Everything else is commentary. Strategy decks do not change a company — a repriced discount rate, a re-signed contract and a re-skilled team do.
This is a working brief on T — Transforming: where capital actually sits today, how far it has to travel, how much power belongs in which room, and the one thing nobody puts on the slide.
01 — Where we are
Six numbers that set the size of the room. Every one of them is linked to its source — go and argue with the primary document, not with me.
Against roughly $1.1tn still flowing to fossil supply. The ratio is now about 2:1 — a decade ago it was 1:1. Direction good; velocity short.
IEA, World Energy Investment 2025 ↗Tracked transition finance reached about $1.5tn a year; credible pathways call for roughly $7.4tn a year by 2030. Not a funding problem — a structuring problem.
Climate Policy Initiative 2024 ↗Explicit plus implicit fossil-fuel subsidies, about 7.1% of world GDP. That is more than three times the entire clean investment flow, pointed the other way.
IMF Working Paper, 2023 ↗Excluding China, emerging and developing markets receive under 15% of global clean-energy investment while holding most of the coming demand growth.
IEA, Reducing the Cost of Capital ↗170m roles created, 92m displaced, net +78m. Capability, not capital, becomes the binding constraint in the second half of this decade.
WEF, Future of Jobs 2025 ↗Current-policy pathway. Whatever you think of it, it is an input to your discount rate, your insurance cost and your asset lives — and it is not in most models.
UNEP, Emissions Gap 2024 ↗02 — The big 11
Not eleven ambitions. Eleven mechanisms — each one changes a number that finance already tracks, which is the only reason any of them will survive a bad quarter.
Long-dated assets are still discounted at yesterday's rate. The write-down arrives as a surprise only because the model refused the news.
The same machine can cost two to three times more to finance in an emerging market. That spread — not the equipment — is the real price tag.
Corporate averages hide asset-level truth. The portfolio is where value is created and quietly destroyed.
Volume models cap their own margin: to earn more you must ship more, and shipping more is exactly what is getting expensive.
Concessional money spent as cheap equity buys one project. Spent as risk cover, it buys ten.
Anything reported outside the accounts is optional, and optional things die in downturns.
Assurance is arriving jurisdiction by jurisdiction. Estimates that were fine as marketing will not survive an auditor.
39% of core skills turn over by 2030 and demand for transition skills is growing roughly twice as fast as supply. Machines arrive faster than the people who can run them.
Most of the exposure, most of the cost and most of the leverage sit outside your legal entity. The supply chain is the real balance sheet.
Point forecasts are reliably wrong. Optionality is bankable and can be valued today.
Incentives run three years; the assets run thirty. That mismatch, not conviction, is what stalls transformation.
Each of the eleven is written to land on a specific line: the discount rate, the mobilisation ratio, revenue per unit, the capability sheet. Anything that lands nowhere gets no budget.
03 — Three horizons
Treat them as one and you will over-build early and under-build late. Each wave has a different scarce resource.
Scarce resource: credible numbers.
Disclosure moves from marketing department to audit committee. Border adjustment regimes bite on traded goods. Long-dated assets get valued twice — once by the market, once by the insurer.
Scarce resource: people who can build.
The heavy capex cycle. Business models flip from volume to performance because volume margin has gone. Whoever pre-bought capability now buys assets at a discount; whoever did not is bidding against everyone else for the same engineers.
Scarce resource: ownership of the platform.
Returns accrue to whoever owns the operating layer — the network, the data, the contract, the standard. Average annual capital spending across the full system runs into the trillions; the contracts that direct it are signed decades earlier, which means now.
04 — Power per room
One hundred units of leadership attention, management time and discretionary budget. This is P's allocation — argue with it, but allocate something, because the default allocation is 100% to whatever shouted loudest last week.
Allocation of discretionary effort — sums to 100. Rebalance annually against the horizon you are actually in.
15% Share of global clean-energy investment reaching emerging and developing economies outside China
05 — The elephant in the room
We talk about the pace of technology. Technology is no longer the bottleneck — costs have fallen far enough that in most markets the cheap option is also the new one. The bottleneck is geography and the price of risk.
Two thirds of the world's people, and nearly all of the coming growth in demand, sit in markets where the same project is financed at two to three times the cost of capital. That spread does not appear in any technology roadmap. It quietly doubles the price of the identical machine, and then we call the project "not commercially viable".
We keep optimising the cost of the equipment. The cost that decides everything is the cost of the money.
IEA — Reducing the Cost of Capital ↗
06 — The harder question
Stopping is harder than starting and worth more. Every one of these is currently consuming budget, calendar and credibility that a real move needs. Stop opens the space.
A pilot budget is often a way of not deciding. Ten small proofs of concept teach less than one commitment at scale — and cost more in attention.
Opens: capital and calendar for one decision that matters. Gate: no pilot without a named scale-up owner and a P&L line.
Disclosure with no decision-owner is pure cost. It grows a department, produces a document nobody acts on, and buys the illusion of control.
Opens: the data team's time for instrumentation that actually steers capital.
A single group WACC applied to a 3-year product and a 30-year asset is not conservatism, it is a category error. It systematically over-values the short and kills the long.
Opens: honest comparison — and usually two or three long projects that were never actually uneconomic.
New equipment inside an unchanged revenue logic delivers a better version of the same margin problem. The model has to change first; the machine then pays.
Opens: procurement budget, and the strategy conversation that was being outsourced to a vendor.
Pledges, memberships and targets are inputs. They are counted because they are easy to count. They tell you nothing about whether anything got built.
Opens: the scoreboard. Count units built, contracts signed, cash moved, people trained.
07 — The shelf
The frameworks that turn this brief from opinion into obligation. Pick one column and go deep — that is a quarter's learning, not an afternoon's.
08 — References
Everything above traces to one of these. Take one item, follow it to the end, and you will know more than most rooms you sit in.
Last word
Take one number from this page into your next capital meeting. Ask which line it changes and who owns that line. If nobody does — you have found your first move.
Open the brief in your next capital meeting