b in FY26 across $5.8b of senior debt, $750m of subordinated notes, $1.7b of hybrids and $1.5b of equity, and S4 and S7 alone require roughly $15b of capex against $10b of expected debt. Net debt to EV of 0.07x today flatters a balance sheet that will gear up materially, and while there are no leverage ratio covenants and no maturities until FY30, an equity raise at $6.53 would be far more dilutive than the FY26 raise. Third, the mix shift to hyperscale cuts both ways: more than 70% of billing is now hyperscale on lower per megawatt rates, which concentrates the customer base and hands pricing power to a handful of global cloud buyers. Fourth, execution on over 400MW of build in two years carries construction, supply chain and cost inflation risk. Finally, a 50bp expansion in data centre cap rates would reduce the value of the $3.2b investment property book by roughly 7%, flowing straight through to book value and to the revaluation line in statutory profit.