An Australian AI infrastructure startup raises $2 billion with Nvidia backing to build and operate data centers for AI training and inference.
Firmus, an Australian AI infrastructure startup, has raised $2 billion in equity at a valuation exceeding $10.5 billion, with Nvidia among its backers. This funding round exemplifies a broader shift reshaping the sector: AI investment has begun to resemble infrastructure project finance rather than traditional venture capital, increasingly underwritten by the companies selling the hardware.
The scale of the raise places it far outside the normal venture capital range. Two billion dollars in a single equity round typically belongs to the realm of public debt markets or strategic corporate investment, not private funding—and this capital is being deployed to build data centre capacity rather than develop software.
Firmus is one instance of a wider pattern. Global startup investment reached a record $510 billion in the first half of 2026, already surpassing the $440 billion invested across all of 2025, with nearly 40 AI startups reaching unicorn status in just six months. Yet the composition of that capital has shifted as much as its volume. A growing share is no longer funding product development, but rather buildings, power contracts, cooling systems and chips. Since compute is AI's binding constraint and compute is a physical asset, capital is flowing toward whoever will build it.
This reframes the nature of these rounds. Traditional venture funding buys equity in a business that might scale at near-zero marginal cost. Infrastructure funding buys a share of an asset with construction timelines, depreciation schedules and utility bills. The AI sector is increasingly performing the latter while calling it the former.
Nvidia's position on Firmus's cap table fits an established pattern. In late July, Nvidia committed $5 billion to Safe Superintelligence, granting the startup access to Nvidia's Vera Rubin systems and a tenfold increase in compute over twelve months. Nvidia is reportedly in talks to guarantee roughly $250 billion in lease and construction financing for OpenAI's Ohio data centre campus, with parallel discussions covering up to $350 billion in chip purchases.
In each case, capital supplied by Nvidia returns to Nvidia as hardware demand. Vendor financing is well-established in capital-intensive industries, and Nvidia has both the balance sheet and strategic interest to fund buyers when compute is the bottleneck. However, this arrangement concentrates risk distinctly. A meaningful share of Nvidia's demand is now underwritten by Nvidia itself, and the value of those positions depends on the same AI boom that drives Nvidia's revenue. When both sides of a transaction are exposed to the identical variable, a downturn does not arrive as one isolated problem.
The timing of Firmus's round occurs in a testing environment. More than $1.3 trillion was erased from the market value of the world's largest semiconductor companies in a handful of trading sessions in late July—Nvidia itself losing roughly $238 billion—driven not by falling demand but by investor questions about whether AI infrastructure spending will produce adequate returns.
Combined 2026 capital expenditure from Amazon, Alphabet, Meta and Microsoft is tracking toward roughly $740 billion, up approximately 77 per cent year-on-year. Meta has moved a $14 billion data centre off its balance sheet to BlackRock-managed funds. Tesla and SpaceX have committed $16.8 billion to a chip factory that could eventually reach $119 billion.
Firmus raising $2 billion at a $10.5 billion valuation signals that private capital markets have not yet reached the scepticism the public markets briefly displayed. Whether that gap closes, and in which direction, is the question underlying every one of these figures.