Artificial intelligence (AI) will contribute more than $15 trillion to the global economy by 2030. According to the University of Adelaide, greater use of AI in key Australian industries will lead to a short-term boost in GDP of more than $200 billion per year and the creation of an additional 150,000 jobs from 2023 to 2030. Although we’ve experienced an explosion of AI, there’s still massive value yet to be created—especially in the industries underpinning our economy.

With foundational models becoming widely available and AI adoption accelerating, the challenge for investors is no longer just identifying AI startups, but rather distinguishing the ones building durable, defensible value. So how can companies build sustainable value and differentiation in a world where models are increasingly commercialised? 

Beyond the hype, what makes an AI company valuable?

The rise of AI has made it easier to create applications that leverage large models and solve a business problem. However, simply integrating a foundational model into a product does not create lasting value. At the application layer, we look for AI companies that meet three core criteria:

1. Domain expertise and data advantage: Foundational models are commodities, and differentiation comes from proprietary datasets and deep industry knowledge. The best AI companies leverage their understanding of specific verticals to create defensible advantages—from biotech to industrial productivity, healthcare to decarbonisation.
   
2. Real productivity and decision-making gains: Customers are trialling AI solutions at an unprecedented rate, but many companies are experiencing high churn. The AI startups that will last are those delivering non-trivial improvements in efficiency, accuracy, or automation—solving real-world problems rather than just demonstrating technical capability.
   
3. Speed and execution: AI has made it faster and cheaper to build, test, and iterate on software products. In the past, the highest-performing SaaS companies would take an average of 60+ months to achieve $30M in annualised revenue. With AI, this has been reduced to 20 months. The companies that succeed will be those that can rapidly experiment, learn from customer feedback and improve their models at an accelerated pace.
   

The pitfalls of AI investing: what we avoid

It’s important to recognise that not all AI companies will build the high-growth sustainable businesses necessary for venture capital. As investors, we remain cautious about:

• Thin wrappers around foundational models: Companies that offer little more than an interface on top of commoditised AI models without a defensible moat, such as domain expertise or deep industry data.
  
• Compute-heavy models without a path to efficiency: Businesses that rely purely on scaling compute spend without clear paths to efficiency and optimisation.
  
• Hype-driven, short-term plays: Startups chasing rapid adoption but lacking a strong product-market fit, long-term differentiation, or sustainable revenue models. As we know, rapid growth may not always reflect long-term commercial viability.
  

Where we see long-term AI value

As we refine our AI investment thesis, we remain focused on companies that align with our long-term strategy—combining cutting-edge scientific discoveries with the world’s biggest challenges to build the next generation of global companies.

Our first focus area is AI’s role in whitespace industries. AI has largely been leveraged for business productivity, such as email summarisation, preliminary research and reporting. But we’re invested in companies that dig a little deeper. There’s an abundance of whitespace industries, like biotech, healthcare and industrial automation, that have massive market potential and overlap with our thematic challenge areas. So in brief, we’re focused on companies using AI to solve problems for high-value, underserved markets that have yet to be upheaved by AI. We back several companies in this space including Lumachain, Kasada, Presien, Pending AI and Regrow. 

Another area we’re interested in is AI which drives unprecedented efficiency. Whether through model compression, optimised inference, or novel architectures, we are looking for startups that push AI forward while reducing its resource intensity. This includes software architecture plays and hardware/semiconductor-focused plays. We’re currently supporting a stealth company working on a hardware solution in this space.

We’re also giving our attention to physical AI—robotics enhanced by generative AI. The intersection of AI and robotics (physical AI) is creating new opportunities for robotics that are more scalable and efficient at completing complex tasks. We are particularly interested in software-driven approaches that enhance robotic capabilities without requiring heavy hardware investments. Portfolio companies like Breaker and RIOS demonstrate our existing belief in this space.

There’s no doubt that the AI sector will continue evolving at a rapid pace. The most successful investors will be those who can separate short-term trends from lasting value. We are committed to backing companies that use AI not just as a feature, but as a fundamental driver of transformation in key industries. Does this sound like your innovation? If so, get in touch! 

‍