Macquarie Technology

Phase A — Understand the business

Company Overview

Macquarie Technology Group is a founder-controlled Australian enterprise-IT and digital-infrastructure operator that has grown out of a boutique business-telco into an owner-operator of secure, sovereign data centres. Founded July 1992 by brothers David and Aidan Tudehope in the wake of Australian telecom deregulation; ASX-listed 1999; renamed from Macquarie Telecom Group → Macquarie Technology Group in May 2023 to signal the pivot from telco to data-centre/cloud.

It reports three operating segments, run through four customer-facing brands:

1. Data Centres (Macquarie Data Centres) — colocation in five owned/operated Sydney facilities, plus the new IC3 Super West build. The crown jewel and the entire investment case. FY25 revenue ~$74m, EBITDA $36.6m (+5.5% YoY).
2. Hybrid IT / Cloud Services & Government (Macquarie Cloud Services + Macquarie Government) — managed Azure, private/hybrid cloud, secure internet gateway, cybersecurity, and colocation sold to 42% of Australian Federal Government agencies. FY25 EBITDA $24.0m (+1.7% YoY).
3. Telecom (Macquarie Telecom) — business voice, data, mobile, SIP trunking, unified comms. The legacy cash cow. FY25 EBITDA $24.0m but guided DOWN to ~$20m in FY26 as NBN wholesale price cuts get passed through to customers.

Group FY25 (year to 30 June 2025): revenue $369.6m (+1.7%), EBITDA $113.6m (+4.1%), EBITDA margin 30.7%, NPAT $34.9m (+5.7%), EPS 135.2c. Eleven consecutive years of EBITDA growth; 22 consecutive half-years of EBITDA growth. 3-yr EBITDA CAGR 8.7%, revenue CAGR 6.1%.

Contract structure: colocation is recurring, multi-year, sticky (high switching cost once racks/network are installed). Government revenue is contracted and compliance-gated. Telecom is recurring but exposed to NBN wholesale repricing. Customer concentration is not publicly disclosed at the name level (our figures empty); the government book is diversified across dozens of agencies. ~450 employees; NRFC deal expected to add ~140.

Plain-terms verdict: a high-quality, sub-scale roll-up of Australian secure IT services, wrapped around a genuinely scarce asset — a certified-strategic, AI-ready Sydney data-centre campus — that the group is spending heavily to expand while the legacy telco funds the lights.

Supply Chain

Upstream inputs → MAQ → end customer, named where disclosable:

• Land & shell — Macquarie Park campus (owned; IC2/IC3 land+buildings acquired FY24) plus a put-and-call option on a large new Sydney data-centre land parcel for a future ~200MW campus. Owning the dirt is a structural advantage vs peers who lease.
• Power — the binding chokepoint. IC3 Super West brings the campus to 63MW with "all end-state power secured". In grid-constrained Sydney, secured grid connection is the scarce input; this is MAQ's single most valuable supply-chain position.
• Construction / fit-out — main works on IC3 topped out; A$350m build cost FY25→phase-1 practical completion. Contractor names n/a. Long-lead electrical/mechanical gear (switchgear, chillers, liquid-cooling CDUs) is the industry-wide bottleneck.
• Compute-adjacent hardware — MAQ is colocation, not a chip buyer; the customer brings GPUs. But IC3 is designed for air, liquid and hybrid cooling at AI rack densities, so MAQ's exposure is to the cooling/power supply chain, not to Nvidia allocation directly.
• Cloud upstream — the Hybrid IT segment resells/manages Microsoft Azure (Azure Managed Services), making Microsoft both a key supplier and, at the hyperscaler level, a potential competitor.
• Connectivity — Telecom segment buys NBN/wholesale access; NBN Co pricing is a direct margin input (see Lens 5).

End customers along the chain: Australian Federal & State Government agencies (42% of federal agencies), large enterprises, and — the future demand thesis — hyperscale cloud + SaaS-with-AI companies wanting to host AI workloads onshore in Australia.

