Australian AI data centre operator NextDC shares shed 44% on revenue delay

NextDC shares fell 44%, erasing about A$3.1 billion in market value, after it cut its 2027 net revenue forecast to A$475 million–A$500 million from A$615 million–A$640 million and EBITDA forecast to A$245 million–A$265 million from A$385 million–A$410 million, excluding affected M3 contracts. The company said A$140 million of contracted revenue would be deferred because of power-supply delays at its Melbourne M3 data centre, shelved new developments and land acquisitions for the foreseeable future, and cut 2027 capex guidance to A$4.5 billion–A$5 billion. The M3 power-supply increase is now expected in 2028; the article also describes weakening investor sentiment toward AI-linked infrastructure.

Analysis

The key read-through is not simply weaker AI demand: it is that contracted data-centre capacity can fail to convert into timely revenue when grid upgrades and local approvals lag. That shifts the underwriting bottleneck from customer demand toward power access and delivery schedules. The delay appears to defer, rather than cancel, M3 revenue, but it still raises the chance of carrying development costs and financing exposure for longer than planned.

For operators, the second-order risk is lower confidence in forward capacity commitments: customers may seek later start dates or more flexible terms, while lenders and equity investors demand evidence of secured power before funding new builds. NextDC’s capex reduction may ease near-term cash needs, but the remaining programme is still substantial; capital recycling could help liquidity while also signaling that some planned sites are less valuable without credible power timelines. Hyperscalers and large tenants may gain negotiating leverage. Operators with energized capacity and diversified power access could take share, though this event does not establish which competitors have spare capacity.

Near term, the 44% fall makes an immediate short unattractive absent evidence of further estimate cuts. Over 1–3 months, watch revised delivery schedules, customer deferrals/cancellations, funding terms and the spending review. Over 6–18 months, community opposition and grid-connection lead times could constrain industry-wide supply and raise the value of already-powered sites. Contrarian angle: this is a strong warning on execution risk, not proof that AI infrastructure demand has collapsed. The thesis weakens if M3 power works restart on schedule and deferred contracts convert without material customer loss; it strengthens if delays spread or contracted capacity is repriced.

Key Decisions for Investors

• Do not buy the post-selloff solely on the assumption that deferred M3 revenue will arrive unchanged. Wait for confirmation of revised power milestones and whether customers retain their contracted capacity.
• Avoid chasing a fresh short in NextDC immediately after the sharp repricing. Reassess after the detailed capex review and updated cash-flow/funding disclosures; a further negative catalyst would be evidence of contract cancellations or additional project delays.
• For relative-value work, compare NextDC with data-centre operators that can demonstrate energized capacity and less dependence on a single delayed upgrade. Use named peers such as Goodman Group, Equinix and Digital Realty as diligence comparables, not presumed beneficiaries; verify available capacity, power access and valuation before positioning.
• Track Jemena’s restart timing, planning approvals, customer conversion, capex commitments and financing conditions. A restart before the stated timetable or unchanged customer commitments would challenge the bearish execution thesis; further slippage or weaker funding access would support it.