Everything About AI
NextFin News - NextDC shares rebounded as much as 8.2% in Sydney on Thursday, a day after a community blockade over power poles in suburban Melbourne triggered the biggest single-session drop since the data centre operator listed in 2010. The swing - a plunge of as much as 44% on Wednesday followed by an intraday recovery of more than 8% - captures the central tension of the moment: investors are being forced to price a hard physical constraint into a sector that has been trading on the assumption that power, land and capital would always arrive on time.
The trigger was a guidance cut of unusual scale. NextDC slashed its fiscal 2027 net revenue outlook to A$475 million to A$500 million from A$615 million to A$640 million and its underlying EBITDA forecast to A$245 million to A$265 million from A$385 million to A$410 million, removing roughly A$140 million of revenue and earnings tied to contracts at its M3 Melbourne site. The company said electricity distributor Jemena has suspended parts of the 66kV line augmentation works needed to feed the expanding facility, does not expect to resume before February 2027, and now sees the next grid-supply increase slipping into fiscal 2028.
By Thursday, the dust had settled enough for buyers to step back in. But the rebound does not resolve the deeper question the week has raised: is this a timing problem that AI demand will eventually absorb, or the first clear proof that the data centre buildout faces a structural bottleneck no amount of contracted demand can fix?
The Plunge and the Rebound
The sequence began with a two-day trading halt starting 19 October, lifted Wednesday morning alongside the ASX announcement titled "M3 Melbourne Power Delay and Revised FY27 Guidance." The market's verdict was immediate and severe. Shares fell as much as 44% in early trade, wiping about A$3.1 billion from the company's market value and making it the worst performer on the S&P/ASX 200; by mid-afternoon the stock was down about 39% at A$5.62 against a previous close of A$9.18, its lowest close since 18 January, leaving a market capitalisation of about A$4.3 billion.
Thursday brought the counter-move. The stock climbed as much as 8.2% in Sydney, trading between A$5.26 and A$5.79 against a previous close of A$5.35, with volume of about 9.9 million shares running at more than two-and-a-half times its average. The rebound was helped by a broader stabilisation in global AI-linked assets after a week that had punished the sector from Seoul to Singapore.
Even so, the shares are down roughly 40% over the past month, and the stock had already fallen for seven straight weeks before the warning. Short sellers, who had made NextDC the 28th most-shorted stock on the ASX with 8.52% short interest as of 14 October, booked an estimated A$250 million in paper profits on the Wednesday rout, market commentary estimated.
What Actually Happened at M3
The mechanics of the delay are unusually mundane for a company at the centre of the AI infrastructure boom. Jemena, the electricity distributor that supplies M3, suspended parts of its 66kV line augmentation works in September 2026 after residents in Melbourne's inner-west suburbs of West Footscray and Yarraville blockaded crews - at times leaving cars and bins across work sites - in protest against the installation of 15-metre-tall concrete power poles carrying 66,000 volts along residential streets. A typical residential pole carries about 400 volts; some of the new structures rise as high as 22 metres, or roughly seven storeys, the local council said.
Jemena has cited safety concerns, restricted site access due to community opposition, and the absence of consent from the local road authority while further consultation takes place. It has told NextDC it does not expect to recommence the works before February 2027. The knock-on effect is political as well as physical: the Victorian Government's planning decision on M3 Stages 6 and 7 is now expected only after the state election on 28 November 2026.
The financial consequence is direct. M3's grid supply increases in stages, and the next increase is what allows NextDC to deliver and bill contracted capacity. With that increase deferred into fiscal 2028, the company has removed the affected M3 contracts from its FY27 guidance entirely. Those contracts contributed approximately A$140 million each of net revenue and underlying EBITDA to the August outlook. Management stressed that the capacity remains under contract, subject to customer rights, and that data hall construction is continuing while it works with the customer on a revised power ramp.
"Jemena has suspended parts of the network works required for the next increase in grid supply to M3 Melbourne," chief executive Craig Scroggie said in the filing. "Jemena has attributed the suspension to community and stakeholder opposition to the works and to the absence of local council consent pending further consultation. On Jemena's current indicative estimate, we expect that increase during FY28."
The company has responded defensively on capital. It cut FY27 capital expenditure guidance to A$4.5 billion to A$5.0 billion from A$5.25 billion to A$5.75 billion, paused all projects not backed by customer contracts, eliminated material discretionary spending, and said it would evaluate capital recycling including potential sales of selected development sites. A detailed review of the capital programme will be announced at the annual general meeting on 20 November 2026.
On liquidity, management moved to head off a funding scare. NextDC reported A$8.675 billion of liquidity at 30 September 2026, said it remains in compliance with financial covenants, and noted that its earliest scheduled debt maturity is December 2029, with a potential convertible-notes holder put option on 17 September 2029. The company also warned the M3 delays could affect its ability to draw on A$6.7 billion of undrawn senior debt.
Why the Market Reacted as If This Were Structural
A 42% single-day wipeout on a guidance cut is not simply a re-rating of next year's earnings. It is the market repricing the certainty of the growth model. Before the downgrade, Goldman Sachs estimated NextDC traded at about 26 times forward FY27 enterprise value to EBITDA, above the 22 times average for global pure-play data centre peers that includes Equinix, Digital Realty and Keppel DC REIT. After the cut, the implied multiple balloons to roughly 37 times even after a 40% fall - meaning the shares could drop 40% and still only converge toward peer valuation.
The deeper issue is the earnings trajectory that justified that premium. Before the downgrade, Goldman had NextDC delivering the strongest FY26-28 EBITDA compound growth rate among its pure-play peers, at 74% versus a peer average of 23%. The entire investment case rested on converting a large forward order book into billable capacity on an aggressive timetable. A one-year deferral of A$140 million of earnings does more than cut FY27: it pushes the compounding machine a year down the track and forces investors to ask whether the next delay is already in the pipeline.
