Goodman Group

Phase A — Understand the business

Company Overview

Goodman Group is a vertically integrated global industrial-property group — it owns, develops and manages logistics and, increasingly, data-centre real estate across Australia, Asia, Europe, the UK and the Americas. It is not a passive landlord REIT; it is best understood as three businesses stacked on one balance sheet, and the mix is the whole story:

• Investment (owned property): rental income from assets Goodman holds directly — ~15% of operating profit.
• Management (fund/partnership fees): recurring fees on ~$75bn of third-party capital run through partnerships (GADP, GAIP, the European DC partnership, etc.) — ~30% of operating profit. This is the capital-light, annuity leg.
• Development (build-and-sell / build-and-hold): the profit engine — ~55% of FY25 operating profit, now dominated by high-margin data centres and multi-storey logistics in land-constrained cities.

The model: Goodman assembles powered land (its "power bank"), develops shells/fit-outs, seeds them into partnerships funded by sovereign and institutional capital, and keeps the management fee plus a development margin — recycling its own equity rather than owning every finished asset. Total portfolio (owned + managed) is ~$85–87bn. External AUM $72.1bn at FY25 → $75.2bn at HY26.

Key products: logistics warehouses; multi-storey "infill" logistics; powered shells and fully-fitted hyperscale data centres. Main customers: e-commerce/3PL logistics tenants historically; increasingly hyperscalers and AI/cloud customers for the data-centre pipeline. Suppliers: the binding one is power — utilities and grid connections; energy availability is management's stated #1 constraint. Competitors: Equinix, Digital Realty, NextDC, Vantage/DataBank (also a JV partner), CyrusOne, Prologis (logistics) and every hyperscaler self-build programme.

Contract structure / payment terms: development income is lumpy and recognised over the build, not recurring; management fees are recurring and sticky; rental income is contracted with occupancy at 95.9%. The data-centre book is still pre-lease — 370MW in WIP but management describes itself as "now in the customer phase," advancing leasing to end-CY2026. That is the single most important sentence in the file: the shells are being built ahead of the tenants.

Supply Chain

Upstream → Goodman → end customer, with the actual chokepoints named:

Upstream inputs:

• Land — Goodman's decades-assembled infill logistics landbank in supply-constrained cities (Sydney, LA, Tokyo, Paris, Amsterdam, Hong Kong, London) is the raw material it converts to data-centre sites. This is a genuine single-source-style advantage: you cannot buy this landbank today at yesterday's prices.
• Power — the true chokepoint. The "power bank" is 6.0GW secured/advanced across 16 cities at HY26 (up from 5.0GW/13 cities at FY25), of which only a fraction is contractually secured. Grid connection queues, substation capacity and utility timelines gate everything. Management: "energy availability poses the most significant constraint".
• Construction / equipment — GCs, and the data-centre fit-out supply chain (cooling, switchgear, generators — i.e. Vertiv/Eaton/Schneider-class vendors, transformers). Goodman is a buyer of, not a seller into, this chain; long-lead electrical gear is an execution risk.
• Capital — the fourth input. Development at this scale is funded by (a) Goodman's own equity (the $4bn Feb-2025 placement ) and (b) partnership capital from sovereign/institutional LPs.

The company: master-plans, develops, and either holds or seeds into partnerships.

Downstream / end customers: hyperscalers and AI/cloud operators (the pre-lease targets), plus logistics tenants (3PLs, e-commerce, retailers) on the industrial book. Named partnership counterparties: a $14bn European data-centre partnership, a $2bn US logistics partnership, and a DataBank JV in Los Angeles (a 32MW facility targeted for 2027).

Chokepoints / single-source dependencies: (1) power — no power, no data centre; (2) hyperscaler demand — the customer set is a handful of names, so concentration on the DC book will be high once leases sign; (3) capital-partner appetite — the fee engine needs sovereign LPs to keep funding; a risk-off turn in that channel would force more on-balance-sheet holding.

Competitive Advantages (moats)

The moat is the land-plus-power bank, not the buildings. Anyone can pour a data-centre shell; almost no one holds 6GW of secured/advanced power across 16 tier-1 global cities sitting under land Goodman already owns. That combination — infill land in the exact metros where power and permitting are scarcest — is the durable edge. Reproducing it requires both a decades-old landbank and years in grid-connection queues.

