AI Progress Stalls, Costs Fall: New Chance for Australia

Published Date : 26/09/2025 

Despite the hype and massive investments, AI progress has slowed, and models are becoming cheaper. This could be a golden opportunity for Australia to build its own AI infrastructure and compete on a global scale. 

Following the news in artificial intelligence recently has been like watching an endless back-and-forth tennis rally between hype and disillusionment.

We're told the AI investment bubble is deflating. Revenues are too low to justify the investment. New AI products — like GPT-5 — are underwhelming. But wait, AI is the next great industrial revolution. Decisions now will shape our future prosperity. We should scrap copyright laws to allow AI companies to train models for free. Your job will be automated. Data centres will consume our resources and pollute our atmosphere. No, we should be bracing for the imminent AI crash. Or maybe not? Australia should build its own AI to compete with ChatGPT and adequately reimburse artists for their work. But these models can cost billions of dollars to train, which we can't afford. And on it goes.

So what is really happening? One way of understanding the mixed signals around AI is to recognize we're looking at two very different visions of the future, each struggling to become reality. One is dominated by big tech, mostly in the US. In the other, the ability to build powerful AI models is much more widely shared, and Australia is not only a consumer of this technology but a producer.

A few years ago, the popular story told about AI was a simple one, if a bit bleak. It went like this: a few big companies would emerge as masters of the new technology and own the means of production for much of the world. AI would be everywhere, embedded in every car, phone, and fridge, every search query and online ad, but the cost of building an AI model would be so high that these few companies would mostly enjoy a monopoly. They would use their models to train even better models, and so on, widening the gap with any potential rivals. This would be very, very profitable. AI companies would be all-powerful. They may even invent and then own 

Frequently Asked Questions (FAQS): 

Q: What is the current state of AI progress?

A: AI progress has slowed, and new models like GPT-5 are not as impressive as expected, despite the high costs of training.

Q: Why are big tech companies still investing heavily in AI?

A: Big tech companies like Nvidia and Meta are investing heavily to maintain their competitive edge and to develop new generations of AI, even though progress has slowed.

Q: How can Australia benefit from the slowed AI progress?

A: Australia can benefit by building its own AI models using open-source technology, which is becoming more competitive and cost-effective.

Q: What are the potential risks of the AI bubble bursting?

A: Some AI companies might be overvalued, and there could be a significant downturn similar to the dot-com crash, affecting the industry and investments.

Q: What are the practical applications of AI that are gaining more interest?

A: Practical applications like coding software, organizational workflows, and basic tasks such as transcribing and summarizing are gaining more interest as the focus shifts from super-intelligence. 

More Related Topics :