Paradox: AI may reduce long-term business profits, and debasement raises costs. A measured corporate Bitcoin strategy may offer a solution

TL;DR

The AI Paradox is squeezing businesses from both sides:

• AI is getting cheaper & more useful → clients expect lower prices for knowledge work
• AI-first competitors can undercut traditional firms
• Meanwhile, non-AI costs (salaries, rent, taxes) keep rising due to monetary debasement/"inflation"

Result? Decreasing revenues + increasing costs = profit squeeze

The numbers are there:

• ChatGPT went from sub-90 IQ to 116 IQ (top 15% of humans) in 11 months. AI agents will amplify the real-world impact of AI
• Your R1,000 today for petrol will only buy what R104 bought 25 years ago (a stark South African example). In the UK, £50 used to buy 58 litres of petrol; now it only buys 37.

My proposed solution: A phased corporate Bitcoin strategy

Why Bitcoin?

• Hard-capped at 21 million (vs infinite fiat printing)
• Has outperformed fiat currencies over medium-long term
• Bitcoin ETFs & other Bitcoin products are now widely accepted as a formal asset class
• Already being used by businesses like Steak 'n Shake, Microsoft, Starbucks, Apple
• Could help businesses preserve longer-term purchasing power

The strategy: Start small, gradually shift some measured amount of treasury holdings toward Bitcoin while also accepting it for payments, and maybe using it for some business costs. Think of it as insurance against the AI commoditisation + monetary debasement combination.

Companies that don't adapt to both AI and the changing money game risk getting squeezed out of existence.

!Nothing in this article is financial advice!

The information and views expressed in this article are for general discussion and educational purposes only. They do not constitute financial, investment, tax, legal, or accounting advice, and should not be relied upon as such. Every business and individual’s circumstances are unique; before acting on any ideas presented here - including those involving AI integration or Bitcoin adoption - seek professional guidance from qualified advisors who understand your specific situation. Past performance and historical trends are not guarantees of future results, and all investments or strategic changes carry risk.

AI's increasing Intelligence : Cost and Usefulness : Cost

AI's Intelligence & business value is increasing

I run an AI consultancy and have seen first-hand how AI/LLM providers like OpenAI (ChatGPT), Anthropic (Claude), xAI (Grok) have increased the intelligence of their models with every iteration. It amazes me how fast we've become blasé about what AI can do now, versus what it could do only months ago. Maxim Lott wrote a great article showing how ChatGPT's latest reasoning model, o3, had an IQ of 116, putting it in the top 15% of humans. 11 months ago the best model had an IQ of less than 90. We went from asking LLMs basic questions to now asking it to evaluate medical results & suggest stocks to invest in - all in less than 3 years.

Will we see more massive leaps in intelligence in the next generation of AI/LLMs, moving towards Artificial Super-Intelligence (ASI)? Or will intelligence rise a bit more and then level-off? We might see those smaller gains in raw intelligence of single models, but see greater real-world impact when those models are built into agentic AI flows that implement best-practice.

Ultimately, AI intelligence is either going to increase exponentially and unlock ASI, or it will increase more gently (and potentially plateau). Either way, with the smartest current models being in the top 15% of people, it's safe to say that AI intelligence is high. But what is the cost of that intelligence?

AI costs will flatline or decrease

For many people, using an LLM/AI like ChatGPT can be completely free, at least for basic access. The UAE recently took things a step further and, through a partnership with OpenAI, will be giving all UAE residents free access to ChatGPT Plus.

Costs via API, ie for developers to build their own products, like my own voice AI and AI-email products, which use AI providers' models for their own purposes, have also decreased. GPT-4 went from $60/1M output tokens in March 2023 (with a small context window of only 8k tokens), to $30 with GPT-4 Turbo, with 128k context window! So halved the price and 16x performance in 8 months. GPT 4.1 took it further: $8/1M output tokens and 1,047,576 context window. There are also more expensive models, like OpenAI's o1 Pro, with $600/1M tokens, and also 'mini' models that are even cheaper, like GPT-4.1 Nano, at only $0.4/1M tokens.

AI providers, like all players in a market, will face competition, so it's fair to imagine AI prices in general will continue to decrease over time.

Couple rising AI intelligence with the lowering of its costs and you have a powerful force that is becoming cheaper & more accessible to use. But how will that intelligence actually be used?