Secondary moats:

• The capital-management flywheel. ~$75bn AUM generates recurring fees and lets Goodman build at scale without carrying every asset — a structurally higher return on its own equity than a pure-play developer, and a relationship moat with sovereign LPs.
• Balance-sheet strength as a weapon. Gearing 4.1–4.3% (look-through ~17%) and ~$5–6.6bn liquidity mean Goodman can secure power and land ahead of demand — a luxury leveraged peers and cash-strapped neoclouds lack. In a capital-intensity race, the lowest cost of capital wins.
• Switching costs / stickiness: modest on logistics (commodity space), potentially high on data centres once a hyperscaler is anchored and fitted.

Bargaining power: strong over capital partners (they need Goodman's origination and land); weak-to-improving over hyperscaler tenants — the DC book is still unleased, so today the customers hold the cards. That balance flips only if power scarcity makes Goodman's powered shells genuinely scarce. Weak over utilities — Goodman needs power more than the grid needs Goodman.

Rivals compare unfavourably on cost of capital (DLR, neoclouds) or on land/permitting optionality, but favourably on operating track record — Equinix and Digital Realty actually run data centres at scale; Goodman is still proving it can lease and operate, not just build.

Segments

our figures is empty; the split below is, by profit contribution and geography, and should be treated as approximate.

By business line (share of FY25 operating profit):

By development mix: data centres = 57% of WIP at FY25 → 73% at HY26. Total WIP $12.9bn (FY25, 57 projects) → $14.4bn (HY26, ~51 projects) → guided ~$18bn by Jun-2026. The project count falling while WIP value rises tells you the mix is shifting to fewer, far larger (data-centre) projects — higher value, higher execution concentration.

By geography: development sites span the US, Europe/UK and Australia; the power bank's 16 cities include Sydney, Melbourne, LA, Tokyo, Paris, Amsterdam, Hong Kong and London. No clean geographic revenue split is available — n/a at segment-P&L granularity.

Trend & cause: the whole group is being re-geared from a logistics landlord into a data-centre developer, funded by the Feb-2025 raise and partnership capital. Development income share is rising because DC yield-on-cost (8.1%) exceeds legacy logistics, and because Goodman is originating more development on its own balance sheet ahead of seeding it into funds — which lifts both the reward and the risk (see Lens 10).

---

Phase B — Measure performance

Earnings Result — HY26 (half to 31 Dec 2025, reported 19 Feb 2026)

The print:

• Operating profit $1.22bn, +8% YoY (one source frames the reported line as +1.6% on a like-for-like operating basis — the +8% is the headline the trade press ran).
• Operating EPS 63.8c; interim distribution 15.0c.
• Operating EBIT ~$1.27bn; property income +17%; occupancy 95.9%.
• Statutory profit $799.8m, swinging back from a ~$202m prior-period loss — the swing driven by positive revaluations (+$252m in the half) after the prior downturn.
• External AUM $75.2bn; WIP $14.4bn, DC = 73%; yield-on-cost 8.1% (up from 7.5%); 370MW DC in WIP, 1,826MW activated in total.
• Power bank 6.0GW / 16 cities; ~0.5GW expected underway by Jun-2026.
• Gearing 4.1%, look-through ~17%; liquidity ~$5.2bn.
• FY26 guidance reaffirmed: +9% operating EPS (→ >$2.6bn operating profit).

Drove it: development income (early recognition of development + performance fees pulled some profit forward ), plus higher management fees on a bigger AUM base and positive revaluations turning the statutory line green.

Balance-sheet flags: cash flow vs. earnings is the thing to watch — a growing share of profit is development profit recognised on assets Goodman increasingly holds on balance sheet pre-completion, so operating profit can run ahead of realised cash until those assets are seeded into funds or sold. Gearing is genuinely low (a real strength), but that is reported gearing; look-through (~17%) and the future funding need for an ~$18bn WIP book are the real leverage story.

Market reaction: shares fell ~6% on the day despite a solid result. The tell: the market was not disappointed by the numbers (income +17%, guidance reaffirmed) but by the absence of a fresh data-centre leasing/partnership catalyst and broad market weakness — a stock priced for conversion needs conversion news, and the half delivered building, not leasing. That is the whole GMG setup in one candle.

