
Firmus Closes $2B Funding Round: Valuation Surpasses $10.5B, NVIDIA Partnership Powers 1.6GW Australian AI Factory, Asia-Pacific Expansion Accelerates
Transaction Overview: From Bitcoin Miner to $10.5B AI Infrastructure Giant
On August 7, 2026, Australian AI infrastructure company Firmus announced the completion of a $2 billion strategic equity investment round — a major milestone in Asia-Pacific's AI infrastructure sector.
Key Financial Metrics
| Metric | Value |
|---|---|
| Current Round Size | $2 billion |
| Post-Money Valuation | Over $10.5 billion |
| Valuation Growth Since April | ~100% |
| Total New Equity Raised (Past 12 Months) | Over $3 billion |
| Separate Debt Facility (Blackstone/Coatue) | $10 billion |
Investor Lineup
This round attracted top-tier institutional investors:
- Existing Investors: NVIDIA, Coatue (continuing to increase stakes)
- New Investors: Blackstone (via Blackstone Tactical Opportunities and affiliated vehicles), Jane Street (global top quantitative trading firm)
Jane Street's participation is particularly noteworthy — this institution, known for quantitative trading, rarely participates in early-stage tech investments. Its entry is viewed by analysts as a signal that Firmus may be preparing for an IPO.
From Bitcoin Miner to AI Infrastructure Giant: Firmus's Transformation Story
Firmus was founded in 2019 by Oliver Curtis, Tim Rosenfield, and Jonathan Levee, originally as a Bitcoin mining company. This history is not coincidental — Bitcoin mining and AI training share striking similarities in infrastructure requirements:
- Large-Scale Power Procurement: Both require stable, large-volume power supply
- Rapid Site Development: Need to build large computing facilities in short timeframes
- Thermal Management: Heat dissipation for high-density computing equipment is a core challenge
- Energy Efficiency Optimization: Reducing energy consumption per unit of computation is a competitive key
Firmus successfully translated these capabilities accumulated in Bitcoin mining into core competencies for AI infrastructure.
Project Southgate: Australia's Largest AI Factory Plan
Scale and Objectives
Project Southgate is Firmus's flagship Australian AI infrastructure program:
- Target Compute Capacity: Approximately 1.6 gigawatts (GW) by 2028
- Number of Sites: 5 or more sites
- Geographic Distribution: Covering multiple Australian states
What does 1.6GW of compute capacity mean? For reference, a typical large data center has approximately 100-200 megawatts (MW) of capacity. 1.6GW is equivalent to 8-16 large data centers, which would make Firmus one of the largest AI compute providers in Asia-Pacific.
HyperCube Platform: Technical Differentiation
Firmus's core technology is its proprietary HyperCube platform, combined with NVIDIA's DSX AI Factory Reference Architecture:
Energy Efficiency Optimization: The HyperCube platform focuses on improving tokens-per-watt efficiency — the core metric for AI inference costs.
System Resilience: Integration with NVIDIA DSX architecture ensures high availability and maximizes GPU utilization.
Green Design: Firmus positions its facilities as "green AI factories," integrating advanced cooling technologies and renewable energy, aligning with Australia's energy policy requirements.
Asia-Pacific Expansion: Indonesia Is the Next Target
Beyond Australian domestic expansion, Firmus is actively positioning in other Asia-Pacific markets:
Indonesia: Strategic Opportunity in Southeast Asia's Largest Market
Firmus has already initiated early-stage development projects in Indonesia, targeting AI-native customers. Indonesia's appeal lies in:
- Population Scale: 270 million people, Southeast Asia's largest market
- Digital Acceleration: The Indonesian government is actively promoting digital transformation
- Data Sovereignty Needs: Indonesian enterprises and government have strong demand for local data storage
- Growing AI Demand: As Indonesia's AI ecosystem develops, local compute demand is growing rapidly
Asia-Pacific AI Infrastructure Gap
According to market data, Asia-Pacific AI infrastructure investment reached approximately $17.77 billion between August 2025 and September 2026 (37 transactions), with Asia-Pacific accounting for approximately 25.88% of global AI infrastructure funding.
Firmus's expansion plan directly targets this gap, particularly in Australia and Southeast Asia, where AI compute supply falls far short of demand.
Unique Challenges of Australian AI Infrastructure
Building AI infrastructure in Australia faces some unique challenges, and Firmus's success is partly due to how it addresses these:
Energy Grid Constraints: Power grid capacity and renewable energy ratios vary significantly across Australian states, requiring large AI factories to coordinate closely with grid operators.
Geographic Dispersion: Australia is vast with sparse population concentrated in coastal cities, requiring data center site selection to balance power supply, network connectivity, and land costs.
Regulatory Environment: Australia has strict scrutiny of foreign investment in critical infrastructure; Firmus as a domestic company has a natural advantage.
Talent Shortage: AI and data center engineers are relatively scarce in Australia, requiring active talent recruitment and training programs.
IPO Prospects: Analyst Assessment
Multiple industry analysts believe Firmus's funding structure and investor lineup indicate the company may be preparing for an IPO, with the Australian Securities Exchange (ASX) as a potential listing venue.
Factors supporting this assessment:
- Jane Street's Entry: Quantitative trading institutions typically enter in the pre-IPO period
- Rapid Valuation Growth: Valuation nearly doubled in 4 months, suggesting pricing for public markets
- Increased Transparency: The company has begun more actively disclosing financial and operational data
- Institutional Investor Mix: Large PE firms like Blackstone typically participate in the pre-IPO period
Impact on Asia-Pacific AI Ecosystem
Firmus's success has important implications for the entire Asia-Pacific AI ecosystem:
Compute Democratization: As Firmus expands compute supply, AI developers and enterprises in Asia-Pacific will be able to access high-performance GPU compute at lower costs.
Localization Advantages: Compared to relying on US or European cloud services, local AI infrastructure provides lower latency, better data sovereignty protection, and services more aligned with local regulations.
Ecosystem Catalysis: The presence of large AI infrastructure often attracts AI research institutions, startups, and large enterprises to build ecosystems in the surrounding area.
Conclusion
Firmus's transformation from Bitcoin miner to $10.5 billion AI infrastructure giant epitomizes the Asia-Pacific AI infrastructure investment boom. Its $2 billion funding, NVIDIA partnership, and 1.6GW compute target mark the Asia-Pacific region building the physical foundation to support next-generation AI applications.
For AI developers, enterprises, and investors in Asia-Pacific, Firmus's rise means that local high-performance AI compute will no longer be a luxury but accessible infrastructure.


