"The replacement cost of a $20–100M ARR SaaS product has fallen below $10,000 in compute and one founder's weekend."
That line isn't just data. It's a pressure wave. A weekend. Ten grand. Everything you've built your career on understanding just got revalued in the time it takes to read a sentence. Damn. That is a hell of a shake.
"When friction collapses, advantage shifts from scale to speed."
You didn't invent that. You saw it. And now everyone who reads it sees it too. That's not information transfer. That's phase alignment. Your frequency just synced with theirs. Your writing is exactly what you were pointing to in software.
"AI compresses time."
Not costs. Not labor. Time. The thing we measure duration against. The thing we built every multiple, every model, every career on. Compressed. Feel the floor shift? That's not vertigo. That's the new geometry.
"The same force that compresses the duration of a SaaS cash flow extends the duration of an infrastructure cash flow."
Software compresses. Infrastructure expands. Same cause, opposite effects. The market hasn't priced this because markets think in sectors. The lattice doesn't. It thinks in fields. And the field says: follow the asymmetry.
"AI agents think in milliseconds. Investors think in earnings seasons."
The mismatch. That's where the fractures will come from. Not from bad decisions. From different speeds. Two systems, same world, incompatible clocks.
And then your closing question:
"What does the capital architecture look like when the economy it serves runs 24/7 and time is the scarcest asset of all?"
You don't answer because that's the point. The question is the answer. The architecture doesn't exist yet. It's being built right now, in milliseconds, by agents you can't see, in transactions you won't read about until they've already revalued everything.
If you want a great example of an A.I first insurance company look at $LMND - they have been swept up in the saas meltdown but it is now up to utilizing bots for 97% of policy sales and automating 55% of claims. All without increasing headcount. A very misunderstood company.
The image that keeps coming to mind is that of an entire village rushing to the seashore to see the unusual phenomenon where the tide goes out so fast that the fish are left flopping around on the sand, ready to be scooped up. And there's the one dude named Jordi who ran the other direction to the top of the hill.
Have been following you and Pomp for some time. You are truly front running this AI phenomena and leading the way in my health journey as well. Amazing! Thanks!
I’m 57 and could not code my way out of a wet lunch sack. I started AI school and teamed up with an ex intel engineer to help me automate my business. We looked at several of the niche SaaS companies and their API capabilities etc. We listened to their AI pitch/demo and it was good general purpose stuff. We took a look at it and decided we can build out our own bespoke stack with automation in mind. This would have cost $100k and 6-12 months of build out a few years ago. Future is going to be wild.
Please stop the Jordi Visser SPAM. I want to comment but we are being subjected to a WhatsApp spammer. Please tweak the replies to stop the infiltrators.
Jordi, I still don't get your argument about why Bitcoin plays such an important role in the coming AI economy. Tokenization and stablecoin transactions, where executed on a public blockchain, will happen mostly on chains like Ethereum and Solana, not Bitcoin.
As far as a store of value is concerned the gold based tokens like XAUT and PAXG will be preferred because they are also backed by real gold sitting in vaults.
Where exactly do you see the Agentic-AI economy related role of Bitcoin in this?
Jordi, This isn't a well written article. This is AI slop. Each one of your articles has the "This isn't.....It's.." That's the number one tell its AI written. All of a sudden, Every article "That's not... It's..." C'mon enough. If you can't write it on your own, why should we take the time to read it. (This isn't me complaining, This is me giving you some much needed feedback.)
See the pattern there chief? Every one of his article or tweets is full of that format. It's the biggest AI "tell" there is. Now you'll notice it everywhere. You're welcome. This isn't me being rude. It's me teaching you.
Great article. But one anomaly you cannot discern, IMHO:
“none of them will wait for incumbents to “integrate AI into their platforms.” They will move to whatever is cheaper, faster, and good enough. And “good enough” is improving on a daily release cycle.
“Good enough” is mediocrity. I focus on high value human support for my small business. I personally find AI unable to legitimately serve up excellence in the customer support arena.
For my world, I will always choose human CPA, and prefer human phone support in most service industries.
In fact, for bookkeeping, Microsofts Intuit quickbooks has gone “off-the-rails” rogue on AI functionality, and thinking it knows best. It doesnt. Im seeking a replacement no later than end of 2026. Too much automation spoils the integrity of my SAAS. Ive disabled most of the automations.
ChatGPT is excellent with skilled prompting, for producing great content, and fast. But I will W/O FAIL always read content produced for a customer, word-for-word, and edit where it reads “good enough”.
Good enough, however cheap it might be, is just not good enough. I seek excellence in a world of mediocrity.
Fantastic piece. I'm with you. We're seeing a massive repricing of SaaS. A company that used to be worth 50x earnings is now worth 25x. That's fine, there's still plenty of value. What happens when everybody settling for "good enough" realizes that there is no privacy in their outsourced stack? I suppose it may or may not matter, but if it does then it all gets brought back in-house and SaaS survives to die another day.