Earnings Calls (sentiment trend)

Across the last ~3–4 calls (FY24 → Q1-FY25 → FY25 → HY26/Q3-2026), the arc is a deliberate narrative migration from "industrial property group" to "digital infrastructure / AI-infrastructure platform":

• Recurring, rising phrases: "power bank," "essential infrastructure," "AI," "customer phase," "energy availability," "capital partners," "yield on cost." The GW number is now the headline KPI management leads with — a REIT that talks in gigawatts.
• Things they started saying: in HY26, "we are now in the customer phase" (i.e. shift from securing power/land to leasing), and "energy availability is the most significant constraint" — both new framing that concedes the two open questions (will tenants come, can power be delivered).
• Things they stopped emphasising: the pure logistics rental story; new-partnership formation ("not planning new partnerships in the near term," pivoting to long-term hold structures and transferring WIP into AUM).
• Tone: confident-to-promotional on the structural demand ("AI is a significant factor in hyperscale customer negotiations"), but increasingly candid on the two gates (power, leasing timing). FX headwinds flagged as a swing factor.

Net: management sentiment is bullish on demand, disciplined on capital, and — tellingly — has begun pre-managing expectations on the timeline ("through to end of this calendar year"). That candour is a modest positive for trust and a modest negative for near-term catalyst density.

Comps

Goodman is a hybrid — a fund manager + developer + landlord — so no single peer is clean. Data-centre/infra REITs are the closest read; note GMG trades on operating EPS (Australian REIT convention), while the US peers quote AFFO, so multiples are not strictly like-for-like. Multiples with source/date; where unsourced, n/a.

Read: on a forward-earnings basis GMG's ~22.7x is below EQIX/DLR/AMT — which looks cheap until you remember (a) US peers earn recurring AFFO from leased, operating assets while a rising share of GMG's earnings is development profit on unleased shells, and (b) GMG's ~1.0% yield is far below the 1.9–3.8% infra-REIT range, so you are paid almost nothing to wait. GMG screens as a growth stock wearing a REIT's ticker: cheaper than peers on forward earnings, dearer on quality-of-earnings and income. The dispersion in analyst targets (below) reflects exactly this ambiguity.

Stock-Price Catalysts

What actually moves GMG over recent years:

• The data-centre pivot narrative (2024–25): the single biggest re-rate driver — GMG became the ASX's flagship "AI-infrastructure" proxy, dragging the stock from the low-A$20s toward the mid-A$30s and a 52-week high of A$37.31.
• The $4bn equity raise (Feb 2025): funded the DC build-out; raises like this cap the stock short-term (dilutionDilutionIssuing new shares, so each existing share owns a smaller slice of the same company.) but were read as a growth signal.
• Earnings prints & guidance: GMG reliably reaffirms/upgrades ~9–11% operating EPS growth; the reaffirmation is priced, so beats need a fresh DC catalyst to move it — hence the −6% on the solid HY26 result (no new leasing/partnership news).
• DC contract wins / new capital partnerships: Macquarie explicitly names these as the upside catalysts. A signed hyperscaler lease or a new sovereign DC partnership would be the next leg.
• Macro / rates & risk-off: as a long-duration property-plus-development name, GMG trades with the AI-infra complex and with rates; the 52-week range A$24.56–37.31 shows how macro-sensitive the multiple is.

Pattern: the market reacts to conversion of the pipeline into signed demand, not to the (already-priced) earnings cadence. GMG is a "show me the tenants" stock.

---

Phase C — Judge people & books

Management

Greg Goodman (Founder & CEO). Founded the group in 1989, CEO since August 1998 — ~27 years at the helm. This is a founder-led company in the truest sense.

