Summary
- Fujitsu agreed to sell its Australian data-centre business to Next Capital, with completion expected later in 2026 subject to customary conditions and approvals.
- The reported perimeter consists of five operating sites: Homebush, Greystanes, Eight Mile Plains, Noble Park and Malaga. North Ryde and its underlying land are excluded.
- Fujitsu will continue to operate the business as usual until completion; customer arrangements and existing data-centre employees are intended to transfer at closing through a jointly managed transition.
- Neither party disclosed the purchase price. The strategic test is whether dedicated ownership can invest in the platform without disrupting the critical workloads it already hosts.
“Nothing changes for customers today” is the easiest sentence in Fujitsu’s announcement to read and the hardest one to deliver.
Next Capital is not buying five inert shells. It has agreed to acquire a functioning national data-centre business whose value depends on power staying available, incidents being handled, certifications remaining effective, employees knowing the sites and customer contracts surviving the change of control. The transaction is signed, but those operating relationships are still on Fujitsu’s side of the handoff until completion.
Continuity is therefore not a courtesy promised alongside the deal. It is the first asset the buyer and seller must preserve.
Five sites move; a sixth stays outside
The reported perimeter includes Homebush and Greystanes in New South Wales, Eight Mile Plains in Queensland, Noble Park in Victoria and Malaga in Western Australia. Next Capital’s own news page identifies those five operating locations and says North Ryde and the underlying land are not included.
That exclusion matters. Fujitsu’s public location page lists six Australian data-centre locations, so a reference to the “Australian business” should not be read as every facility associated with the company. The deal creates a five-site operating platform spanning Sydney, Melbourne, Brisbane and Perth, while leaving a separate Sydney asset outside the reported package.
The parties have not published a detailed asset schedule, individual property tenure, customer-by-customer consent requirements or site-level financials. Industry reports cite capacity, utilisation and valuation estimates with different scopes. Those estimates should not be turned into transaction facts.
What is public is the control boundary: five live sites, an excluded North Ryde asset, customer arrangements intended to move at completion and staff expected to move with them.
An agreement is not yet a handover
Fujitsu expects completion later in the year, subject to customary conditions and approvals. Until then, it says it will operate the business as usual. That sequence separates three states that are often collapsed in acquisition headlines.
The first is signing. The buyer and seller have agreed the transaction and defined a future transfer.
The second is the pre-close period. Fujitsu remains responsible for the platform. Customers still need ordinary service, changes, incident response and contractual accountability while regulatory and other conditions are addressed.
The third begins at completion. Customer arrangements are expected to transfer to the new business through a transition managed jointly by Fujitsu and Next Capital. Existing data-centre employees are also expected to move. Only then does the promise of an independently operated Australian platform become an operating fact.
This structure contains more execution risk than moving title to buildings. Some contracts may require notices, consents or novation mechanics. Operational tooling, vendor agreements, access controls and escalation procedures may need to be separated from Fujitsu systems. The public statements do not disclose the details, so they cannot be assumed complete.
Fujitsu is separating the facility layer from the service layer
Fujitsu describes the sale as a capital-allocation choice. It wants to invest more deeply in modernising critical systems, cyber resilience, sovereign AI, high-performance computing and quantum computing. It says the data-centre platform, meanwhile, will benefit from dedicated commercial ownership and investment.
The logic is a vertical separation. The physical layer supplies controlled space, power, cooling, security and connectivity. The technology-services layer designs, integrates and operates the systems customers place on top. A company can consider the second layer strategically central while deciding that a specialist owner is better placed to fund the first.
That does not make the five sites obsolete or peripheral. Their workloads may be critical to the very modernisation and resilience services Fujitsu wants to sell. The separation succeeds only if customers can continue using the facilities while Fujitsu changes where it commits capital.
Nor does Fujitsu’s reference to sovereign AI make this an acquisition of five AI campuses. The sites are established enterprise-grade facilities with varied workloads. Any future AI upgrades will depend on available power, cooling, density, customer demand and investment after closing. The announcement supplies a strategy, not a completed retrofit plan.
The missing price is less useful than the operating perimeter
Neither Fujitsu nor Next Capital discloses the consideration. Next Capital’s page carries an attributed report that the price was just below an estimated A$200m replacement value, but that is not a figure confirmed by the parties and should not be treated as the purchase price.
The buyer says the acquisition sits in its A$375m fifth fund. That identifies the investment vehicle, not the amount earmarked for this deal or the capital available for post-close upgrades. Next Capital’s earlier ownership of Australian operator iseek gives it relevant sector history, but a prior investment does not establish the plan for these sites.
Without audited portfolio figures or a disclosed price, valuation multiples would be guesswork. The more informative public evidence concerns what must move: five operating sites, customer arrangements and the workforce that maintains continuity.
That workforce is particularly important. Data centres are asset-heavy but operationally dependent on local knowledge—power systems, maintenance history, security processes and customer-specific change procedures. Transferring existing employees reduces the risk that ownership changes faster than operational memory.
Expansion comes after a clean transfer
Dedicated ownership may give the new platform a clearer reason to invest in capacity, efficiency and customer acquisition. It can also expose the business to the usual private-equity questions: how much capital will be committed, which sites can support higher density, what returns are required and how much operational flexibility customers retain.
None of those answers is in the signing announcement. The immediate milestones are more prosaic and more important: approvals, completion, contract transitions, staff transfer, uninterrupted service and a clear operating identity outside Fujitsu.
If those steps work, Next Capital receives a live platform rather than an integration problem, while Fujitsu can pursue its services strategy without severing the infrastructure relationships beneath it. If they do not, the first symptom will not be a missed expansion target. It will be friction for customers who were told that nothing changes today.