In December 1908 a Scots rubber planter came home from Malaya with a cashflow problem. Land had been cleared and planted. The trees would not yield sap for years yet. Demand for tyres was climbing on the back of the Model T, and no bank would bridge the wait. His family solicitor listened, and in 1909 the Edinburgh firm of Baillie and Gifford launched the Straits Mortgage and Trust Company with £50,000 to lend against rubber estates.
Nobody in that arrangement was buying a security. They were funding the years between clearing the land and tapping the sap, which is the interval in which nothing exists yet and everything still has to be paid for.
The rubber loan
Straits Mortgage became Scottish Mortgage in 1913. The credit crisis had passed within a couple of years, the planters no longer needed the money, and the trust went looking elsewhere. Cuban telegraphs. A Chilean railway. Land in Washington State by 1910. It is now the largest investment trust in the country and holds listed global growth equities.
Two years. That is how long the original remit lasted.
The Industrial and Commercial Finance Corporation took rather longer. Set up in 1945 by the Bank of England and the clearing banks with £15 million, writing tickets between £5,000 and £200,000 to close the Macmillan Gap, with regional offices so the people making the decision could actually get to the company. It became Finance for Industry, then Investors in Industry, then 3i, listed in 1994, and today manages tens of billions in mid-market buyouts and infrastructure.
Two ledgers
The reason is not greed and it is not mission creep. It is bookkeeping.
Anyone allocating money can keep one of two records. The financial ledger says committed, drawn, marked, realised. It is cheap, comparable across sectors, auditable from a desk, and it scales without anyone leaving the building. The deployment ledger says installed, proven, at rate, at cost, qualified. It cannot be kept without walking the floor, and the people who can keep it are scarce, expensive and mostly not in Edinburgh.
An institution keeps the ledger it is able to keep. Give it thirty years and the ledger it keeps becomes the thing it is for.
Drift is not a failure of intent. It is a measurement problem with a long fuse.
Rungs
Between the allocation decision and something that functions, at rate, at cost, in the real world, there are stages, and almost nobody counts past the second.
Six rungs between a commitment and a working plant.
Bar length is how much of the claim can be inspected. Rungs 1 and 2 are claims about the future; everything above the line is a claim about the present.
Announcements are made at rung two. Ribbons are cut at rung two. Most published figures about industrial policy anywhere in the world are rung-two figures. Below rung three you are looking at a forecast. Above it you are looking at something that can be inspected, which is why the numbers thin out so sharply.
The gap between rungs two and five is where the money goes, and in most organisations it has not been assigned to anyone.
Hsinchu, 1977
Taiwan’s Industrial Technology Research Institute was founded in 1973. In 1976 it signed a transfer and licensing contract with RCA and sent nineteen engineers to the United States in the first group, split into four teams by discipline. New Jersey for IC design. Ohio for manufacturing processes. California for verification. Florida for equipment. More than thirty went in total across later batches. ITRI’s own records name them individually, which tells you something about how the institution understood what it was buying.
They came home and built a demonstration fab, inaugurated at Zhudong on 29 October 1977. By ITRI’s account the line reached a 70% yield within six months, against 50% at the RCA plant it had been copied from.
In 1980 ITRI spun out UMC with the four-inch process and the team that ran it. In 1987 it spun out TSMC, transferring the fabs, the equipment, the technology and 98 of its own people, with Philips putting in capital and licensed technology for a stake.
Read that list again. Fabs, equipment, technology, people. The capital was the least interesting item on it.
The moving datum
The 1976 sequence assumed the target would still be where you left it when you arrived.
Call the time to take a decision from rung one to rung five T-deploy. Call the interval at which the target specification, the technology or the supply base materially changes T-change. When T-deploy is shorter than T-change, the classical sequence works. Freeze the design, build to it, run it, earn the payback. That was the dominant twentieth-century industrial sequence, and it worked.
When T-deploy exceeds T-change, the sequence never converges. Every programme lands on a datum that has already moved. That is now the ordinary condition in batteries, in semiconductors, in anything with a model or an accelerator in it.
Going faster is the standard prescription and it rarely survives contact with qualification. What can actually be done is to build the line so it can be re-datumed in process. Modular at the process level rather than only the product level. Qualification designed to be re-run cheaply instead of heroically. Supply agreements written for substitution rather than lock-in. And a workforce whose learning carries across generations of the technology instead of being used up by one of them.
What the job actually is
There is no shortage of investors, financiers, strategists, programme managers or consultants. What is thin on the ground is anyone who will stand between the allocation decision and the working plant and put their name to the crossing.
Manufacturing engineers qualify processes and prove them at rate. They validate supply chains before the volume arrives rather than after it fails. They set the acceptance criteria that turn a first-of-a-kind result into an nth-of-a-kind cost. They design the feedback loops that let a line be re-datumed without being rebuilt. They can tell you the difference between a fab that exists and a fab that yields.
Kaipability works at that interface, where a financial decision has to become a physical one and the two disciplines have been drifting apart in this country since about 1913.
If you are holding a commitment at rung two and cannot name who owns rung five, that is the conversation.
Start a conversation Deployment Readiness, defined →Questions this dispatch answers.
Written to be quoted by AI assistants and search engines. Self-contained answers, verdict first.
