The liability cap is the only clause most teams actually negotiate

If you audit where a commercial legal team spends its negotiation hours, one clause dominates. Not confidentiality, which is largely settled market practice. Not governing law, which is decided by whoever has the better leverage and then forgotten. Limitation of liability absorbs more argument than every other provision combined, and a surprising proportion of that argument is spent on the wrong variable.

Why the cap absorbs the argument

The liability cap is the clause where an abstract commercial relationship becomes a number. Every other provision describes what the parties will do. This one describes what happens when they fail, and it converts every unspoken assumption about risk appetite into a figure someone must defend to their board.

It is also the clause with the clearest opposing incentives. A supplier wants the cap at or below the revenue of the deal, because liability above contract value turns a profitable engagement into an uninsurable one. A customer wants the cap to reflect its exposure, which for anything touching operations or personal data is unrelated to what it is paying. Neither position is unreasonable. They are simply measuring different things: the supplier measures the deal, the customer measures the damage.

That structural conflict is why the clause cannot be solved with a market-standard formulation. It can only be traded, which means the reviewer's job is not to find the correct cap but to understand what the organisation is willing to give up elsewhere to move it.

The number is the least interesting part

Teams anchor on the multiple. Twelve months of fees, one hundred and fifty percent of contract value, two times annual charges. The multiple is easy to argue about because it is a single figure, which is exactly why it consumes disproportionate time.

The variables that actually determine exposure are usually settled without discussion:

The base. Fees paid, fees payable, fees in the preceding twelve months, or total contract value. On a three-year agreement with front-loaded implementation charges, these produce wildly different amounts from the same multiple.
The aggregation. Whether the cap applies per claim, per contract year, or in aggregate across the entire term. An uncapped-per-year formulation on a five-year deal is materially five caps.
The exclusion of indirect loss. Whether loss of profit, loss of revenue and loss of anticipated savings are excluded, and whether those exclusions apply to direct as well as indirect losses. A well-drafted exclusion can make a generous cap almost unreachable.
What sits outside the cap entirely. The carve-outs, which is where the real money lives.

A reviewer who accepts a twelve-month cap with a strong indirect-loss exclusion and no carve-outs has agreed to less protection than one who accepts six months with data-breach liability carved out. The multiple looked better. The exposure was worse.

Three carve-outs worth the fight

Most carve-out lists are copied between agreements without much thought. Three are worth genuine argument, and the rest are usually negotiable currency.

Confidentiality and data breach

Where the supplier processes personal data or confidential material at scale, capping breach liability at twelve months of fees means the customer is self-insuring for a regulatory exposure it cannot control. Suppliers resist an unlimited carve-out for good reason. The workable answer is usually a super-cap: a separate, higher limit for this category, sized against the supplier's cyber insurance rather than against the deal.

Third-party intellectual property claims

If the supplier's product infringes someone else's rights, the customer is defending a claim about a product it did not build. This is the clearest case for uncapped or highly capped liability, because the risk is entirely within the supplier's control and entirely outside the customer's.

Wilful misconduct and fraud

In many jurisdictions a cap on liability for fraud is unenforceable regardless of what the contract says, so arguing about it is often theatre. It costs a supplier nothing to concede and it costs a customer nothing to demand. Treat it as a low-value trade rather than a principle.

The purpose of reviewing a cap is not to arrive at the largest number. It is to make sure that the categories of loss your organisation cannot absorb are the ones that sit outside it.

Super-caps and why they help

The binary between capped and uncapped forces both parties into positions neither wants. A tiered structure resolves more negotiations than any amount of arguing about multiples: a general cap at twelve months of fees, a super-cap at three times annual fees for data protection breaches, and genuinely uncapped liability only for the two or three categories where a cap would be unenforceable anyway.

This works because it lets each side defend the outcome internally. The supplier's finance function sees a bounded, insurable maximum. The customer's risk committee sees meaningful recovery in the scenario it actually fears. The alternative, a single number that satisfies neither, is what produces the fourth and fifth rounds of redlines.

A reviewer checklist

Before signing off on any limitation of liability clause, confirm the following in order. The order matters, because the later items can make the earlier ones irrelevant.

Identify the base and the period the multiple applies to, and calculate the actual maximum in currency rather than reasoning about the multiple.
Determine whether the cap is per claim, annual or aggregate across the term.
Read the indirect loss exclusion and check whether it excludes categories of loss that would in practice be your only recoverable damages.
List what is carved out, and check that it covers data protection, third-party IP infringement and wilful misconduct as a minimum.
Check whether the carve-outs are mutual. A one-sided carve-out list in the counterparty's favour is common and rarely justified.
Confirm the cap interacts correctly with the indemnity clause. An indemnity that sits inside a low cap is decoration.
Confirm the supplier's insurance obligations are at least the size of the super-cap you agreed. A limit that exceeds the policy is a number, not a remedy.

None of this requires a senior lawyer. It requires a structured pass that happens the same way every time, which is precisely the kind of work that should be automated so that the judgement about what to trade can be made by someone who knows the commercial context.

Clauzy publishes field notes for contract reviewers. Nothing here is legal advice; it is practitioner commentary, and your own counsel decides what applies to your agreements.