The Tailor at Zara Prices

#157 - Reflections on investing and life from building my holding company

Dear reader,

At 20, I used to walk into Marks & Spencer and look at their clothing section with horror. Who buys this stuff? Shapeless jumpers, dull colours, cuts designed to hide any human form. Dad clothes. Clothes for people who’ve given up.

At 40, I have to admit: M&S is kind of compelling. Honest quality, fair price, zero pretence. Won’t change your life, but does the job.

Of course, if I could have a bespoke suit tailored to my body at M&S prices, I wouldn’t hesitate. That’s always been the trade-off: pay for the tailor, or accept the compromise. - In tech, we’ve just arrived here. And almost nobody’s noticed.

The first shift

The biggest shift in startup world was from custom-made to fast fashion.

Before SaaS, software was bespoke. Expensive, slow, built for you. Enterprise projects costing hundreds of thousands, taking months, requiring dedicated teams. The digital tailor.

Then came Salesforce, HubSpot, Notion, Slack. The Zara of tech. They standardised everything, slashed prices, democratised access. Good enough for everyone, which means perfect for no one. But the price was right and the speed unbeatable.

This shift paid massively. A whole generation of VCs and founders got rich on it.

The second shift

Now AI introduces a new paradigm: bespoke at Zara prices. Custom-made software, built by a capable tailor, at £29.90. If you think we’re not there yet, you’re wrong.

My email, WhatsApp, calendar, CRM, to-do list, notes - all integrated into one solution that Claude built for me. I didn’t buy a product. I didn’t subscribe to anything. I described how I work, what I need, my patterns - and got a system that adapts to me instead of the other way around.

This paradigm shift is still underappreciated. But it’s happening, it’s accelerating, and in my view it’s inevitable.

Who makes money in this new world

Three categories.

1. Picks and shovels

The classic gold rush play. Data centre suppliers, electricity providers, chip makers, contractors building the infrastructure. They don’t care who wins the AI race. They sell to everyone. This always works.

2. Information asymmetry arbitrageurs

People who exploit the gap between the world as it is and the world as it was.

I have a friend. Two people, €500k revenue. With AI, they do growth hacking like they’re a team of 20. Is it AGI-proof? No. Will it last forever? Probably not. But right now, they’re printing money while everyone else philosophises about what AI means.

The window won’t stay open. But while it’s open, it’s very open.

3. Professional services enhanced by AI

This one’s not making big money yet. But I believe it has potential.

Accountants. Lawyers. Event organisers. Consultants. People who close the feedback loop between AI and the outside world.

The problem with these businesses was always the ceiling. A good accountant can only handle so many clients. Want to grow? Hire more accountants. But each new hire brings complexity, fixed costs, quality risk. Linear growth is a trap.

AI breaks the ceiling. Same structure, 10x the clients. The repetitive work - and in accounting, 70% is repetitive - gets done at zero marginal cost. The freed-up time goes to high-value advice, client relationships, strategy.

A “non-scalable” business becomes suddenly very scalable.

Where should VCs invest?

Picks and shovels - yes. Deep tech VCs backing infrastructure, private equity investing in the data centre value chain. Not trivial, but the logic holds. Reasonable match.

Information asymmetry plays - I don’t see how these are fundable. And let me be clear: I’d put 80% of today’s hot startups in this bucket. Lovable, Fyxer, the current darlings. They’re playing the old B2B SaaS playbook in a world where that playbook is expiring. They might show flashy growth, but the asymmetry will close. I don’t think they’re worth much. At least that’s my view.

Professional services 2.0 - this is where it gets interesting. These might be worth backing. But here’s the paradox: if AI makes everything an order of magnitude more capital efficient, how much capital should founders actually raise? Can the VC model even work for this strategy?

I’m not sure it can.