When Carillion was liquidated in January 2018 it was not just one big failure — it was a detonation that ran down the supply chain. The most instructive part is how a single insolvency forced hundreds of smaller, otherwise-viable companies under. GalimAI’s data is built to see exactly this kind of linked exposure: who is connected to whom, and where one company’s strain becomes another’s.
What GalimAI’s own data reveals
A failure like Carillion does its damage through connections. GalimAI maps 463,022 property-owning companies and ties each to its named directors, its charges and its distress markers — the raw material for seeing how strain travels. The companies that failed in 2018 were rarely the weakest on paper; they were the ones over-exposed to a single counterparty that stopped paying.
That is what makes the GalimAI map useful beyond a headcount. Our analyses of stacked distress signals, accelerating dissolutions and the regional distress map turn linked, second-order risk into something visible — the kind of concentration that turns one big insolvency into many small ones.
What happened: the Carillion collapse, in plain terms
Carillion entered compulsory liquidation on 15 January 2018 owing close to £7bn against just £29m of cash. As a major government contractor it sat at the top of a vast supply chain — more than 30,000 suppliers and subcontractors were exposed to its failure.
The contagion was immediate. Construction insolvencies in the first quarter of 2018 jumped to around 780, up a fifth on the same period in 2017, with specialist subcontractors hit hardest. Total construction insolvencies for 2018 rose about 13% to 2,954, with analysts attributing much of the spike to Carillion’s fallout.
The public backdrop
| Indicator | Figure | Note |
|---|---|---|
| Debt at collapse | ~£7bn | Against ~£29m cash |
| Suppliers exposed | 30,000+ | Top of a deep supply chain |
| Q1 2018 construction insolvencies | ~780 (+20% YoY) | Subcontractors worst hit |
| 2018 total | 2,954 (+13%) | Fallout attributed to Carillion |
The lesson is concentration risk: a company’s health depends on who it depends on. GalimAI’s map is where that dependency becomes legible — the directors, charges and signals that show where a counterparty failure would land.
The most plausible mechanism
The channel is counterparty exposure. When a large payer stops paying, subcontractors carrying that work lose receivables they cannot replace, and thin margins turn into insolvency within months. The tight timing — a visible Q1 2018 spike immediately after a January collapse — is the signature of contagion rather than coincidence. We read it as a strong, well-evidenced correlation with a clear mechanism, while noting that many affected firms were already thinly capitalised.
Sources
The proprietary figures in this study (the 463,022 companies, 1,000,000+ owners and the distress signals) are GalimAI first-party data. The public background figures are drawn from:
- Carillion blamed for Q1 spike in construction insolvencies - Global Construction Review
- Carillion - Wikipedia
Frequently asked questions
What did Carillion's collapse do to subcontractors?
It triggered a wave of failures down its supply chain. Construction insolvencies in Q1 2018 rose about 20% year on year to roughly 780, with specialist subcontractors hardest hit, and the 2018 total rose 13% to 2,954.
How big was Carillion's failure?
Carillion was liquidated in January 2018 owing close to £7bn against about £29m of cash, with more than 30,000 suppliers and subcontractors exposed.
What does GalimAI's data add?
GalimAI maps 463,022 property-owning companies linked to their directors, charges and distress signals - the connections through which a single failure spreads. That turns second-order, supply-chain risk into something you can see in advance.
Did Carillion cause the 2018 insolvency spike?
The timing and counterparty mechanism line up tightly - a Q1 spike immediately after a January collapse - making it a strong, well-evidenced correlation. Many affected firms were also thinly capitalised, so it was rarely the only factor.