What would a $270b AI investment look like for Australia?
Welcome to my new substack – a thought experiment on a nation building investment on AI instead of AUKUS submarines.
Welcome to my new Substack!
I’m going to be regularly posting here as an early stage venture capitalist from Australia – sharing what I’m seeing and hearing on AI, investing, startups and technology Down Under.
I co-founded and launched Galileo Ventures just over 3 years ago now with our $10m Fund I. It’s a total privilege to invest and work with exceptional founders at the cutting edge – I like to say we back entrepreneurs at the edge of the forest on a lot of macro and tech trends.
But to kick things off here is a thought experiment 🧠
A few weeks the Australian government poured $1B into PsiQuantum – a bold move to build a Quantum computer and a ‘moonshot project’ for Queensland. It’s received a lot of criticism as being a big bet on one company, some of it I disagree with as I think Australian government needs to make more bold bets on technology.
So, what would a bold nation building investment in AI look like – a technology that has immediate implications?
Imagine if we spent the money for 5 nuclear submarines1 – a total of $270 billion – on AI instead?
A $270 billon investment in AI
These a 5 potential implications:
We would be the biggest AI investor in the world – outpacing the USA and UK and firmly establishing AU as serious contender.
We could train our own frontier AI models (and build our own chips) – Frontier models are expensive to train, with some estimates now seeing costs going over $1b to train2, and attract a chip manufacturing industry in Australia from Taiwan. For example, the US just pledged $US5B co-investment for their Semis.
We could transform industrial competitiveness – Fund R&D into global solutions, boost our VC sector, attract international co-investment and expand our fledgling tech sector with more high-paying jobs.
Rapidly improve our way of life –We could become a leader in adopting AI in healthcare, education and social services becoming a pioneer developing advance services available to everyone, effectively and cheaply.
Finally, we could (probably) solve our defence capability. Such as developing the first fleet of autonomous vessels to monitor and deter in our waters.
Of course, I know such a nation-building investment by a small country government is naive. But should it be?
The AUKUS agreement which means AU can access nuclear sub tech is under a lot of scrutiny – one for the huge estimated cost, slow pace of delivery and if it will be actually relevant for AU defence in 25 years time.
There a lot of estimations but the CEOs of LLM companies themselves are now saying $1b training runs are on the horizon. See this interview.