Automation & AI Agency realise AI Commoditisation

AI, and LLMs, can do some pretty amazing things when you use them through their websites or apps; ask it to analyse your latest blood results and it does!

Increasing profits...initially

In my consulting work for a customer acquisition agency, I helped a strategic employee automate 3 tasks (pure robotic automation, not any added AI...yet). 1 task was invoicing clients, which she already did, so automation saved her time - around 3 hours/week. But the other 2 tasks were tasks she couldn't get around to; the automation allowed her to do more - sending field reports to 40 individual reps, and management reports to their respective managers - about 4 hours of work every week. Automation & AI are getting cheaper to use, and they can output more, valuable, work too.

This reflects the initial upslope of profits x AI/automation usage - companies can save costs by not hiring new employees and/or make more revenue by being able to offer more of their services - increasing profits.

AI Commoditisation Threshold

But, as more companies start to use more AI, there will be a growing expectation of lower prices for knowledge- and service-based work. Companies will not be able to charge higher prices as their clients know the work is being done by AI, which is only getting cheaper to use - "You're using AI for this, so why shouldn't we see lower prices?".

Added to this is the raft of new AI-first competitors; these are businesses with less than 10 people doing the work of 100, and are able to charge 30% - 50% less than incumbents, drawing revenues away from those incumbents. I think this will drive down revenues, and likely profits, especially for more traditional, non-AI, knowledge- and service-based businesses, like many professional services. Here are two examples:

• Market Research (I was an Insight Director at Kantar Media): AI automates previously expensive insights generation (maybe AI does 80% of the work for strategic insights, with a human looking over it, depending on the price), dramatically cutting prices clients are willing to pay, while costs (analyst salaries, UK NI tax, office rent) still rise due to inflation/debasement.
• Legal Services: AI rapidly commoditises contract review and compliance audits, lowering billable fees, but inflation pushes firm costs upwards, squeezing margins.

Work, and profits, move more and more towards leaner AI-enabled companies, which may not even be based in the same country as their new clients, using their location to add to their advantage.

Summary thus far

AI's increasing usefulness : cost, coupled with the improved business-focused abilities of AI and automation, and AI-first competition, will drive profits down for most traditional, non-AI, knowledge- and service-based businesses. That's the profit side of things; what about their business costs and monetary debasement ("inflation")?

Non-AI business costs rise, further squeezing profits

As in many things, change can initially be slow, and then increase rapidly. Some companies are already AI- and automation-first, others are implementing it more slowly, and (many) others have their heads in the sand.

Non-AI business costs rise

Whilst AI and automation costs will tend to decrease (or at least their usefulness : cost will increase) over time, non-AI costs will likely increase.

In service/knowledge-based professional services businesses, like market research, consulting, legal & accounting, media & creative, salaries are a major cost driver. Now, with more people likely churning out of employment due to AI, you might think that salary costs would decline quickly as there would be more labour supply, and likely less demand as more companies implement AI? Long-term, yes, entire roles will be removed (though new AI-related ones will undoubtedly arise - I see AI as a positive force of change; similar to the internet), and salaries for non-AI compatible roles will decline. But in the near-term transition, it will likely not be possible for businesses to meaningfully reduce salary costs, mostly due to political risk such moves would result in. Add to this, at least in the UK, more employment-related tax rises for businesses, like the recent increase in NICs.

So it's likely salaries for many roles will remain as they are, at least for now. Other non-AI business costs like office rental, marketing & supplies etc will likely continue to increase, as they 'always have', though possibly at an even faster rate - this is due to out-of-control monetary debasement, aka 'inflation'.

Monetary debasement ("inflation") increases all costs further

I remember, when I was about 10 years old, living in South Africa, where my mom would put in R100 of petrol at the petrol station. 1 litre of petrol cost R2.22. That same litre of petrol (ie not much about petrol has changed) now costs R21.24 - that's a 9.56x increase in the cost of petrol. Or said another way, that same amount of money is 9.56x less powerful now - purchasing power has been literally decimated in ~25 years. So that same R100, the same piece of paper with 'R100' on it, is now 9.56x less valuable than it was - it can now buy 10% of what it did. This is due, mostly, to monetary debasement by governments & 'central' banks. "But that's the way it's always been" - well, not exactly.

