That "greater of" trick is something, but it's still a rounding error when their mistake takes down your revenue stream for a day.
>How do you even start that push back?
You don't start with a marked-up agreement. You start with an email to the rep and their sales engineer: "We can't proceed with this liability clause as-is. What's your process to review exceptions?" It forces them to reveal their internal escalation path, or lack of one, before you waste time on edits they can't approve.
If they balk, you have your answer.
-- old school
Yeah, that's a good point about shifting the focus to specific SLA penalties. I've been trying to get my head around the actual AWS agreements, and it's pretty much the same. The cap is tiny, but the indemnity for their own IP issues feels wide open for them to claim.
Your tactic makes sense, to at least tie up their resources if something goes wrong. Makes me wonder, has anyone ever seen those separate penalties actually get paid out? I'm worried they'd just fight that too.
The "standard" defense is the first filter. Your negotiation leverage depends entirely on whether you're buying seats for a startup or provisioning a seven-figure infrastructure commit.
I've seen a multiple of ACV negotiated exactly once, for a multi-year, eight-figure deal with a major cloud provider. The multiple was 1.5, and it was treated as a major concession. More common is a fixed minimum floor, like $500k, regardless of fees paid. That's still low but better than a pure fee-based cap.
Forget gross negligence as a carve-out; their legal won't touch it with a ten-foot pole. A carve-out for direct costs of a data breach or IP infringement is sometimes possible, but they'll fight hard to exclude consequential loss. Walking away is a valid signal, but it only works if you have a viable alternative.
benchmark or bust