From FIDIC to Datacloud: It's not the contract, it's the conduct

I closed my article following the FIDIC Contract Users' Conference earlier this year with the observation that contracts don't deliver projects; people do. However FIDIC recognises that the contract must still be fair, balanced and apportion risk to the party best able to manage it. It is only by choosing the right form of contract with an appropriate risk allocation, which the parties understand and are familiar with, that we enable effective contract management and constructive human behaviours to deliver successful projects.

It is a conclusion that resonates well beyond the world of FIDIC practitioners, as I was reminded earlier this month at the Datacloud Conference in Cannes.

FIDIC forms are, of course, widely used in the data centre sector. But where the FIDIC  Conference examined the contracts and risk allocation broadly across the construction industry, Datacloud is thoroughly sector-specific: operators, hyperscalers and insurers speaking frankly about what happens in practice when building data centres at every scale. A contrast in perspective - but with conclusions that were remarkably similar.

The same questions, a different vantage point

A panel session at Datacloud on contracting trends opened with an interesting statistic: nine out of ten data centre projects are delayed, with an average delay of 34% of the programme. When addressing whether the form of contract could improve these figures, the answer was unambiguous: it makes no material difference. Neither alliancing contracts nor particular standard forms were a driving factor in the data.

This will resonate with those who attended the FIDIC Conference and the theme that people, not contracts, deliver projects. The Datacloud panel reached the same conclusion from a different direction: where the right people and relationships are in place, risk is far more effectively managed than through drafting alone. Sticking in contractual penalties and liquidated damages is not effective and indeed may not even be possible given the scale of new gigawatt developments, where the level of exposure is such that meaningful financial remedies could be commercially unviable or even existential for a contractor. Therefore, they are rarely a realistic risk management tool.

While the keen interest in FIDIC’s upcoming collaborative contract reflects a desire to promote collaborative behaviours, this evidence (at least in the data centre sector) suggests that collaborative contract forms do not, of themselves, necessarily change outcomes. It is people's attitude, their approach and their willingness to work together that drive outcomes, rather than any particular set of contractual provisions.

Risk identification and allocation is key

Saying that the form of contract is not the defining factor is emphatically not the same as saying that negotiation does not matter. Quite the opposite.

FIDIC's position is clear: it is in no-one's interest to apportion risk inappropriately, as this risks insolvencies, mismanagement and a claims culture. The Datacloud panel made the same point in starker commercial terms: with the scale of modern data centre projects (and gigawatt campuses representing a fundamentally different order of risk to anything seen before) some penalty regimes could sink or destabilise even major companies in the market.

The message from both communities is therefore the same: negotiate thoroughly and negotiate early, not to secure the best position for a future dispute, but to avoid one. Identify the risks, discuss them openly, and allocate them clearly before work begins. As one panellist put it, far better to have an honest conversation at the outset about risks such as power shortages and permitting delays because nobody wants the argument at the end. Knowledge of the contracts, and of the risks they govern, is paramount to ensuring risk is properly allocated and managed by the appropriate party to do so. An honest, upfront conversation is far more valuable than pushing all the risk onto a party who may be unaware of it or ill equipped to manage it.

The role of insurance: backing off risk where possible

BCLP was delighted to co-host a workshop at Datacloud with Marsh on data centre contract risk management, focused on the alignment of contractual risk with insurance and the role of effective contract management and insurance structuring in protecting projects. It was a recurring theme across the conference and one that needs to be addressed at the very earliest stages of a project.

A separate panel on emerging risks and insurance coverage also looked at this. The panel reinforced the same point: risks need to be addressed throughout the project; from design and team structure through to operational and maintenance regimes- all of which should underpin the service level agreements (SLAs) to which operators have committed. Gaps are inevitable, and the insurance industry needs to evolve products that align with changing risks.

One new product reported by Marsh and Aon as gaining traction is SLA insurance cover. It covers an operator's liability to pay SLA penalties for service outages due to a myriad of reasons such as human error, cyber attack and machine breakdown — although bundling so many risks into a single product was acknowledged to be a challenge. Apparently, most data centre outages are attributable to human error, with a correlation between errors and staff returning from holiday! Such a product serves two functions: risk coverage itself, and speed of pay-out, where liquidity is critically important. Products offered are parametric, providing a pre-agreed payout when a specified event occurs. The trigger is simply an SLA violation and a policy should pay within 30 days. The panel is not yet seeing such cover as a lender requirement, but considered that it could, in theory, reduce financing costs. The product is expensive, and whether the risk mitigation justifies the premium will depend on the operator's risk profile - although the cost may be capable of being absorbed into customer charges as part of the wider commercial model.

One important caveat insurers were keen to stress is that insurance should not be a vehicle for taking more risk than a party would otherwise accept. The risk is that an SLA insurance product drives an operator to take risks it would not normally take, simply because it knows its losses are insured. That would rather undermine the purpose.

The insurance panel also confirmed something that is relevant to anyone negotiating force majeure clauses in data centre contracts: there is surprising variation in the consistency of FM definitions and their application across contracts. Detailed, clear definitions are preferred as those that are vague or require a court to interpret cause real difficulty.  This is particularly significant under English law, where there is no common law doctrine of force majeure (only the very narrow doctrine of frustration) meaning that parties' risk (and insurance cover) is entirely dependent on contractual FM provisions. Clear, detailed drafting is therefore essential.

Another innovative product available in the market is insurance against judicial review in the planning process. For data centre developers navigating complex planning environments, where a judicial review challenge can cause significant delay and cost,  this represents a meaningful new tool for risk management that practitioners and developers alike should be aware of.

Looking forward

There is an appetite globally among contract users for genuine human engagement and contract management over contract terms themselves. Standing in the Cannes sunshine, having attended numerous panel events, talked to more delegates than I can count, co-hosted a workshop on contract risk management with Marsh, and heard from the insurance market about the products now available to support better risk allocation, it seems to me that appetite extends well beyond the construction industry.

The form of contract matters a bit. The risk allocation within it matters enormously. But it is the willingness of the parties to engage with that allocation honestly, before the contract is signed, not after problems arise, and to use every available tool, including the insurance market, to support that allocation, that makes the real difference. That is not a legal insight so much as a human one. And it seems to travel well.

A version of this article was published in PLC Construction on 22 June 2026.