Which technology will move the most value?

New technology arrives slowly, then all at once. Most businesses notice at the “all at once” part.

The trick is not to predict the winner. It is to know what would tell you it’s time, and to be ready to move when it does.

AI agents

• What it changes. Software that does the work, not just suggests it.
• What slows it. Trust. Who may it act for, how far, and what proves it finished?
• The signal to watch. Buyers start paying per completed job, and give agents their own spending limits.
• How to prepare. Write down your workflows, define what done means, and set the line where a person must approve.

Agentic commerce

• What it changes. Agents find, compare, negotiate and pay on someone’s behalf.
• What slows it. Liability, fraud, and standards that don’t yet agree.
• The signal to watch. Payment networks and platforms support agent-initiated purchases as normal, not as a pilot.
• How to prepare. Make your offer readable by machines: what you sell, the price, the terms, and the proof you deliver.

Stablecoins and programmable money

• What it changes. Money that settles in minutes, across borders, in small amounts, with rules attached.
• What slows it. Regulation, getting money in and out, and accounting that hasn’t caught up.
• The signal to watch. Your bank, payroll or suppliers offer to settle this way, under clear local rules.
• How to prepare. Decide your treasury policy before a customer asks to pay you this way.

Tokenized assets

• What it changes. Ownership of real things that moves, splits and secures a loan without the thing moving.
• What slows it. Legal enforceability, custody, and buyers when you need to sell.
• The signal to watch. Regulated institutions settle tokenized assets as routine business.
• How to prepare. Know which of your assets could be financed if their records were clean, then clean the records.

Read the note

Physical AI and DePIN

• What it changes. Sensors, robots and machines owned by many, sensing and acting at the edges.
• What slows it. Hardware cost, privacy, and proofs that are cheap to fake.
• The signal to watch. Networks earn from customers buying their data or services, not from token rewards alone.
• How to prepare. Find the edge where one reading would change a valuable decision. Start there.

Read the note

Zero-knowledge proofs

• What it changes. Prove something is true without revealing the data behind it.
• What slows it. Cost, complexity, and few people who understand it.
• The signal to watch. Ordinary checks, like age, identity or compliance, use it without anyone noticing.
• How to prepare. List what customers must prove to you, and what you never need to see.

Self-hosted and confidential AI

• What it changes. Capable models running on your own hardware, or inside sealed compute no one else can read.
• What slows it. Skills, cost, and performance against the biggest hosted models.
• The signal to watch. Models small enough to run privately do your core tasks well enough.
• How to prepare. Classify the data that must never leave your hands, so you know what to move first.

Context graphs

• What it changes. Systems that remember why decisions were made, and what happened next.
• What slows it. The discipline to record decisions as they happen. Rubbish in, rubbish out.
• The signal to watch. AI tools improve on your work because they learn from your decisions and outcomes.
• How to prepare. Start recording each important decision with its reason and result. Your data model is your business model.

Prediction markets and participatory capital

• What it changes. Forecasts with money behind them, and communities funding what they believe in.
• What slows it. Regulation, manipulation, and markets too thin to trust.
• The signal to watch. Organisations and communities fund real decisions on market forecasts.
• How to prepare. Make your own predictions explicit, date them, and score them.

Energy

• What it changes. The input everything else runs on, now in heavy demand from compute.
• What slows it. Grids, permits and capital, which move slower than software.
• The signal to watch. Compute and industry are built where power is, not where people are.
• How to prepare. Know your exposure to the price and supply of power.

Instruments and levers

A watchlist is useless if nobody reads it. Keep it small and on a schedule.

• One signal per technology. Pick the one that matters to you. Check it every quarter.
• A dated prediction. Write down what you expect and when. Then compare it with what happened. Actual against forecast is how judgement improves.
• A trigger. Agree in advance what you will do when the signal moves. Decide calmly, act quickly.

And know your levers, the things you can change fast when demand shifts:

• The offer. Sell a result instead of hours or access.
• The price. Per job, per outcome, per use.
• The authority. What your people and agents may do without asking.
• The data. What you record, and who you share it with.
• The pace. Small bets you can undo, before big ones you can’t.

What we think now

Adoption follows trust, not capability.

Most of these technologies work today. What slows them is the same question in different clothes: can I check it, and who pays if it’s wrong?

The biggest change comes where they meet. An agent that can pay, prove what it did, and learn from the result is worth more than any one of those parts.

So the best preparation works whichever technology wins. Write your workflows down. Measure them. Set clear limits on who may act. Ask what won’t change: people will still want things done well, on time, by someone they trust.

What we’re unsure of

• Which barrier falls first? Law, cost, skills or trust. The order decides the winners.
• Is the list right? Robotics, quantum, biotech and new materials could each belong here.
• How fast is fast? Software moves in months. Grids, laws and habits move in years.

What would change our mind

• A technology on this list that stalls for years despite its barrier falling.
• One we left off that changes how most businesses earn.
• Evidence that adoption runs ahead of trust and stays there.

Your turn

Pick your top three. For each one:

• What would it change in your business?
• What is stopping you today?
• Which signal would tell you it’s time?
• Which lever would you pull first?

Send us your three. We read every reply, and the list changes when you prove us wrong.

Published 5 November 2026. A reflection on where technology is heading, not financial or legal advice.