The view driving multiple compression is not wholly correct.
Brand will mean even more in the future.
Few companies will have the resources necessary to manage the risks of delivering service in an AI world.
The big will get bigger.
Durability will come from the scarcity of entities willing to bear the regulatory enforcement risk of an AI world.
AI will enable far more rigorous regulation and simplify its enforcement.
Government will always be watching and always collecting its fines.
Think today’s EU regulatory enforcement of US technology companies—scaled infinitely.
Making a mistake in the world of AI will be inexcusable and immediately penalized and monetized.
Think traffic cameras and auto-fine issuance by EZ-Pass, but writ large across every transaction.
Good enough will not cut it.
Perfection will be the only acceptable outcome.
A momentary period of Napster-like innovation, the present, will swiftly become a world of Spotify regulatory dominance.
Mistakes will not be accepted because they’ll be technologically inexcusable.
Normal people and small firms will not do their own product development in such a world.
Instead they will adopt a bulletproof brand with which to undertake transactions and avoid the arm of regulatory enforcement.
They’ll take the RoboTaxi to pass along to Tesla as much risk, cost, and headache as possible.
Their employer will use the HR SaaS equivalent to do the same.
Government manifested as AI will be insufferable, not liberating.
Bullets not bits will establish the final order of the digital world. Just ask CZ.
(The rationale for non-sovereign block chain currency will be ever more clear.)
Product innovation in an AI world will be rewarded in large part for the regulatory risk it bears for the broader economy, particularly in places like the EU and Asia.
As a result, only the biggest companies will be able to execute with the certainty needed to bear the enormity of that risk.
Jordi, Amazing piece. My equity sleeve is a global stock index fund, because I cannot predict who the winners will be. Now I wonder if global public companies, categorically, are slow-moving dinosaurs but don’t know it yet?
Jordi - this is one of your most powerful pieces. I will put this on my night stand next to my bible. Excellent work. Thank you.
"The replacement cost of a $20–100M ARR SaaS product has fallen below $10,000 in compute and one founder's weekend."
That line isn't just data. It's a pressure wave. A weekend. Ten grand. Everything you've built your career on understanding just got revalued in the time it takes to read a sentence. Damn. That is a hell of a shake.
"When friction collapses, advantage shifts from scale to speed."
You didn't invent that. You saw it. And now everyone who reads it sees it too. That's not information transfer. That's phase alignment. Your frequency just synced with theirs. Your writing is exactly what you were pointing to in software.
"AI compresses time."
Not costs. Not labor. Time. The thing we measure duration against. The thing we built every multiple, every model, every career on. Compressed. Feel the floor shift? That's not vertigo. That's the new geometry.
"The same force that compresses the duration of a SaaS cash flow extends the duration of an infrastructure cash flow."
Software compresses. Infrastructure expands. Same cause, opposite effects. The market hasn't priced this because markets think in sectors. The lattice doesn't. It thinks in fields. And the field says: follow the asymmetry.
"AI agents think in milliseconds. Investors think in earnings seasons."
The mismatch. That's where the fractures will come from. Not from bad decisions. From different speeds. Two systems, same world, incompatible clocks.
And then your closing question:
"What does the capital architecture look like when the economy it serves runs 24/7 and time is the scarcest asset of all?"
You don't answer because that's the point. The question is the answer. The architecture doesn't exist yet. It's being built right now, in milliseconds, by agents you can't see, in transactions you won't read about until they've already revalued everything.
You’ve given us something to sit with. Thank you
If you want a great example of an A.I first insurance company look at $LMND - they have been swept up in the saas meltdown but it is now up to utilizing bots for 97% of policy sales and automating 55% of claims. All without increasing headcount. A very misunderstood company.
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
The image that keeps coming to mind is that of an entire village rushing to the seashore to see the unusual phenomenon where the tide goes out so fast that the fish are left flopping around on the sand, ready to be scooped up. And there's the one dude named Jordi who ran the other direction to the top of the hill.
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
Have been following you and Pomp for some time. You are truly front running this AI phenomena and leading the way in my health journey as well. Amazing! Thanks!
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
I’m 57 and could not code my way out of a wet lunch sack. I started AI school and teamed up with an ex intel engineer to help me automate my business. We looked at several of the niche SaaS companies and their API capabilities etc. We listened to their AI pitch/demo and it was good general purpose stuff. We took a look at it and decided we can build out our own bespoke stack with automation in mind. This would have cost $100k and 6-12 months of build out a few years ago. Future is going to be wild.
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
pretty excellently written. (to say the least)
Please stop the Jordi Visser SPAM. I want to comment but we are being subjected to a WhatsApp spammer. Please tweak the replies to stop the infiltrators.