There is also a financing dimension that the equity market is now scrutinising. Data centre developers are capital-intensive by nature, and NextDC is among the most levered of the group. UBS, in August forecasts, expected the company to remain loss-making until at least FY31 despite EBITDA growing more than fourfold, with depreciation and amortisation of A$500 million to A$575 million a year, losses peaking at A$225 million in FY28, and long-term debt more than doubling in FY27 from A$3.28 billion to A$7.93 billion before peaking at A$11.17 billion in FY30. The warning that M3 delays could impair access to A$6.7 billion of undrawn senior debt turns a physical delay into a potential funding constraint - and funding constraints, unlike construction delays, compound.
Citi added in a note that the power delays could affect earnings beyond fiscal 2027, citing delays to approvals for the M3 expansion and a longer-than-expected timeline for transformer upgrades. Tom Beadle, an analyst at RBC Capital Markets, put the broader risk plainly: "Whilst out of the company's control, delays to power supply and development applications relating to NXT's M3 data centre is evidence that community concerns around data centre developments present a real risk to the timing of delivery of contracted capacity for data centre owners."
The Cyclical Read Versus the Structural Read
This is the judgment the rebound forces investors to make, and the two readings point in opposite directions.
The cyclical case is straightforward. The A$140 million of affected contracts are still contracts. The capacity is still under construction. The demand side of the equation - hyperscaler and AI compute orders - has not changed. If Jemena resumes work in February 2027 and the grid connection lands in FY28 as now guided, the deferred earnings are recovered rather than lost, and the whole episode becomes a timing variance that FY28-29 absorbs. Under this reading, the 42% drop was a panic overreaching a one-year deferral, and Thursday's rebound is the first step toward normalisation.
The structural case is harder to dismiss, and it is what the sector rout this week has been about. The constraint is not demand, capital or technology. It is social licence - the consent of the communities in whose backyards the infrastructure must sit - and that is not something a balance sheet can buy quickly. The M3 protest is not an isolated incident. Across Australia, Green Square was forced to pull a proposed Western Australian project, Goodman Group shelved a A$1.2 billion facility in Sydney's Lane Cove, and private developer Firmus abandoned a planned A$44 billion stock market float amid the same backlash. Globally, the pattern is visible in the past week alone: Blackstone-backed AirTrunk postponed a Singapore REIT IPO on 13 October, CoreWeave's bonds fell below 80 cents on the dollar, South Korea's Elice Group plunged 32% on its debut before recovering, and Singapore's DayOne Data Centers is weighing whether to delay or shrink a US IPO.
The structural argument is not that AI demand will fade. It is that the transmission mechanism from contracted demand to billable revenue has a new choke point, and that choke point sits outside the operator's control. A hyperscaler can sign a 10-year contract on Monday; a local council consultation process can stretch the power connection across an election cycle on Tuesday. When the critical path runs through community consent and political approval rather than through procurement and construction, the variance on delivery timelines widens - and valuation multiples that assume predictable conversion compress.
On the evidence so far, the call is mixed but tilts structural at the margin. The M3 delay itself is cyclical - a deferral, not a cancellation. But the cause is structural: community opposition to data centre power infrastructure is a durable, replicable risk that has now been proven to move earnings, and it has appeared simultaneously across multiple jurisdictions. The market is right to treat this as a regime change in execution risk, even if NextDC's specific contracts survive.
The Counter-Thesis, and What Would Break It
The strongest argument against the structural read is that the market is conflating a single-site planning dispute with a sector-wide impairment. NextDC's problem is geographically specific - one 66kV line in Melbourne's inner west - and the company's response has been to pause only uncontracted projects while continuing to build contracted capacity. If the Victorian planning decision lands after the 28 November election and Jemena resumes in February, the FY28 recovery path could be steep, and the shares may look cheap at 37 times forward EBITDA relative to peers growing at a third of NextDC's pre-downgrade rate.
That argument holds only if the delay is contained. The signal that would falsify the "timing problem" thesis is concrete: if Jemena has not recommenced the M3 works by the end of February 2027, or if the company reports at its 20 November AGM that additional sites beyond M3 are facing community or council-driven power delays, the structural read is confirmed and the current valuation floor is illusory. A second marker: if NextDC cannot access its A$6.7 billion of undrawn senior debt by the next reporting period, the funding constraint has moved from caveat to reality, and the equity will be repricing leverage risk, not just timing risk.
What to Watch
In the short term, the rebound's durability depends on whether the broader AI infrastructure selloff stabilises. NextDC's steep monthly decline has tracked a global de-rating of data centre assets, and a single good day in Sydney does not reverse that tide. The 20 November annual general meeting is the first scheduled catalyst, when management will outline the outcome of its capital expenditure review and any capital recycling plans.
Medium term, the Victorian state election on 28 November 2026 is a political inflection point: a planning decision on M3 Stages 6 and 7 after the poll could either unlock the project or extend the delay into FY29. Jemena's February 2027 restart target is the operational line in the sand.
Long term, the sector's valuation regime depends on whether operators can institutionalise community engagement as a permitting capability rather than a public-relations exercise. NextDC said it would build on its existing community outreach through a board-overseen engagement plan for M3's construction and operation. Investors will be watching whether that plan converts into resumed works - or whether the next protest, in the next suburb, becomes the next guidance cut.
The rebound tells you the market is willing to buy the dip. The protest tells you the dip may not be a one-off. NextDC's contracts may yet deliver - but the era of assuming that power will arrive on schedule for Australia's AI buildout is over, and every data centre developer now prices a community-consent risk premium that did not exist a year ago.
Explore more exclusive insights at nextfin.ai.