Chokepoints / single-source dependencies: (1) Sydney grid power — secured for 63MW, but the next 200MW campus needs a fresh connection; (2) Microsoft Azure as both supplier and competitor to the cloud arm; (3) NBN Co as monopoly wholesale input to Telecom. Names present → this lens passes.

Competitive Advantages (moats)

Moat stack, strongest first:

1. Sovereign / "Certified Strategic" compliance moat — MAQ's entire data-centre portfolio holds the Australian Federal Government's Certified Strategic designation under the Digital Transformation Agency's Hosting Certification Framework. This is a regulatory licence to serve classified/government workloads that hyperscalers and generic colo cannot easily replicate. It is why MAQ has 42% of federal agencies. As data-sovereignty rules tighten (AI, critical infrastructure), this moat widens. This is the durable edge.
2. Secured Sydney power + owned campus — 63MW of secured end-state power on owned land in a supply-constrained market. Power is the true scarce resource in AI infrastructure; MAQ has it locked.
3. Switching costs — once an agency's or enterprise's secure workloads, network, and compliance certifications sit inside a Macquarie facility, migration is costly, risky, and re-certification-heavy. Reflected in a world-leading NPS of +71 — the highest of any ASX-listed company, which is both a moat (retention) and a marketing engine.
4. Vertical integration — MAQ can sell a government customer secure colo + managed cloud + cyber + telco as one accountable stack; hyperscalers and pure-play colos can't match the full-service, single-throat-to-choke model for regulated Australian buyers.

Bargaining power: Strong over customers in the sovereign/government niche (few compliant alternatives). Weak upstream — MAQ is price-taker on NBN wholesale, on Azure, and on grid connection terms. Against pure-play data-centre peers (NextDC), MAQ is far smaller in MW and lacks NextDC's hyperscale-scale relationships and balance-sheet firepower — so its moat is depth in a defensible niche, not breadth.

Fragility: the moat is national (Australia-only) and niche (sovereign/regulated). It does not protect the commodity colo or the shrinking telco. Scale-wise MAQ is a minnow next to NextDC/AirTrunk/Global Switch.

Segments

FY25 segment detail (year to 30 June 2025), all ``:

Reconciliation note: segment EBITDA ($36.6 + $24.0 + $24.0 = $84.6m) is below group EBITDA ($113.6m); the gap is corporate/unallocated and inter-segment items — MAQ's group EBITDA definition differs from the sum of the three reported segment lines (FY24 showed the same pattern: DC $34.7m + Hybrid IT $50.8m + Telecom $23.6m against group $109.1m, where FY24 "Hybrid IT" appears to have bundled cloud+government+data-centre differently). Segment taxonomy shifted between the FY24 and FY25 presentations — treat cross-year segment comparisons with care; this is a real provenance caveat, not a rounding issue.

By geography: effectively 100% Australia. No meaningful international revenue. This is a single-country play — a bull point (sovereign moat) and a bear point (no diversification, one regulator, one grid, one currency).

The segment story in one line: Data Centres is the small-but-accelerating future, Government is the stable annuity, Telecom is the shrinking-but-cash-generative past that funds the build.

---

Phase B — Measure performance

Earnings Result (latest print — FY25, reported 27 Aug 2025)

The print, all ``:

