NEXTDC (ASX: NXT) — Investment Research Report

3-Year / 5-Year Price Scenarios

Base EPS: $0.2 — Analyst normalized EPS proxy, not reported statutory EPS: FY25 and 1H26 statutory results remained loss-making because depreciation and financing costs are heavy during the build-out. Targets use a normalized earnings conversion of the contracted platform and are highly uncertain.

0. Information Richness & AI Limitations

Grade A for disclosure, not for certainty. NEXTDC's investor centre lists the FY25 annual report and the newer 1H26 results announcement and interim report. The primary filings are detailed, while the live market quote is independently available through Yahoo Finance and the ASX company header. Investment certainty is materially lower than information richness: the company is in a capital-intensive expansion phase, statutory earnings are negative, and long-range returns depend on power, funding, construction and customer decisions. The analysis is a dated educational snapshot as of 2026-07-31, not personalized financial advice.

1. Data & Cross-Validation

Yahoo Finance reported NXT.AX at A$13.505 on 2026-07-31; the ASX company header independently reported A$13.505 and market capitalisation of A$9.726B. Dividing that market cap by price implies approximately 720.2M shares, and A$13.505 × 720.2M = A$9.726B, a close reconciliation. The Yahoo chart showed a 52-week range of A$10.91–18.22.

Primary financial facts (AUD): the official NEXTDC investor-centre index lists FY25 and 1H26 as published. The FY25 results announcement for the year ended 30 June 2025 reported total revenue of A$427.209M, attributable loss of A$60.543M and basic loss per share of 9.59 cents. The same filing's operating highlights and the official investor centre reported FY25 net revenue of A$350.2M and underlying EBITDA of A$216.7M. The 1H26 results announcement for the half-year ended 31 December 2025 reported total revenue of approximately A$205.5M, net revenue of A$189.2M, underlying EBITDA of A$115.3M and net loss after tax of A$39.4M. Neither period declared a dividend.

Sources: NEXTDC financial reports index, FY25 Results Announcement, FY25 Annual Report, 1H26 Results Announcement, and 1H26 Reports and Accounts.

2. Business Essence — Duan Yongping

One line: NEXTDC develops and operates carrier-neutral data centres, selling secure powered capacity, cooling, connectivity and interconnection to cloud, enterprise, government and digital-platform customers.

The economics are recurring once a facility is built: customers colocate critical infrastructure, connect to multiple clouds and networks, and face meaningful operational risk in switching. Pricing power is local rather than universal because power availability, land, permits and network ecosystems constrain supply. Operating leverage is high: fixed facilities and energy infrastructure support incremental revenue, but the model first requires very large construction and fit-out investment. FY25's record A$1.7B capital expenditure and 1H26 expansion programme show both the opportunity and the cash-flow burden.

3. Moat — Buffett

The moat is built from scarce powered sites, permits, network connectivity, security certifications, operating reputation and customer ecosystem density. Interconnection makes a facility more useful as more carriers, clouds and software platforms are present; this creates a local network effect. High switching costs arise from migration risk and downtime sensitivity, although hyperscalers retain negotiating power and can build or lease elsewhere.

The moat is potentially widening as AI demand makes power and ready-to-connect capacity scarce. It can narrow if competitors secure power faster, customers insource, or distributed architectures reduce data-centre intensity. NEXTDC's 1H26 contracted utilisation of 416.6MW and forward order book of 296.8MW are evidence of demand, not proof that every planned project will earn attractive returns.

4. Reverse Thinking & Risks — Munger

Inversion: the stock can lose money even while revenue grows if capex, depreciation, interest and dilution grow faster than per-share cash earnings. The most important disconfirming evidence would be falling contracted utilisation, weaker order-book conversion, project delays, or positive EBITDA failing to become positive free cash flow. The historical analogy is a utility-like infrastructure build-out with technology-cycle valuation: scarce assets can be excellent businesses, but overbuilding and leverage turn growth into poor shareholder returns.

5. Management — Duan Yongping + Buffett

Management has established a credible Australian data-centre platform, expanded into Auckland, Kuala Lumpur and Melbourne, and continued to sign capacity contracts. The FY25 record contracting and the 1H26 senior debt upsizing to A$2.3B indicate access to capital and customer confidence. The test is capital allocation: management must preserve returns while funding Sydney, Melbourne, Kuala Lumpur and other projects.

Incentives are aligned through ordinary equity, but growth narratives can encourage overbuilding. A management-quality upgrade requires repeated on-time commissioning, contracted utilisation ahead of spend, disciplined leverage and less reliance on equity issuance. The business should remain competitive after a CEO change because the core assets, sites and customer ecosystem matter more than a single personality.

6. Industry & Civilizational Trend — Li Lu

Data centres are the physical layer of cloud computing, AI inference and training, digital government and enterprise software. NEXTDC sits between utilities and digital platforms: it monetises power, real estate, cooling and connectivity while customers supply the compute hardware. The total addressable market is constrained by electricity, transmission, water, land, permits and skilled construction, which makes local supply scarcity more important than global server demand alone.

The long-run trend is favourable but not linear. AI can increase rack density and power demand, while new chip architectures, liquid cooling and distributed compute can change facility requirements. The best civilisational position is not simply owning megawatts; it is owning permitted, connected, reliable capacity in markets where customers need low-latency and sovereign infrastructure.

7. Valuation & Scenarios — Buffett + Duan

Statutory P/E is not meaningful because FY25 and 1H26 reported losses. I therefore use a deliberately explicit normalized EPS proxy of A$0.20, representing an analyst judgment about eventual earnings conversion rather than a reported number. This is the largest uncertainty in the model. The scenarios apply 20%/10%/−5% normalized EPS growth and 45x/35x/25x multiples over three years, producing targets of A$15.55, A$9.32 and A$2.14; probabilities of 30%/50%/20% give a weighted target of A$9.75, or −27.8% from A$13.505.

Over five years, the bull/base/bear cases use the same growth assumptions and 45x/35x/20x multiples, producing A$22.39, A$11.27 and A$3.10. Probabilities give a weighted target of A$12.97, or −4.0% before any dividend (none was declared in the cited periods). The bull case is plausible only if AI demand, capacity delivery and pricing compound while funding remains available. The base case assumes strong growth but multiple compression; the bear case captures stalled demand, cost overruns and dilution. The current price therefore offers no obvious margin of safety despite a strong industry runway.

8. Decision Memo

Decision: Watch / no position. A patient buyer should prefer A$9.50–11.50, or wait for evidence that new capacity is being commissioned on time and funded without damaging per-share economics. Existing holders should monitor order-book conversion, contracted utilisation, net debt, project returns and free cash flow. The thesis is invalidated by persistent utilisation deterioration, material customer concentration, repeated equity funding or returns below the cost of capital. This is an analytical decision, not personalized financial advice.

Four-Master Commentary

The following are analytical simulations, not real quotations.

Buffett: A data centre can be a wonderful recurring business when the site, power and customers are hard to replicate. But I would not pay for five years of perfect execution before the company has converted its build-out into durable per-share cash earnings.

Munger: Invert it: growth can still destroy value if management overbuilds, debt compounds and customers capture the economics. The question is not whether AI is important; it is whether NXT earns a return after all the capital required.

Duan Yongping: The customer relationship and interconnection ecosystem are valuable, but a good business is not automatically a good buy. I want to see disciplined capital allocation and a price that leaves room for mistakes.

Li Lu: Digital civilisation needs physical compute infrastructure, and power scarcity may make the best locations more valuable. The long-term trend is strong, but technology and capital cycles mean the entry price still determines the investment outcome.