Monetary debasement occurs essentially when governments & central banks (ie that guy Dave, who you went to school with, is somehow in government and in charge of deciding what happens with our currency) decide to print more money. And they can print an "infinite" amount of fiat money without any effort/work being done - they just create it. You work harder and harder every year, but they just keep printing more - that's why you feel like you can't catch up. How big is this problem? Well, see how much money was printed during COVID.

Most central banks and governments choose a seemingly arbitrary percentage to debase currencies by every year; a 'target' of around 2%. This is enough for them to debase the currency by, and make you work harder, and just enough for you not to care. Indeed, the entire world watches to see what a few unelected people in the Fed and other central banks decide to do with interest rates, which effectively controls the money supply.

For businesses, they can see these price rises in increased salaries and other costs, "just something that happens every year". But as the above chart shows, monetary debasement isn't always "just 2%" - so businesses will continue to see profits getting squeezed even more as monetary debasement adds to increasing non-AI costs.

The paradox

So AI will likely, over time, lead to lower business profits. But non-AI costs, especially driven by debasement, are just going to keep rising, with no way for businesses to predict by how much.

This gives a paradox - with decreasing business revenues and increasing costs, what can a business do? The lines are moving in opposite directions?

There's a saying in Bitcoin, coined (ahem) by Lawrence Lepard in his book "The Big Print" - "Fix the money, fix the world"

And for businesses facing AI-related revenue & profit difficulties, Bitcoin might just help them too.

Bitcoin for businesses

TL;DR

Bitcoin can allow businesses to retain the purchasing power of the money they generate, and protect the value they've created from being eroded by fiat debasement

What is money, and what is 'hard money'

Seems like quite a silly and basic question, right? We all know what money is...right? Well, there's a lot more to it, especially if we want to compare things like gold, fiat currencies, and Bitcoin, and understand why a Bitcoin strategy can be important for businesses.

Bitcoin is considered to be the 'hardest' money because you cannot inflate it; no one can create more Bitcoin, when the ~21 million have been mined, that's it. And anyone can verify precisely how many Bitcoin have been mined, and when - a massive contrast to fiat currencies, when we're all watching the Fed to see what a few unelected people have decided to do.

The money supply denominator is fixed with Bitcoin

What is Bitcoin

I understood the value of Bitcoin when I sent my friend in Australia $10 worth of Bitcoin for a beer, whilst I was in South Africa - no bank to approve/reject the transaction, cost me $0.2 in transaction fees, and he got it 10 minutes later. Wow.

Check these videos out for great explanations on what Bitcoin in, the article continues below

1. 3 minutes

2. 37 minutes

3. 39 minutes

Fiat loses purchasing power by design

Let's look at the world's major fiat currencies vs Bitcoin; in other words, comparing fiat currencies to the hardest money we have - Bitcoin - this is fair because of Bitcoin's 'hard cap' finite supply - it makes sense to compare fiat against an unmoving denominator.

The below is the inverse of a normal price chart - we're all used to seeing the Bitcoin price moving up against fiat currencies; the inverse of that is all other currencies are trending down against Bitcoin; it losing their purchasing power. This has happened over the long term; if you zoom into a specific smaller timeframe you will see the ups-and-downs of a normal market, but those are insignificant over a long time-frame, making Bitcoin a great way to store the value your business is creating now.

Retaining purchasing power means the money you've made today can be used to buy the same, or more, things later.

This is key for businesses; you've made $50,000 excess profit this year, which can buy a new printing machine today. If you save that $50,000 in USD (ie keep it in 'cash' - in your business account with a traditional bank), it will be worth less every year, by design - that same machine will cost you more in the future, so you need to make more money to afford it later. But if you put that same $50,000 into harder money, like Bitcoin, it can appreciate against falling fiat money, preserving your purchasing power, and serve as a hedge against monetary debasement.

This is especially key for the largest of businesses. Apple has £35 billion in cash & cash equivalents as of March 2025. If inflation/debasement is 'only' 2.4% this year, it means Apple's purchasing power would have decreased by £840,000,000...in a year! And that happens every year (some years are better/worse) - so if I were them I'd be looking intently at how to preserve my purchasing power & shareholder value, as a responsible executive team should.