• Revenue $369.6m, +1.7% YoY — modest; the group is capacity-gated on its best segment and repricing-pressured on its telco.
• EBITDA $113.6m, +4.1%; margin 30.7% (+90bps) — margin expansion despite flat-ish revenue = mix shift toward higher-margin data-centre/cloud and cost discipline. Eleventh straight year of EBITDA growth.
• NPAT $34.9m, +5.7%; EPS 135.2c.
• Segment drivers: Data Centres +5.5% EBITDA (the accelerator, but capacity-capped pre-IC3); Cloud/Government +1.7%; Telecom roughly flat but about to decline.
• Guidance / outlook: Telecom EBITDA guided DOWN to ~$20m in FY26 (from $24m) on NBN price pass-through — a rare explicit negative guide and the key near-term earnings drag. Group-level FY26 numeric guidance n/a. IC3 Super West phase-1 (initial 6MW IT load) commissioning September 2026 — first incremental data-centre revenue lands in FY27, not FY26.
• Balance sheet (30 June 2025): cash $16.1m + deposits $75.0m (~$91m liquidity), undrawn debt facility $450m, plus the A$200m NRFC hybrid (first $100m drawable ~Mar 2026, second tranche by Mar 2027). Effectively net cash / very lightly geared before the growth draw — a fortress position going into a heavy capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs. cycle. Explicit net-debt figure n/a.
• Cash flow: operating cash flow $109.9m (after $25.5m of tax covering FY24+FY25). Strong cash conversion. Capex $150.1m — Growth $111.0m (incl. $106.0m on IC3 Super West), Customer-growth $18.9m, Maintenance $20.2m. MAQ is now FCF-negative at the group level because of the build ($109.9m opcash − $150.1m capex ≈ −$40m ) — this is deliberate growth spend, funded by the facility + NRFC, not distress.
• Dividend: 0.00% yield — MAQ pays no dividend currently; capital is redirected entirely to the data-centre build.
• Market reaction / what's priced in: despite the AI-infra narrative, MAQ shares are down ~31% over the trailing 12 months and underperformed the ASX All Ords by ~35%. The market is not paying up for the IC3 story yet — it is discounting execution risk, the FY26 telecom drag, zero near-term DC revenue, and dilutionDilutionIssuing new shares, so each existing share owns a smaller slice of the same company. memory (see Lens 8).

Unusual vs own history: first time the group is meaningfully FCF-negative and first explicit negative segment guide (Telecom) — both are consequences of the strategic pivot, not operational deterioration. The tape disagrees with the fundamentals' quality: an 11-year EBITDA-grower down 31% is the whole setup.

Earnings Calls (sentiment trend)

No transcripts on the shelf (transcripts=0) and MAQ (small-cap ASX) transcripts are not cleanly scrapeable, so this is directional from results-announcement language ``:

• Consistent management refrain across FY24→FY25→1H26: "consecutive halves/years of EBITDA growth" (now 22 halves) — management anchors the story on unbroken profitable-growth track record, a deliberate contrast to loss-making growth-stage data-centre peers.
• Rising emphasis (FY25 into 2026): AI-ready capacity, liquid cooling, sovereignty, hyperscale + SaaS demand — the vocabulary has shifted decisively from "telco/managed services" to "AI infrastructure." The IC3 build and the NRFC sovereign-investment win dominate recent commentary.
• New candour: explicit acknowledgement of the NBN/Telecom margin headwind — management is pre-warning the market rather than hiding it, which reads as credible.
• What they've stopped saying: the "telco" identity — the May-2023 rename and the messaging both bury Telecom as legacy.

Tone trajectory: confident and increasingly AI-forward, but grounded/under-promising on the near term (flagging the Telecom dip, no hype guidance on IC3 lease-up). Consistent with founder-operators managing a multi-decade compounder rather than a momentum story. Sentiment: steady-positive, low-promotional. Caveat: this is inferred from written releases, not a heard-tone analysis of Q&A — a real limitation without transcripts.

Comps

Peer table — MAQ vs Australian data-centre / digital-infra peers. Multiples are `` with source/date, or n/a. No multiple is fabricated.

Read: the comp table is the thesis. MAQ trades at ~16x EV/EBITDA on blended group EBITDA that is still ~75% telco+cloud; NextDC — a pure-play, loss-making, hyper-growth data-centre developer with 667MW contracted vs MAQ's ~34MW live — commands ~61x. The market values a marginal MW of contracted AI capacity at NextDC at a vast premium to a marginal MW inside MAQ. The entire bull case is that IC3's 47MW, once leased, should be valued closer to the pure-play multiple than the telco multiple — a re-rating that MAQ's group structure currently suppresses. The P/E of ~50x already embeds some of this optionality; EV/EBITDA of 16x does not. The gap between those two multiples (rich P/E, cheap EV/EBITDA) is the single most important number in this dossier — it says the equity is priced for growth while the enterprise is priced for a utility.