1. Track record: built Goodman from an Australian industrial landlord into a top-tier global logistics platform (~$85bn+) and is now executing a bet-the-franchise pivot into data centres. Delivered ~9–11% operating EPS growth through multiple cycles; navigated the 2022–23 property-value downturn (a $3.4bn revaluation drawdown across group + partnerships) without breaching the balance sheet. Quantified: FY25 operating profit $2.31bn (+13%), NTA per security $11.03 (+25%).
2. Tenure & skin in the game: Greg Goodman holds ~1.82% directly, ~A$1.07bn of stock, and added 630,000 performance rights in Nov-2024 — real, large, aligned ownership. KMP are required to hold fixed remuneration in Goodman securities. No single external holder >10% (State Street ~8.85%, Vanguard ~8.90%, BlackRock ~8.58%) — index/institutional register, founder retains outsized influence.
3. Capital allocation: the standout — disciplined, capital-light, counter-cyclical. Keeps reported gearing ~4%, recycles capital through partnerships rather than owning everything, raised equity before the DC build rather than levering into it, and grows fee AUM. Reinvestment-led (low payout: DPS 30c vs OEPS 118c → ~25% payout, funding growth). ROE mid-teens on a low-leverage base is a genuinely good return for property.
4. Red flags: the honest ones are strategic, not governance — (a) a founder deep into a bet-the-company pivot into a business (operating data centres for hyperscalers) Goodman has not previously operated at scale; (b) rising on-balance-sheet development lifts risk if the cycle turns; (c) low ~1% yield concentrates the return in capital growth, i.e. in the pivot working. No related-party-deal, comp-scandal or accounting-restatement flags surfaced.
5. Archetype: owner-operator founder — long horizon, skin in the game, willing to reshape the franchise. For this stage (scaling a capital-intensive new vertical) a founder with a low-gearing instinct is close to the ideal operator; the risk is founder conviction outrunning the leasing evidence.

Forensic Red Flags

Forensic lens on the accounting. Ground truth: no filings on disk (ASX filer, no CIK), so this is + structural reasoning on the REIT-developer model.

• Revenue recognition / earnings quality (the main flag): a rising share of operating profit is development profit (~55% in FY25), and HY26 explicitly benefited from "early recognition of development and performance income". Development profit recognised over a build — especially on assets held on balance sheet before completion — is lower-quality than recurring rent or fees: it depends on completion, on end-valuations, and on eventually finding a buyer/partner. Operating EPS can run ahead of realised cash. This is the single item a forensic analyst underlines.
• Fair-value / revaluation dependence: statutory profit swings on property revaluations (−$202m prior period → +$799.8m HY26 on +$252m revaluations; a $3.4bn drawdown in the downturn). The statutory P/E of 37.3x vs operating P/E ~22.7x is the visible gap between IFRS mark-to-market earnings and management's "operating" measure — neither is wrong, but the reader must know which is which.
• Off-balance-sheet / look-through leverage: reported gearing ~4% flatters the true picture; look-through gearing ~17% captures partnership debt. Fees and asset values in the ~$75bn managed book carry leverage that doesn't hit GMG's own balance sheet — a feature (capital efficiency), but one that hides the group's real economic leverage.
• Cash flow vs earnings: with WIP guided to ~$18bn and DC assets increasingly held pre-lease, watch for operating cash conversion lagging reported operating profit and for development-asset carrying values (are pre-let assumptions embedded in valuations?).
• SBC / performance rights: KMP performance rights and equity-linked pay exist but were not flagged as excessive; dilution modest.

Regulatory findings (required):

• SEC EDGAR (LR + AAER): none possible — Goodman has no CIK and does not file with the SEC. Zero SEC findings.
• Non-SEC / web search: a targeted search for "Goodman Group" (ASIC OR DOJ OR settlement OR fine OR penalty OR investigation) returned no material enforcement action against GMG.AX. Note: the search surfaces abundant US litigation against "Goodman" the HVAC/air-conditioning manufacturer (an $803m coil class action, a $5.55m DOJ fire-hazard settlement) and unrelated entities like "Goodman Group Lawyers (in liquidation)" — none of these is the ASX-listed Goodman Group. Provenance care matters here.
• Item 3 / Legal Proceedings equivalent: no material litigation disclosure surfaced in the FY25 reporting coverage.
• Verdict: No material regulatory or legal findings — verified via SEC EDGAR EFTS (no CIK, N/A), targeted web search (ASIC/DOJ/settlement), and FY25 reporting coverage as of 2026-07-06.

---

Phase D — Project & stress-test

Forward Projection

Built bottom-up from FY25 actuals + management's own guidance. No our model create in this unattended watchlist run (per skill rules).