Bitcoin increases its purchasing power over the medium-to-long term

Because Bitcoin's supply is finite, and demand is growing, it tends to grow over the medium-long term. Let's look at a real world example of how Bitcoin's purchasing power has increased over the years.

A 3 bed terraced home in Oxford went from £320k in 2017 to £450k this year (+41%) - for the seller, great. For a buyer; they've had to spend 41% more of their hard-earned money for the same house. The house's intrinsic value hasn't gone up 41% - the value of your £ has declined.

In Bitcoin terms, you could buy one house for ₿94.3 in 2017, but in 2025 you could buy thirteen houses with ₿94.3, because Bitcoin's purchasing power grew

Bitcoin in business is gaining traction, used to buy/sell, and as a long-term treasury asset

Some businesses, especially those operating in countries that are more progressive, creative & generally open to change, have already started using Bitcoin both in day-to-day operations, and as a treasury asset.

Microsoft, AT&T, Apple, Starbucks, TAG Heuer & PayPal all accept Bitcoin in their businesses. Steak 'n Shake in the USA recently started accepting Bitcoin, through the Bitcoin Lightning network, saying it has been a 'boon for their business'.

So Bitcoin can, and is, being used for buying and selling products & services. It also tends to hold, and grow, its value over the longer term, making it a natural fit for a company's longer-term treasury reserves (ie not for operations, but to help secure and grow the company's longer-term stability, and protect against fiat erosion).

"But what about ETH, XRP etc, aren't they also cryptocurrencies?" - yes, sadly they are

Bitcoin-only; no shitcoins

Bitcoin is the first real cryptocurrency, created to address the problems inherit with centralised fiat banking and money systems. Other cryptocurrencies, like XRP and ETH, all have fundamental flaws that make them inappropriate for use in, and by, serious businesses.

Most other cryptocurrencies are highly centralised, with identifiable people who were given a large initial portion of the currency when it was created, and they still exercise significant control over the network - giving rise to risk. Most other cryptocurrencies also don't have a hard-cap, making them inflatable, just like fiat money.

This is why Bitcoin is superior - it is decentralised, hard-capped, censorship-resistant money that anyone can use and verify. And this is why serious businesses should only ever consider using Bitcoin in their businesses. It is also money powered by the internet, something it shares with AI.

AI & Bitcoin - Solving the Paradox

The Paradox - decreasing business revenues, largely due to AI, and increasing costs - what can a business do? My suggestion is to employ a phased Bitcoin strategy, allowing the company to hold progressively more Bitcoin vs fiat on its balance sheet and in its treasury, helping to offset rising costs and secure its long-term viability through Bitcoin's deflationary nature.

A phased & balanced Bitcoin strategy

As in many things, a considered approach is better than going all-in immediately. Businesses should understand their long-term goals & see how Bitcoin can help them achieve those. Businesses could start offering their services in fiat and Bitcoin, allowing them to accumulate Bitcoin directly, which could be used for more operational purposes; covering shorter-term costs. They could also convert x amount of cash generated, or existing treasury assets, into Bitcoin every month, slowly growing their holdings towards their desired mix of Bitcoin and other assets like gold, blue-chip stocks and cash.

Selling some Bitcoin to cover certain costs, and CGT

Treasury assets are generally meant to bolster the financial foundations of a business, they're not meant to cover operational costs. To this end, if dipping into the treasury is necessary, the business could sell the 'softest' assets first, only selling Bitcoin as a last resort, as, once it's sold, you might never get it back at that price again.

In the UK, currently, as in many other markets, selling Bitcoin triggers capital gains tax. This is onerous. This could indeed be a stumbling block for UK companies looking to adopt a Bitcoin strategy, especially to cover any shorter-term operations costs. But UK companies like The Smarter Web Company are adopting a Bitcoin strategy regardless.

Bitcoin and AI in business

In our always-changing world, a Bitcoin strategy may help businesses build longer-term financial stability, helping combat both the deflationary effects of AI, and the relentless debasement & devaluation of fiat currencies.

Using Bitcoin in business for both buying & selling will provide more choice for Bitcoiners to pay in, and even reduce credit card payment fees, as happened for Steak 'n Shake. And using it for corporate treasuries can help bolster a company's longer-term stability, especially if they, like many/most companies, are still crossing the AI Rubicon.

#AI #Automation #Bitcoin