Stock-Price Catalysts (>5% moves, last ~5 years)

Directional (no clean intraday move-attribution dataset on the shelf); all ``:

• 2021 (up): July-2021 announcement of IC3 Super West kicked off the data-centre re-rating; MAQ rode the global 2021 data-centre/AI-infra enthusiasm to multi-year highs.
• May 2023 (identity catalyst): rename to Macquarie Technology Group — signalling the DC pivot; re-anchored the equity story on data centres.
• Capital raisings (dilution catalysts): completed a ~A$160m follow-on equity offering to help fund the build — equity raises are historically the sharp down catalysts for this name (the market marks the dilution).
• March 2026 (up catalyst): A$200m NRFC sovereign hybrid — a first-of-its-kind, non-dilutive, government-sovereign endorsement of the sovereign-cloud thesis. Validated both the funding path and the strategic moat.
• 2025→2026 (down): the ~31% trailing-12-month decline despite these positives — driven by higher-for-longer rates (a capex-heavy, no-dividend infra builder is rate-sensitive), the FY26 Telecom guide-down, and impatience with the multi-year gap to IC3 revenue.
• Sept 2026 (upcoming catalyst): IC3 phase-1 (6MW) commissioning — the first hard proof the asset works and can be leased.

Pattern the market actually reacts to: (1) data-centre capacity/leasing news (up), (2) equity dilution (down), (3) rates/macro (this is a duration asset), and (4) sovereign/government validation (up). Notably it reacts less to the steady EBITDA beats — the market has decided this is a data-centre optionality story, and prices the option, not the annuity.

---

Phase C — Judge people & books

Management

CEO David Tudehope; Head of Data Centres/Hybrid IT (Executive Director) Aidan Tudehope — co-founders, brothers, in the seats since 1992.

1. Track record: built Macquarie Telecom from two founders' savings in 1992 into a ~A$1.6–1.8B-market-cap enterprise-IT and data-centre group over 30+ years. 22 consecutive half-years and 11 consecutive years of EBITDA growth — an exceptional operating record on the ASX. Successfully engineered the strategic pivot from commodity telco to sovereign data centres, and landed a first-of-its-kind sovereign government hybrid (NRFC).
2. Tenure & skin in the game: ~34 years. Founders hold ~41% of the company (insider ownership including other insiders ~44–53% of the register). This is one of the most aligned management teams on the ASX — a genuine owner-operator setup. David Tudehope is described as the most bullish insider.
3. Capital allocation: reinvestment over distribution — no dividend; cash and raised capital funnelled into owned land + the IC3 build. Historically disciplined (long profitable-growth streak implies no value-destroying M&A). The judgment call now is whether ploughing FCF + A$200m hybrid + up to A$450m facility into a single 47MW asset is the right bet — the quality of this allocation won't be provable until IC3 leases.
4. Red flags: low. NPS +71 and long tenure argue against promotional behaviour. No related-party or comp scandals surfaced. The main governance watch-item is simply founder control — 41% ownership + a tiny float means minority holders have limited say and low liquidity; the founders' priorities are the company's.
5. Archetype: founder-operators, builder-type, long-horizon compounders — exactly the archetype you want running a multi-year, capital-intensive infrastructure build, provided you trust their capital-allocation judgment. For this stage (betting the balance sheet on AI data centres), founder conviction + aligned ownership is a feature; the risk is concentration and the absence of an external check.

Forensic Red Flags

Act as a forensic analyst — findings, all `` (no filings on shelf; FY25 statements are image-scans, unparsed):

