Firmus Lands $10B Blackstone Debt To Power Australia AI Data Centers

Firmus said on February 9, 2026 it finalized a $10B debt funding package led by Blackstone and backed by technology investor Coatue Management, adding a large new pool of private credit to its buildout plans. The financing lands at a moment when markets are re-rating infrastructure tied to artificial intelligence, and when higher interest rates make the cost and duration of debt a direct driver of equity and credit risk.

The company said the debt will fund the next phase of Project Southgate, aimed at building AI training and inference capacity across Australia with data centers developed in collaboration with CDC Data Centres and Nvidia. The project targets up to 1.6 gigawatts of capacity over the next three years, a scale that brings energy, grid access, and equipment supply to the front of investor models.

$10B Private Debt Deal Reprices The Funding Channel For AI Builds

A financing package of $10B changes the capital stack discussion for large data center projects, because private lenders can move faster than public bond markets but typically demand tighter covenants and more control over cash flow waterfalls. For Blackstone, the deal expands exposure to a theme it frames as “picks and shovels,” and the size signals that private credit is willing to underwrite multi-year buildouts tied to AI demand rather than near-term contracted revenue alone.

For Firmus, the debt comes after equity funding raised in 2025 totaling A$830 million, which reduced early-stage balance sheet pressure before a heavier construction cycle. The mix of equity first, then debt, is a common structure when projects need land, permitting, and power agreements in place before lenders step in with scale, and it can lower dilution risk if timelines hold.

In broader markets, large private credit commitments can influence pricing for comparable infrastructure borrowers by anchoring expectations for leverage and duration. When $10B of supply is absorbed by one transaction, the marginal buyer in other credit deals can demand a higher spread, particularly if competing AI builds also need long-tenor funding in 2026.

Project Southgate’s 1.6 Gigawatt Target Raises Execution Stakes

Capacity targets near 1.6 gigawatts place the project in the category of national-scale infrastructure rather than a single-campus expansion. That matters because grid connection studies, transformer lead times, and substation upgrades can become gating items, and delays typically push interest expense higher before revenue begins to offset it.

The collaboration structure is also market-relevant. CDC Data Centres provides development and operational experience, while Nvidia’s role points to a compute stack designed around high-density training and inference. That linkage can pull spending into a defined supply chain, including servers, networking, liquid cooling, and power conversion equipment, where bottlenecks have already been visible in 2025–2026 ordering cycles.

Power Demand Becomes A Variable For Utilities, Inflation, And Rates

A buildout measured in gigawatts forces investors to model power as both a cost and a constraint. If Firmus secures long-term power purchase agreements at fixed prices, it can reduce margin volatility and stabilize debt-service coverage. If power is sourced at floating rates tied to wholesale markets, cost variability can rise quickly, especially during peak demand periods.

At the macro level, concentrated data center power demand can affect regional electricity forwards and grid upgrade capex, which can feed into inflation-sensitive sectors through higher utility costs and construction demand. That is one channel through which AI infrastructure can influence rate expectations, not by changing policy directly, but by changing the distribution of demand across energy and industrial inputs.

Equity Spillovers Extend Beyond Tech Into Contractors And Materials

Even if Firmus is privately held, the financing can move listed peers and suppliers by strengthening the narrative that data center pipelines in Australia remain financeable at scale. Contractors, electrical infrastructure providers, and specialized cooling and power equipment makers tend to benefit when a large project is de-risked by committed funding, because it converts a concept pipeline into a procurement timeline.

Semiconductor and networking suppliers can also see a read-through. When project funding powers capacity expansion, it supports multi-quarter demand visibility for accelerators, switches, storage, and interconnect, even if deliveries remain constrained by lead times. In Tredu risk mapping, that tends to lift sensitivity in tech equities to capex headlines, while pushing volatility toward names most exposed to capital intensity and power availability.

Foreign Exchange And Credit React Through Risk Appetite And Funding Costs

For foreign exchange, large-scale domestic infrastructure investment can support the Australian dollar at the margin if it signals sustained capital formation and imported equipment demand, though day-to-day direction remains dominated by global rates and commodity terms of trade. The more immediate market response usually shows up in credit: when lenders demonstrate willingness to finance AI-linked builds, spreads for adjacent infrastructure themes can tighten, but only if power and permitting risks look containable.

If costs rise, the credit channel can reverse. Higher construction inflation or grid delays can lift interest carry and weaken coverage metrics, creating a wider dispersion between projects with secured power and those still negotiating access.

Base Case: Funding Supports A Steady Three-Year Buildout

The base case is that the new debt package is drawn in stages aligned with milestones, keeping interest carry proportional to delivered capacity and limiting balance sheet strain. Under this path, capacity ramps toward the 1.6 gigawatt target over three years, and spending remains supportive for data center equipment and construction without forcing a repricing in Australian power markets.

A concrete trigger for this outcome is signed grid connection and power procurement progress that reduces the probability of idle assets, alongside visible equipment delivery schedules that avoid long slippages.

Upside Scenario: Faster Customer Commitments Pull Forward Returns

The upside scenario requires large customer commitments that raise utilization quickly once new centers are live, improving unit economics and accelerating cash flow. If AI demand remains strong and offtake is secured earlier than expected, the project can justify additional build phases, and debt terms can become less restrictive in refinancings.

Triggers include multi-year compute capacity agreements, smoother supply deliveries for high-density equipment, and stable power pricing that preserves margins.

Downside Scenario: Grid Bottlenecks Or Capex Fatigue Slow The Ramp

The downside scenario is driven by constraints rather than demand, with grid bottlenecks, permitting friction, or equipment lead times pushing schedules back and raising financing costs. A separate downside risk is a pause in AI capital spending by major buyers that reduces urgency, leaving new facilities underutilized while interest expense accrues.

Triggers include delays in substation upgrades, higher-than-expected electricity costs, or a sustained pullback in AI infrastructure ordering that forces projects to phase more slowly.