Anchors: FY25 operating EPS 118.0c (+9.8%); FY26 guidance +9%; ~2.04bn securities out; management running toward ~$18bn WIP with DC yield-on-cost 8.1%.

Logic: the base case is essentially "management hits ~9% because they always have and the WIP backlog supports it" — a defensible extrapolation of a low-gearing, fee-plus-development compounder. The bull–bear spread widens sharply from FY27 because that is when the data-centre bet resolves: bull = the 6GW power bank leases up and converts development WIP into recurring managed-asset income and fees; bear = leasing slips or hyperscaler capex cools, on-balance-sheet DC shells sit unleased, and development income (the ~55% engine) decelerates just as the asset base is heaviest. EPS growth is guided and bankable near-term but bet-dependent medium-term.

Provenance guard: forward multiples/consensus EPS in dollar terms beyond guidance are off the stated growth rate — I did not source a clean Visible-Alpha-style consensus EPS series, so treat FY27–28 as scenario arithmetic, not sourced consensus.

Brier-loggable base call (not logged this run): "GMG.AX FY26 operating EPS ≥ 128c (i.e. ≥ +8.5%), p ≈ 0.80, resolves 2026-08 (FY25 result date)" — high confidence given the reaffirmed guide.

Bull vs Bear

Bull case. Goodman is the best-capitalised, best-located data-centre developer on the planet that isn't a hyperscaler. It holds ~6GW of power under prime infill land in 16 cities where power and permitting are the scarce goods, funds the build with ~4% gearing and a de-risked $4bn war chest, and monetises three ways (develop, manage, own) so returns on its own equity dwarf a pure-play developer's. As AI-driven demand collides with a power-constrained world, powered land becomes the choke-point asset, and Goodman owns more of it than anyone. Convert even half the pipeline and the fee-AUM base steps up permanently, re-rating the recurring earnings. Founder-operator with A$1bn of skin in the game, ~9–11% EPS growth through cycles, NTA compounding (+25% in FY25). The market pays ~22.7x forward — below US infra-REIT peers — for arguably the highest-quality growth in the group.

Bear case (permanent-impairment risks):

1. The shells are unleased. The DC book is pre-lease ("customer phase"); if hyperscaler leasing slips or AI capex intentions cool, Goodman is left holding on-balance-sheet, capital-intensive assets earning nothing — and development profit (~55% of the P&L) decelerates hard just as WIP peaks at ~$18bn.
2. Development earnings are cyclical and lower-quality. A property-plus-development stock priced at ~23x forward is priced like a compounder but earns like a cyclical developer; a downturn compresses both the earnings and the multiple (double-hit). The prior cycle already showed a $3.4bn revaluation drawdown.
3. Yield / power / execution. Yield-on-cost (8.1%) can compress if construction and power costs rise or if DC supply outruns demand; energy delivery is the binding constraint management itself flags, and long-lead grid/electrical timelines can slip 12–24 months.

Pre-mortem (18 months out, thesis broken): It's early 2028. Hyperscaler leasing came slower and at lower rents than modelled; two flagship DC shells sit substantially unleased on balance sheet; a broad AI-capex "digestion" year cooled sentiment across the complex; development income fell as a share of profit and cash conversion visibly lagged reported OEPS; the multiple de-rated from ~23x to the mid-teens even as EPS merely flatlined. The stock is down 30–40% not because Goodman did anything reckless, but because it was priced for flawless conversion and got merely-good.

Are multiples too high? Relative to US infra-REIT peers, no (cheaper on forward earnings). Relative to the risk that ~55% of earnings is unleased-development-dependent and you're paid ~1% to wait, arguably yes. The multiple embeds successful conversion.

Contrarian view (what the market refuses to see): The bulls treat the 6GW power bank as demand secured; it is supply optionality, and optionality is only worth something if exercised at a profit. The market is under-weighting the timing gap between building powered shells and signing hyperscaler leases — and over the next 12 months that gap, not the total-addressable-market story, is what sets the stock. Conversely, the bears under-weight that powered land in tier-1 cities is a genuinely scarce, appreciating asset even if leasing slips — the downside is cushioned by real-asset value in a way a neocloud's GPUs are not.

Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks the model: the model needs continuous partnership-capital appetite and hyperscaler leasing. Choke either — a risk-off turn in sovereign LP allocations, or a hyperscaler capex pause — and the flywheel stalls: development profit (the ~55% engine) is the most exposed line, and it's the one holding up the ~9% growth guide.
• Revenue concentration: legacy logistics is diversified, but the growth is concentrating into a handful of hyperscaler tenants across a few very large DC projects (WIP value rising, project count falling). Once leases sign, tenant concentration on the fastest-growing book will be high; lose or delay one anchor and a whole project's economics wobble.
• Why the moat may be weaker than bulls think: Goodman is a superb developer and capital allocator, but an unproven data-centre operator. Equinix and Digital Realty actually run these assets at scale for enterprise and cloud customers; Goodman's edge is land and power, not operations — and hyperscalers increasingly self-build or dictate terms, capturing the operating margin and leaving Goodman as a powered-shell landlord (thinner, more commoditised than the bull case assumes).
• Most dangerous competitor bulls underestimate: the hyperscalers themselves (self-build) and well-capitalised specialists (Vantage/DataBank — note DataBank is both partner and potential competitor). If power scarcity eases even modestly, Goodman's powered-shell premium erodes fast.
• Worst capital-allocation / accounting concerns: none egregious, but the shift to on-balance-sheet development pre-lease and early recognition of development/performance income are the two levers that flatter near-term OEPS while raising real risk. The gap between statutory P/E 37x and operating P/E ~23x is where the skeptic lives.
• Assumptions that must hold for today's price: (1) ~9% OEPS growth persists 3+ years; (2) the 6GW power bank leases up at healthy rents on schedule; (3) no cap-rate/rate shock re-rates real assets down; (4) capital partners keep funding. Break any one and the ~23x forward looks generous.
• If growth disappoints 20–30%: OEPS growth halving from ~9% to ~4–5% plus a multiple de-rate to mid-teens implies ~30–40% downside — the pre-mortem path. The ~1% yield offers almost no downside cushion.
• Single scenario that permanently impairs: a multi-year AI-capex "digestion" period in which hyperscalers stop signing new large leases while Goodman holds a heavy, part-unleased DC WIP book — forcing write-downs, stalling development income, and permanently resetting the growth algorithm from "compounder" to "cyclical developer." Plausibility: moderate — not a base case, but far from tail, and it's the one that turns a great company into a bad stock.

Management Questions (ordered by information value)

1. Of the 6GW power bank, how many MW are under signed hyperscaler leases today vs. secured-power-but-unleased — and what's the expected quarterly leasing run-rate through CY2026?
2. What share of the ~$18bn WIP is held on Goodman's own balance sheet vs. in partnerships, and where do you expect that split to be at completion?
3. How much of HY26 operating profit came from "early recognition" of development/performance income, and how should we think about the cash-conversion of operating EPS over the next two years?
4. On the DC book, are you the operator or the powered-shell landlord — and how does the economics (and your captured margin) differ between fully-fitted vs. powered-shell deals?
5. What is the stabilised return on incremental equity on a typical data-centre development vs. your legacy logistics developments?
6. If hyperscaler leasing slips 12–18 months, what happens to FY27–28 development income and to your on-balance-sheet carrying values?
7. What's your cap-rate / valuation sensitivity — how much does a 100bp move in cap rates move group NTA and look-through gearing?
8. You said you're not forming new partnerships near-term; at what pipeline scale do you need fresh sovereign capital, and what's the appetite you're seeing?
9. Where are the grid-connection and long-lead-equipment bottlenecks most acute, and where has your delivery timeline actually slipped?
10. How concentrated will the DC tenant base be at stabilisation, and what's your exposure to any single hyperscaler?
11. What return threshold would make you walk away from a data-centre site you've secured power for?
12. How do you protect development margin if construction and power costs keep rising faster than achievable rents?
13. What's the succession plan, and how does the founder-led capital-allocation discipline institutionalise beyond Greg Goodman?
14. How should investors reconcile the ~1% distribution yield with a capital-intensive growth phase — is a payout reset on the table?
15. What single leading indicator should we watch to know the data-centre bet is working (or not) 12 months before it shows in earnings?

---