• Revenue recognition: colocation and managed-services revenue is recurring and low-risk to recognise; no aggressive-rev-rec signals surfaced. Cannot verify contract-liability/deferred-revenue movements without the audited statements — flagged as an open item, not a red flag.
• Cash flow vs earnings: operating cash flow ($109.9m) comfortably exceeds NPAT ($34.9m) — the gap is heavy D&A on the infrastructure base, which is normal and healthy for a data-centre owner. No earnings-quality red flag; the reverse — earnings are conservatively stated relative to cash.
• Capitalisation risk (the one to watch): with $150m capex and $106m going into IC3, the accounting question is how much cost is capitalised vs expensed, and the depreciation schedule once IC3 is commissioned. A large new asset base will depress reported NPAT via D&A even as EBITDA/cash rise — watch for a widening EBITDA-to-NPAT gap in FY27+ that is mechanical (depreciation), not deterioration. Not a red flag, but the number most likely to confuse a casual reader.
• Balance-sheet structure: the A$200m NRFC "hybrid" — perpetual, callable, subordinated, unsecured, non-convertible — sits in a grey zone between debt and equity. How it's classified (equity vs liability) and its coupon materially affect leverage optics and NPAT (if coupon runs through P&L). This is the single most important accounting item to scrutinise in the FY26 accounts.
• Receivables/inventory vs revenue: no data on the shelf; government/enterprise receivables are typically clean. Open item.
• SBC / non-GAAP: MAQ reports statutory EBITDA/NPAT; no evidence of heavy SBC add-back gaming. Founder ownership reduces reliance on option comp.
• Goodwill/intangibles: not sourced; low risk given organic (not acquisitive) growth history.

Regulatory findings (required sub-section):

• SEC (EDGAR LR + AAER): regulatory/regulatory-findings.md returns 0 findings — MAQ has no CIK and does not file with the SEC (Australian issuer). No EDGAR search possible. ``
• Australian regulators (web search): One historical action — in 2016 the ACMA sanctioned Macquarie Telecom over Integrated Public Number Database (IPND) breaches (inaccurate records for ~142,499 services, 2010–2015). Resolved via an enforceable undertaking — NO monetary fine — committing to data-process upgrades and an independent audit. No recurrence found; no ACMA/ACCC action against the company in 2024–2026.
• ⚠️ Misattribution guard: the December-2025 ASIC A$35m short-sale-misreporting fine belongs to Macquarie Securities (Macquarie Group / MQG, the investment bank) — a different company. It is NOT Macquarie Technology Group. Do not attribute it to MAQ.AX.
• 10-K Item 3 (Legal Proceedings): n/a — no SEC filings; Australian annual-report legal-proceedings note not parseable from the image-scan. Open item.
• Conclusion: No material current regulatory or legal findings. One dormant 2016 telecom-data undertaking (no fine, resolved). Verified via SEC EDGAR EFTS (0, no CIK), targeted web search of Australian regulators (ACMA/ACCC/ASIC), as of 2026-07-06. Clean for the issuer itself.

---

Phase D — Project & stress-test

Forward Projection (FY26 / FY27 / FY28 EPS)

Built bottom-up from FY25 actuals + disclosed guidance. Output ``; every input labelled. AUD. Fiscal years end 30 June. No our model logged (watchlist/unattended mode — SKILL says skip the create step in the loop).

Base-case build:

• FY25 actuals: revenue $369.6m, EBITDA $113.6m, NPAT $34.9m, EPS 135.2c, ~25.8m shares.
• FY26: Telecom EBITDA −$4m (guided $24m→$20m); Data Centres +$2m (still capacity-capped pre-IC3, phase-1 only commissions Sept-2026 so ½-year, minimal FY26 contribution); Cloud/Gov +$0.5m. Net EBITDA roughly flat-to-slightly-down ~$112–114m. NPAT pressured further by rising D&A + NRFC hybrid coupon once drawn ($100m at, say, an ~8% coupon ≈ $8m annual cost ) → FY26 NPAT ~$28–32m, EPS ~110–125c.
• FY27: first meaningful IC3 revenue as phase-1 (6MW) leases and further MW commission; Data Centres EBITDA steps up (at ~$1.8m EBITDA/MW, each incremental 6MW ≈ +$11m EBITDA at full lease ). Telecom stabilises at the new ~$20m base. EBITDA ~$125–140m; NPAT ~$32–40m, EPS ~125–155c — but heavily dependent on IC3 lease-up pace.
• FY28: IC3 ramps toward stabilisation; multiple MW tranches contracted. EBITDA ~$150–175m; EPS ~150–200c if leasing tracks; the operating leverage on a paid-for, secured-power asset is large.

