SuperX FY2025 Results: Operations Briefing and What's Coming in FY2026
SuperX AI Technology Limited (NASDAQ: SUPX) closed FY2025 as a transition year-moving from interior design to full-stack AI infrastructure. The reported numbers reflect legacy work with a late-quarter bump from initial AI equipment sales. Heading into FY2026, the company has capital, new joint ventures, and an execution plan centered on AI data center buildouts.
FY2025 at a glance (legacy-heavy results)
- Revenue: $3.6 million, including roughly $1 million from AI servers and related IT equipment in June 2025.
- Net loss: $21.2 million, driven mainly by non-cash items and higher staffing and professional costs tied to AI R&D ramp-up.
- Liquidity and assets: $17.2 million in cash and cash equivalents; $52.1 million in total assets at year-end.
- Note: FY2025 results primarily reflect legacy services, with partial AI contributions after the rebrand to SuperX.
For detailed disclosures and risk factors, review the company's latest Annual Report on Form 20-F filed with the SEC. EDGAR search
What changed after year-end (July-October 2025)
- Electrical systems: "SuperX Digital Power" JV with Zhonhen Electric focused on HVDC for AI data centers; launched the Panam-800VDC end-to-end solution (new builds) and Aurora-800VDC retrofit solution (existing sites).
- Thermal management: "SuperX Cooltech" JV with Chengtian Weiye covering CDUs, Micro-Channel Liquid Cold Plates, and integrated cooling systems.
- GPU supply and design access: Majority stake in MicroInference, a Singapore-based NVIDIA Partner Network Solution Provider, to strengthen supply and solution design. Learn more about the program context here: NVIDIA Partner Network.
- Global services: "SuperX Global Service" JV with Teamsun to deliver deployment, maintenance, and managed services at scale.
- Footprint: Japan AI Supply Center planned to be operational by end of 2025; U.S. subsidiary in Silicon Valley targeted to be operational by end of 2025 for R&D and market delivery.
- Product releases: Next-gen AI servers (XN9160-B200, XN9160-B300) and an All-in-One Multi-Model Server series. New rack-scale AI platform based on NVIDIA GB300 chips. A Modular AI Factory solution aiming to compress data center deployment cycles to under six months.
What this means for Operations
If you manage data center capacity, procurement, or deployment, the signal is clear: SuperX is organizing a full lifecycle offering-from electrical and cooling to servers, integration, and services. Below are the operational angles to evaluate now.
- Electrical architecture: 800VDC implies a different distribution strategy, safety protocols, and testing procedures. The "retrofit" option suggests brownfield paths without a full rebuild, but still demands facility audits and staff training.
- Cooling shifts: Liquid solutions (CDUs, cold plates) change floor layouts, serviceability, water treatment, and monitoring. Plan for leak detection, maintenance windows, and spare part stocking.
- GPU allocation and lead times: NPN access may help with supply, yet you should still lock procurement windows and stage deliveries to match installation crews and commissioning schedules.
- Faster deployment promises: A data center-scale modular approach can shorten delivery, but only if permitting, site readiness, utility interconnects, and compliance documentation are in sync. Build a cross-functional timeline up front.
- Global services and SLAs: With a services JV, ask for standard operating procedures, escalation paths, and on-site response times by region. Validate RMA logistics and spares pools before go-live.
- Vendor stability and financing: The company reports more than $70 million raised since March 2025 and agreements for over $170 million more in October 2025. Confirm support commitments and financing options for multi-quarter rollouts.
FY2026 roadmap: checkpoints for customers
- Integration of acquisitions and JVs: Request a unified product catalog, compatibility matrix, and reference architectures that cover electrical, cooling, servers, and services.
- AI factory deployments in key markets: Ask for pilot outcomes, performance data, and total cost envelopes for greenfield and retrofit projects.
- Partner ecosystem: Clarify relationships with silicon vendors, OEMs, and financial institutions; explore leasing or consumption models if capex is constrained.
- Service coverage: Verify on-the-ground teams, language support, and change-management alignment in your regions.
Quick action list for Ops leaders
- Run site surveys for 800VDC readiness and liquid cooling feasibility (including safety, compliance, and training needs).
- Model thermal and electrical loads for rack-scale deployments and define commissioning/acceptance test plans.
- Engage procurement now to secure GPU allocations and coordinate logistics with installation milestones.
- Finalize SLAs: response times, escalation policies, spares, and RMA processes; document roles and handoffs.
- Upskill your team on AI data center operations and tooling. If you need structured learning paths, see job-focused options here: Complete AI Training - Courses by Job.
Key dates to watch
- Japan AI Supply Center: expected operational by end of 2025.
- U.S. subsidiary (Silicon Valley): expected operational by end of 2025.
Note: This update includes forward-looking statements from the company. For risks and assumptions, see its latest Form 20-F on the SEC's website. EDGAR search
About SuperX
SuperX AI Technology Limited provides AI infrastructure solutions spanning hardware, software, and services for AI data centers. Offerings include high-performance AI servers, 800VDC solutions, high-density liquid cooling, AI cloud, and AI agents. The company is headquartered in Singapore and serves global institutional clients. More at superx.sg.
Your membership also unlocks:
