Latest / The Cloud Business Podcast with Fexingo: AWS, Azure, GCP, and Enterprise Infrastructure / The One Percent Rule That Is Rewriting Cloud Contracts
Transcript
- Lucas: If these conversations are useful for what you're building or running, you might want to stick around for the end — but first, let's talk about something that's quietly rewriting cloud contracts right now. Luna: I'm guessing this isn't about GPU clauses again, because we've done a bunch of those. Lucas: No, this one is more subtle — and arguably more consequential. It's what I'm calling the one-percent rule. A growing number of cloud service agreements now include a clause that says: for any service disruption, regardless of duration or severity, the customer's only remedy is a credit equal to one percent of their monthly bill for that service. Luna: Wait — so if my entire production environment goes down for six hours, I get one percent off? That's basically nothing. Lucas: Exactly. And that's the point from the provider's perspective. Historically, cloud SLAs were structured around uptime percentages. Five-nines, four-nines, and if you dropped below a threshold, you'd get a tiered credit — ten percent for 99 percent uptime, twenty-five percent for 95 percent, that kind of thing. Luna: Right, so the worse the outage, the bigger the credit. The one-percent rule flattens that entirely. Lucas: It does. And it first started appearing in AWS's enterprise agreements about eighteen months ago — quietly, in the fine print of renewed contracts. Now Azure and GCP are both piloting similar language in their select-tier agreements. Luna: A couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. Lucas: Yeah, it's a small thing that makes a big difference for us. And speaking of small things that make a big difference — this one-percent rule is exactly that. A one-line change that could shift billions in liability. Luna: So who pushed for this? Was it the hyperscalers trying to cap their risk, or did customers ask for it? Lucas: It's almost entirely provider-driven. The official rationale is simplification: one flat credit, no complex uptime calculations, no disputes over whether an outage counts toward the SLA. But the real effect is that it caps the provider's maximum liability at one percent per incident, no matter how catastrophic. Luna: And for a large enterprise spending, say, ten million a year on AWS, one percent is a hundred thousand dollars. That's not nothing, but if an outage costs them millions in lost revenue, the credit is a rounding error. Lucas: Exactly. I talked to a fintech company — mid-market, PCI workloads on AWS — that had this clause slipped into their renewal last quarter. They didn't catch it because the contract was hundreds of pages. Their previous SLA would have given them a twenty-five percent credit for an outage that took down their payment processing for four hours. Luna: Let me guess — under the new clause, they got one percent. Lucas: They did. Their finance team calculated that the new clause reduced their expected annual SLA credits by about forty percent, even though their actual reliability didn't change. The provider's liability just shrank. Luna: So what's the defense? Are there any carve-outs? Lucas: There are some. The one-percent rule typically doesn't apply to certain critical services — like core database or networking — where providers still offer tiered credits. And some contracts have a cumulative cap: if you have multiple incidents in a month, the total credit can't exceed, say, five percent. Luna: But for most compute and storage, it's one percent flat. Lucas: Right. And the trend is spreading. GCP's new 'simplified SLA' for their standard tier includes a flat two percent credit for any outage below 99.9 percent uptime. Azure's enterprise agreement now has a similar provision for their 'basic support' tier. Luna: So what should a CFO or procurement lead look for in their next renewal? Lucas: First, read the SLA section carefully — don't assume the language is the same as last time. Look for any mention of 'fixed credit' or 'single percentage'. Second, ask for an exception for production-critical workloads. If you're running PCI, HIPAA, or real-time trading systems, negotiate a separate SLA with tiered credits for those services. Luna: We've talked before about the importance of benchmarking your actual uptime. If you have your own monitoring data, you can push back on the provider's numbers. Lucas: That's key. One analyst I spoke with called the one-percent rule 'the most consequential fine print change since reserved instances.' Because reserved instances changed how you paid; this changes what you get when things go wrong. Luna: And it's happening at a time when cloud outages are still frequent enough that enterprises worry about them. Just last month, a major Azure region had a six-hour latency spike that affected thousands of customers. Lucas: Right. Under the old SLA, those customers might have gotten ten or fifteen percent credits. Under the new one, it's one percent. The providers are effectively transferring more risk back to the customer. Luna: So what's the counterplay? Can customers just say no? Lucas: It depends on your leverage. If you're a small shop, you probably can't negotiate custom SLAs. But if you have a seven-figure annual spend, you can push back. Some enterprises are now requiring that any new contract include a 'critical workload addendum' that carves out certain services from the one-percent rule. Luna: And there's always the multi-cloud argument — if one provider's SLA is too weak, you can shift workloads to another. Lucas: Exactly. But that requires architectural flexibility. The providers know that once you're locked into their ecosystem, it's hard to leave. So the one-percent rule is a test of how much lock-in they can exploit. Luna: Is there any regulatory angle here? Could regulators step in if this becomes standard? Lucas: It's possible, but unlikely in the near term. Cloud SLAs are still mostly governed by contract law, not industry regulation. However, if a major outage hits a critical infrastructure provider — say, a bank or a hospital — and the credit is trivial, there could be political pressure. The UK's FCA has already started asking questions about cloud concentration risk. Luna: So we're at an inflection point. The one-percent rule is here, but it's not universal yet. Enterprises that act now can negotiate around it. Lucas: Right. And the window is closing. As more providers adopt it, the baseline will shift. My advice: audit your current contracts, benchmark your uptime, and start the conversation with your account rep before renewal. Luna: Good advice. I'll be checking our own contracts this week. Lucas: Same here. That's it for today — next time, we'll look at how cloud providers are starting to charge for API calls that were previously free. Luna: Sounds like another fine print adventure. Lucas: It always is. Thanks for listening.