We keep hearing that artificial intelligence is going to be the great deflationary force of our lifetime. The theory is simple: robots do the work cheaper, software writes itself, and the cost of everything plummets. But if you look at what is happening on the ground in Australia right now, the reality is looking a lot more expensive.
The Reserve Bank of Australia recently moved their benchmark interest rate to 4.85%. That is an eighteen-year high. For those keeping score, that means borrowing costs are the most restrictive they have been since before the 2008 financial crisis. The culprit? It is a mix of high energy costs and, surprisingly, the massive capital expenditure required to fuel the AI boom.
The infrastructure tax
For founders and builders, this is a wake-up call. We usually think of AI as a digital expense—a line item in a cloud computing bill. But for a national economy, AI is a physical burden. Data centers do not run on vibes; they run on massive amounts of electricity and cooling. Australia is finding out that as tech companies rush to build out the capacity for generative AI, they are competing for the same energy and labor resources as everyone else.
This creates a feedback loop that central banks hate. Tech companies have deep pockets and are willing to pay a premium for power and specialized talent to get their chips spinning. That drives up the baseline cost for every other business in the country. When the cost of doing business goes up across the board, the central bank reaches for the only tool it has: the interest rate lever.
Why builders should care about 4.85%
If you are building a startup in the Web3 or AI space, high interest rates are your biggest headwind. It is not just about the cost of a business loan. It is about the opportunity cost for investors. When the RBA keeps rates high, venture capital starts to look for safer returns elsewhere. Why bet on a high-risk crypto protocol or a nascent AI agent when you can get a guaranteed, historically high return on government debt or savings?
For those of us in the trenches, this means the era of cheap experimentation is officially over. Every bit of compute you use and every developer you hire is now being priced against an economy that is trying to cool down. If your business model relies on "free" money or subsidized cloud credits that are eventually going to expire, you are in a dangerous spot.
Energy is the new bottleneck
The Australian situation highlights a growing rift between the software world and the physical world. We have spent the last decade acting like software has zero marginal cost. But AI has changed the math. The energy required to train and run large language models is staggering. In Australia, where energy prices were already a sensitive political and economic issue, the added strain of data center expansion is pushing the grid to its limits.
This is where the skeptics among us get a bit loud. We are being told that AI will solve the climate crisis and optimize our grids, but in the short term, it is doing the exact opposite. It is acting as a massive vacuum for energy, driving up prices for the average household and forcing the central bank to tighten the screws on the whole population.
The Founder's perspective
So, what do we do with this? First, stop assuming that tech is inherently deflationary in the near term. We are in the "build-out" phase. This is the heavy lifting part of the cycle where we are digging trenches and laying cable. It is expensive, it is messy, and it is inflationary.
- Efficiency is a feature, not an afterthought: If you are building AI tools, the winners will be those who can do more with less compute. High energy costs mean that lean models will eventually outperform bloated ones on a margin basis.
- Geography matters: Australia's rate hike shows that local infrastructure constraints have global impacts on the tech ecosystem. Where you host and where you build actually matters again.
- Capital discipline: At 4.85%, your burn rate is your biggest enemy. There is no bailout coming from the central banks anytime soon.
The hard truth is that the AI revolution is currently being subsidized by higher interest rates for everyone else. We are building the future on a very expensive credit card.
A Reality Check on the AI Hype
We often talk about AI in these abstract, ethereal terms. We talk about "the cloud" as if it is some magical dimension. The RBA's move reminds us that the cloud is just someone else's computer in a hot room that needs a lot of power. When that power gets expensive, the whole economy feels it.
I am still long on the potential for these technologies to transform how we work, but I am skeptical of the timeline for when these efficiencies will actually hit the bottom line of the average person. Right now, the average Australian is paying more for their mortgage because tech companies are in an arms race to build bigger and hungrier data centers.
The Takeaway
The RBA hitting an 18-year high is a signal that the "tech boom" is no longer insulated from the rest of the economy. For builders, the message is clear: the market is demanding that you prove your value. If you are consuming massive amounts of energy and capital, you better be producing something that justifies the cost. The days of growth at any price are gone, replaced by a world where the cost of power and the cost of money are finally catching up to the hype.
Read the original at Cointelegraph →