Three paths (EPS, AUD cents):

The forecast that matters isn't EPS — it's IC3 lease-up. MAQ's near-term GAAP EPS will be suppressed by depreciation and the hybrid coupon even as the enterprise valueEnterprise valueWhat it would cost to buy the whole company: its market value plus its debt, minus the cash you would get with it. Often very different from market cap. builds. This is why the market uses EV/EBITDA and MW-contracted, not P/E, for this name. Base-case scoreable claim (not logged): MAQ FY27 group EBITDA ≥ A$125m, p≈0.55.

Bull vs Bear

Bull case. MAQ owns something genuinely scarce: a certified-strategic, AI-ready, 63MW-end-state Sydney data-centre campus with power secured, plus a put-and-call on a ~200MW next campus, funded by a fortress balance sheet (net cash + A$450m facility + A$200m non-dilutive sovereign hybrid) and run by 41%-owner founders with an 11-year unbroken EBITDA-growth record and the highest NPS on the ASX. Australia is legislating toward data sovereignty exactly as AI drives onshore-hosting demand — a secular tailwind straight into MAQ's compliance moat. The re-rate lever is explicit: MAQ trades at ~16x EV/EBITDA while pure-play NextDC trades at ~61x; as IC3's 47MW leases to hyperscalers/SaaS-AI customers, the data-centre segment's earnings and implied value should migrate toward the pure-play multiple, and a sum-of-the-parts that values the DC arm at even half NextDC's multiple implies material upside from A$63.82. Analyst consensus is Buy, avg target ~A$84.78 (5 buys / 2 holds, 7 analysts) — ~33% above spot.

Bear case (permanent-impairment risks). (1) IC3 lease-up disappoints — MAQ is a 34MW minnow entering a market where NextDC (667MW contracted), AirTrunk (Blackstone-owned), and Global Switch fight for the same hyperscalers; if the big AI tenants pick scale players, IC3's 47MW leases slowly and the whole re-rate thesis stalls with a depreciating asset dragging NPAT for years. (2) Sub-scale disadvantage is structural — MAQ cannot match peers' capex ($2.2B NextDC raises) or hyperscale relationships; its niche (sovereign/government) may be too small to justify a pure-play multiple, so the SOTP re-rate may simply never arrive. (3) Telecom secular decline + NBN repricing permanently resets a quarter of group EBITDA lower. Pre-mortem (18 months out, thesis broken): IC3 commissioned on time but leasing crawled (hyperscalers went to NextDC/AirTrunk), the NRFC coupon and depreciation crushed reported EPS, rates stayed high punishing the duration asset, and the stock de-rated further toward a telco-plus-empty-DC multiple — the market concluded MAQ built a beautiful data centre nobody rushed to fill. Are multiples too high? P/E ~50x is demanding for a business whose EPS is about to be depreciation-suppressed; EV/EBITDA ~16x is not demanding if you believe the DC re-rate. The whole debate is which multiple is "right."

Contrarian view — what the market is refusing to see: the market is treating MAQ as a sub-scale also-ran to NextDC and pricing it near a telco multiple. What it under-weights is that sovereignty is a moat NextDC's hyperscale model doesn't fully own — 42% of federal agencies and a Certified-Strategic-only portfolio is a defensible, high-margin annuity that becomes more valuable as AI-sovereignty rules tighten. MAQ may never be NextDC's size, but it doesn't need to be to justify a re-rate off 16x — it needs IC3 to lease and the sovereign premium to be recognised. The −31% year has arguably created the setup.

