Is AI the new gold rush?

1. Is AI the new gold rush?

The last 12 – 18 months has seen a decisive shift in the venture capital landscape. AI has moved from being a talking point to a funding priority with all the hallmarks of a gold rush - frenzied investment, soaring prices and some speculation. The pace and breadth of transactions suggests a genuine step-change rather than just a cyclical uptick.

Capital allocation to AI has accelerated sharply both globally and within Australia. According to Crunchbase’s reporting on U.S. scaleup activity in 2025, AI is attracting the lion’s share of venture capital funding. While not all VC is flowing into AI, the allocation shift is material and is reshaping competition for high‑quality deals. Reporting from Cut Through Ventures points to AI as one of the most active categories by deal count and dollars in 2025 with 'AI and Big Data' consistently ranked top among investors’ 'most exciting sectors'. LinkedIn’s 2025 list of 'Top 20 Australian Startups' includes seven AI companies compared with just two such companies in 2024. This enthusiasm is reflected in funding trends as AI entered the top five most funded sectors for the first time (behind Fintech, Climate Tech, and Medtech) and ranked second by deal count in the Q2 2025 Cut Through Ventures report.

For investors, this is translating into increasingly competitive funding rounds and increasing valuations. Nonetheless, the outlook is that the opportunity for investors is significant with VC funds targeting AI across the full company lifecycle, from pre‑seed to growth stage. Early‑stage momentum is evident in rounds such as Rampersand’s investment in Keeyu and Blackbird’s backing of Enhance Labs, signalling a strong appetite to back emerging AI platforms. We are seeing an unprecedented level of investment at the seed stage with pre-revenue investments in the sector. At the later stage, there is similar growth with Harrison.ai and Heidi each raising north of $100 million this year. AI infrastructure is also booming with Firmus Technologies raising $330 million at a valuation above $1 billion and then tapping the market in the same year seeking another $500 million at more than double the initial valuation.

2. NVIDIA investment in Firmus

“During a gold rush, sell picks and shovels”

Ashurst advised NVIDIA on its investment into infrastructure company Firmus Technologies as part of Firmus’s $330 million capital raising in September 2025. The round valued Firmus at approximately $1.9 billion and establishes it as Australia’s newest unicorn.

NVIDIA’s participation underscores the strategic alignment between Firmus's renewable‑powered AI infrastructure ambitions and NVIDIA’s energy‑efficient AI computing. The raised funds will support Firmus’s rollout of Project Southgate to build four AI data centres (described as 'AI factories') across Australia and to partner with CDC Data Centres to retrofit their sites with Firmus technology. In focusing on infrastructure for AI, Firmus is staying true to the old business adage: “during a gold rush, sell picks and shovels.” Historically, those supplying the tools and infrastructure have secured stronger profits than prospectors, a pattern some investors expect to persist in AI.

One of the key considerations for any AI investment is understanding the technology. Firmus develops liquid‑based cooling designed to reduce energy consumption by around 30% and water usage by up to 99% relative to conventional systems, based on the company’s public disclosures. The value in the Firmus technology becomes apparent when you consider that for every 20 to 50 questions asked to a large language model (for example, ChatGPT), the equivalent of an estimated 500ml bottle of water is used for cooling the servers that generate responses. With an estimated 2.5 billion questions asked to ChatGPT daily, the importance of the Firmus technology becomes clear.

In November 2025, Firmus received commitments for a further $500 million equity raise at more than double the valuation of the earlier round to further strengthen its balance sheet and accelerate the rollout of Project Southgate.

3. Key areas of focus with AI investments

Compared to other regions, Australia has been relatively slow to adopt a clear approach to AI regulation. This has led to some uncertainty that the Australian Government has been seeking to dispel through its economic reform roundtables. The regulatory landscape still remains unsettled, but signs point to a less prescriptive regulatory model (compared to the European Union's detail-heavy AI Act), with sector-specific regulations being adopted.

In practice, this means clarification on what might be considered 'high risk' uses of AI, as well as incorporating accountability for AI influenced decisions, privacy and cross border data rules and regulatory alignment across jurisdictions. Successful AI companies need to take into account not just the pace of technological change but the potential regulatory headwinds.

One particular area of focus for AI companies has been Australia's copyright laws. Australia has strict copyright laws which prohibit unauthorised copying with very limited exemptions. This can be contrasted with the position overseas where there are broader exemptions (for example, fair use in the United States) and models are being considered where copyright infringement as part of training AI systems is allowed unless copyright owners opt out.

An exemption for copyright infringement for AI companies was proposed during the Government's Productivity Commission meetings in August. However, the government has indicated it has no plans for a copyright exemption to be introduced. Therefore, companies developing AI in Australia at the moment need to be careful to ensure they have the proper permissions and licences in relation to their training materials. This issue is likely to continue to generate headlines as AI companies seek to develop models in Australia and overseas.

Investors and governments are coalescing around a pragmatic approach to building out AI infrastructure. For example, the Government of New South Wales has been seeking to accelerate approvals for major tech projects such as large-scale data centres to support the continuing AI-led boom. The goal of regulation should be to facilitate long-term confidence and investment by developers, cloud providers and data centre operators and allow risks to be managed without sacrificing access to global AI technology. If achieved, this offers Australia more flexibility and innovation opportunities than the EU model and creates a stable, investable environment for AI and digital infrastructure.

4. 'Navigating the Legal Landscape AI in Australia'

Ashurst published a legal deep dive report on 30 June 2025 on how existing legal frameworks in Australia apply to the rapidly evolving field of AI covering: (i) consumer privacy, (ii) employment, (iii) misinformation / disinformation; and (iv) foreign interference law. For more information, see this link to the report.