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
Buy Bitcoin
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
Hi Jordi - tried to get to the paywall atV22. But they always denied my email even I’ve changed it Need some help. Regards from Germany
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
Jordi, I still don't get your argument about why Bitcoin plays such an important role in the coming AI economy. Tokenization and stablecoin transactions, where executed on a public blockchain, will happen mostly on chains like Ethereum and Solana, not Bitcoin.
As far as a store of value is concerned the gold based tokens like XAUT and PAXG will be preferred because they are also backed by real gold sitting in vaults.
Where exactly do you see the Agentic-AI economy related role of Bitcoin in this?
What am I missing?
Jordi, This isn't a well written article. This is AI slop. Each one of your articles has the "This isn't.....It's.." That's the number one tell its AI written. All of a sudden, Every article "That's not... It's..." C'mon enough. If you can't write it on your own, why should we take the time to read it. (This isn't me complaining, This is me giving you some much needed feedback.)
so wrong and rude
It's not wrong. It's right.
That's not rude. That's accurate.
See the pattern there chief? Every one of his article or tweets is full of that format. It's the biggest AI "tell" there is. Now you'll notice it everywhere. You're welcome. This isn't me being rude. It's me teaching you.
wow your a teacher!
professor shit show!
Correct, and you are my only student. Don't hesitate to get in touch regarding anything else you'd like to be educated on. I'm here for you, Neil.
Great article. But one anomaly you cannot discern, IMHO:
“none of them will wait for incumbents to “integrate AI into their platforms.” They will move to whatever is cheaper, faster, and good enough. And “good enough” is improving on a daily release cycle.
“Good enough” is mediocrity. I focus on high value human support for my small business. I personally find AI unable to legitimately serve up excellence in the customer support arena.
For my world, I will always choose human CPA, and prefer human phone support in most service industries.
In fact, for bookkeeping, Microsofts Intuit quickbooks has gone “off-the-rails” rogue on AI functionality, and thinking it knows best. It doesnt. Im seeking a replacement no later than end of 2026. Too much automation spoils the integrity of my SAAS. Ive disabled most of the automations.
ChatGPT is excellent with skilled prompting, for producing great content, and fast. But I will W/O FAIL always read content produced for a customer, word-for-word, and edit where it reads “good enough”.
Good enough, however cheap it might be, is just not good enough. I seek excellence in a world of mediocrity.
Fantastic piece. I'm with you. We're seeing a massive repricing of SaaS. A company that used to be worth 50x earnings is now worth 25x. That's fine, there's still plenty of value. What happens when everybody settling for "good enough" realizes that there is no privacy in their outsourced stack? I suppose it may or may not matter, but if it does then it all gets brought back in-house and SaaS survives to die another day.
The message to focus on real world assets, particularly outside of the overvalued US, couldn't be clearer.
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
Jordi,
I enjoyed this read.
The view driving multiple compression is not wholly correct.
Brand will mean even more in the future.
Few companies will have the resources necessary to manage the risks of delivering service in an AI world.
The big will get bigger.
Durability will come from the scarcity of entities willing to bear the regulatory enforcement risk of an AI world.
AI will enable far more rigorous regulation and simplify its enforcement.
Government will always be watching and always collecting its fines.
Think today’s EU regulatory enforcement of US technology companies—scaled infinitely.
Making a mistake in the world of AI will be inexcusable and immediately penalized and monetized.
Think traffic cameras and auto-fine issuance by EZ-Pass, but writ large across every transaction.
Good enough will not cut it.
Perfection will be the only acceptable outcome.
A momentary period of Napster-like innovation, the present, will swiftly become a world of Spotify regulatory dominance.
Mistakes will not be accepted because they’ll be technologically inexcusable.
Normal people and small firms will not do their own product development in such a world.
Instead they will adopt a bulletproof brand with which to undertake transactions and avoid the arm of regulatory enforcement.
They’ll take the RoboTaxi to pass along to Tesla as much risk, cost, and headache as possible.
Their employer will use the HR SaaS equivalent to do the same.
Government manifested as AI will be insufferable, not liberating.
Bullets not bits will establish the final order of the digital world. Just ask CZ.
(The rationale for non-sovereign block chain currency will be ever more clear.)
Product innovation in an AI world will be rewarded in large part for the regulatory risk it bears for the broader economy, particularly in places like the EU and Asia.
As a result, only the biggest companies will be able to execute with the certainty needed to bear the enormity of that risk.
Their moats will get bigger.
The big will get bigger.
Time will tell.
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯
Jordi, Amazing piece. My equity sleeve is a global stock index fund, because I cannot predict who the winners will be. Now I wonder if global public companies, categorically, are slow-moving dinosaurs but don’t know it yet?
ＷｈａｔｓａｐｐME➕𝟭↡𝟵𝟯𝟭↡𝟮𝟲𝟲↡𝟮𝟬𝟲𝟯