Devil's Advocate (short-seller)

You are a skeptical short-seller. The case against MAQ.AX:

• The whole thesis rests on one unbuilt-out asset leasing to customers who have better options. IC3's 47MW is a rounding error against NextDC's 667MW contracted. The hyperscalers driving AI demand sign with scale players who can deliver hundreds of MW across regions. MAQ is pitching "sovereign + SaaS-AI onshore" — a real but narrow slice. If IC3 lease-up is slow, you own a capital-intensive telco with a half-empty new building depreciating against earnings for a decade.
• Revenue concentration risk cuts both ways: the crown-jewel DC segment is ~$74m of a $370m group — you're paying a growth multiple for <20% of the business while 65%+ is low-growth telco + steady-state cloud. And Telecom (guided down) is structurally melting.
• The moat may be weaker than bulls think: "Certified Strategic" is replicable — AWS/Azure/Google all run sovereign-cloud programs for Australian government, and NextDC also holds government certifications. MAQ's 42%-of-agencies today is not contractually permanent.
• Most dangerous competitor bulls underestimate: NextDC — vastly better capitalised (A$8.4B liquidity, A$2.2B raises), more MW, deeper hyperscale relationships. In a land-grab for AI capacity, capital and scale win; MAQ risks being out-built in its home market.
• Capital allocation is a single concentrated bet: no dividend, FCF-negative, drawing a A$450m facility + a A$200m perpetual hybrid — all to fund one 47MW asset. If it doesn't lease well, that's a balance-sheet stretched for a miss, and the perpetual hybrid's coupon is a permanent EPS tax.
• Valuation if growth disappoints 20–30%: strip the DC re-rate optionality and value MAQ as a slow-growth Australian IT services group on ~10–12x EV/EBITDA → materially below A$63.82. The P/E of ~50x has nothing to fall back on if IC3 underwhelms.
• Single scenario that permanently impairs: hyperscalers standardise on NextDC/AirTrunk for Australian AI capacity, MAQ's IC3 leases to <50% over three years, and sovereign-cloud commoditises as the big-3 clouds win government mandates — MAQ becomes a stranded-ish 63MW campus attached to a declining telco. Plausibility: moderate — not a fraud or blow-up risk, but a very real "good asset, wrong scale, slow fill" de-rating. Liquidity trap: 41% founder-owned, ~25.8m shares — thin float means a re-rate down is hard to exit.

Management Questions (ordered by information value)

1. IC3 Super West — what is the current signed/contracted MW and the committed pipeline, and what leasing pace do you expect for the 47MW over FY27–FY28? (The single answer that most changes the thesis.)
2. Who are the target IC3 customers — what mix of hyperscale cloud, SaaS-with-AI, and government — and how many are in advanced/contracted negotiations today?
3. How should investors think about the sum-of-the-parts value of the Data Centres segment vs the group's ~16x blended EV/EBITDA, given NextDC trades at ~61x?
4. How is the A$200m NRFC hybrid classified (equity vs liability), what is its coupon/step-up, and what annual P&L cost should we model once fully drawn?
5. What is the expected depreciation trajectory and EBITDA-to-NPAT gap in FY27–FY28 once IC3 commissions, and how do you want the market to value the business through that period?
6. On the next ~200MW campus (the put-and-call land) — timing, estimated capex, power-connection status, and how you'll fund it without dilution.
7. How defensible is the Certified Strategic / sovereign advantage as AWS/Azure/Google expand their own government-sovereign offerings — what keeps 42% of federal agencies with MAQ?
8. What is the FY26–FY28 group EBITDA and capex outlook, and when does the group return to positive free cash flow?
9. Quantify the NBN/Telecom headwind beyond FY26 — where does Telecom EBITDA stabilise, and is there a floor?
10. What are the power and grid-connection risks for the 63MW end-state and beyond — is all end-state power for the current campus firmly secured and priced?
11. Given the ~31% share-price fall and the balance-sheet capacity, would you consider a buyback or is every dollar committed to the build?
12. Return targets: what unlevered yield-on-cost / ROIC do you underwrite for IC3, and what lease-up assumption underpins it?
13. Customer concentration in Data Centres and Government — what is the single-largest-customer share, and how contracted/recurring is it?
14. How does liquid-cooling / AI-density design at IC3 differentiate you technically from NextDC and AirTrunk for the same tenants?
15. Succession and founder control — with ~41% held by the founders after 34 years, what is the long-term ownership and leadership plan?

---