Latest / The Cloud Business Podcast with Fexingo: AWS, Azure, GCP, and Enterprise Infrastructure / Cloud Contracts That Fine You for Data Gravity Waste
Transcript
- Lucas: So we just got the full text of a 2026 AWS Enterprise Discount Program renewal — the kind that covers a mid-sized fintech with about three petabytes across US East, EU West, and Asia Pacific. And buried in the fine print was a new clause: a data gravity waste penalty. Luna: Data gravity waste — that's the idea that once data accumulates, it's more expensive to move it than to just leave it and pay the storage fees. But now they're fining you for it? Lucas: Exactly. The clause specifically penalizes customers who store the same dataset in more than two regions without a documented business justification. For this fintech, it meant an extra $180,000 a quarter — because they had redundant copies in three regions for what AWS deemed 'non-essential' redundancy. Luna: Wow. And I imagine the justification has to be pretty specific — not just 'we want low latency everywhere'. Lucas: Right. The contract required a detailed mapping of each redundant dataset to a compliance requirement, a disaster recovery SLA, or a latency-sensitive workload. Generic 'multi-region for resilience' didn't cut it. And here's the kicker: the penalty wasn't a one-time fee. It was a monthly surcharge on every gigabyte of what they called 'gravitated excess' — 15 percent above the standard egress rate. Luna: So they're basically monetizing the inertia that data has. The harder it is to move, the more they charge you for keeping it around. Lucas: And it's not just AWS. We're seeing similar language in Azure's latest Enterprise Agreement updates and GCP's CUD — committed use discount — renewals. Microsoft calls it 'storage adjacency optimization', Google calls it 'data efficiency surcharge'. But the mechanism is the same: a penalty for holding data in too many regional buckets without a clear operational need. Luna: I'm curious what the origin story is here. Are cloud providers actually losing money on data storage, or is this a way to push customers toward their higher-margin compute services? Lucas: Both, actually. Storage margins have been compressing for years — especially with the price wars between the three big players. Meanwhile, egress fees have always been a profit center. This clause essentially extends that profit logic to the data itself. If you're going to store three copies, you're going to pay for the privilege of not moving them. Luna: And it aligns with the broader trend of cloud contracts becoming more punitive. We've done a bunch of episodes on new penalty clauses — GPU decommissioning fees, data egress audits, carbon inefficiency penalties. This feels like the next logical step. Lucas: It does. But I think there's something more subtle happening here. Cloud providers are trying to prevent a specific kind of behavior: the 'set it and forget it' approach to data. In the early days of cloud, companies would spin up buckets, replicate data, and never audit what they actually needed. The hyperscalers want to train customers to be more disciplined — and they're using the contract to enforce that discipline. Luna: But how do you negotiate against this? If I'm a procurement lead at a large enterprise, I'm not going to just accept a 15 percent surcharge on my redundant storage. Lucas: You're right. And we've seen some effective carve-outs. The most common is a disaster recovery exception — you can store a full copy in a second region for DR, as long as you can prove you actually test failover annually. Another is compliance: if a regulator like the SEC or the EU's GDPR requires local data residency, you can keep a copy in that region without penalty. But you have to list those regulations explicitly in the contract. Luna: So it's not just about the fine — it's about the administrative burden of documenting every copy. That's a hidden cost too. Lucas: Exactly. The fintech I mentioned ended up spending about 40 engineer-hours just cataloging their data stores to justify which copies were exempt. And after that exercise, they found they could delete about 30 percent of their redundant data entirely — it was old logs, test databases, and staging environments no one was using. Luna: So the clause actually forced them to clean house. Maybe that's the unintended upside — better data hygiene. Lucas: That's the argument the hyperscalers are making. They frame it as a 'shared responsibility for cost efficiency'. But the cynic in me notes that the penalty rate — 15 percent above standard egress — is higher than what it would cost to just move the data to a cheaper storage tier like AWS S3 Glacier. So the provider wins either way: either you pay the surcharge, or you pay to move the data to a less accessible tier that you'll then pay retrieval fees for later. Luna: It's a clever trap. You're either charged for staying or charged for leaving. Lucas: And that's why FinOps teams are now getting involved in contract negotiations. A few years ago, FinOps was mostly about right-sizing instances and reserved instances. Now they're auditing contract language for these kinds of clauses. I spoke to a FinOps manager at a large retailer who told me they now run a 'data gravity audit' before every renewal — they map every dataset, its regional distribution, and its business justification, and they use that as leverage to negotiate lower penalty thresholds. Luna: What kind of leverage do they have? If the hyperscaler insists on the clause, can the customer really push back? Lucas: Yes, but it depends on scale. If you're a $50 million annual spend customer, you can often negotiate a higher threshold before the penalty kicks in — say, five petabytes instead of two. Or you can negotiate a lower surcharge — 10 percent instead of 15. Some customers have even gotten a 'data gravity holiday' for the first twelve months of a new contract, giving them time to rationalize their storage. Luna: So the penalty is real, but it's also a starting point for negotiation. That's actually more interesting than a hard-and-fast rule. Lucas: Exactly. And the negotiation itself reveals something about the customer's maturity. If you walk in with a clear data map and a business case for each redundant copy, you have a stronger hand. If you show up and say 'we need multi-region for everything', you're going to get hit with the full penalty. Luna: I wonder how this plays into the broader multi-cloud trend. If you're on two or three clouds, data gravity waste could be even worse — you might have copies in AWS, Azure, and GCP for the same workload. Lucas: That's a great point. And so far, the penalty clauses only apply within a single provider's ecosystem. They can't fine you for storing data on a competitor's cloud. But that might actually push companies toward single-cloud strategies to avoid the complexity of managing cross-provider penalties. At least with one provider, you only have one set of rules to navigate. Luna: So the penalty clause could ironically reduce multi-cloud adoption, even if the original intent was just to cut waste. Lucas: That's the irony. And it's not just about penalties — it's about the administrative overhead. Every new cloud adds a new contract, new compliance requirements, new audit cycles. The data gravity penalty clause is one more reason to consolidate. Luna: Let's talk about the specific numbers in that AWS contract. You mentioned $180,000 per quarter for a mid-sized fintech. How did they calculate that? Lucas: It breaks down like this: they had three petabytes of data, of which about one petabyte was deemed 'excess' — stored in three regions without a documented justification. The surcharge of 15 percent on the standard egress rate for that one petabyte came out to roughly $60,000 per month. And because it was billed monthly, it added up to $180,000 per quarter. Luna: And the standard egress rate is around $0.09 per gigabyte for the first ten terabytes, then tiered down. So 15 percent on top of that adds up fast. Lucas: Exactly. And remember, this is a fintech — they're not even storing video or large media files. This is mostly transaction logs, customer data, and compliance archives. Imagine what a media company or a genomics firm would face. Luna: So this is likely to hit heavily regulated industries hardest — finance, healthcare, insurance — where data retention requirements force multiple copies. Lucas: Yes, but those industries are also the ones best positioned to negotiate exceptions, because they have clear regulatory mandates. The real pain is for mid-market companies that don't have dedicated legal or compliance teams to document every copy. They're the ones who end up paying the surcharge without realizing it until the bill arrives. Luna: And that's where the role of the cloud broker or managed service provider becomes crucial. If you're a mid-market company using an MSP, they should be flagging these clauses and helping you negotiate. Lucas: Absolutely. In fact, some of the larger MSPs are now offering 'contract risk assessments' as a standalone service. They'll review your cloud agreements and highlight exactly these kinds of hidden penalties. I've heard of one case where an MSP saved a client $400,000 a year just by renegotiating the data gravity waste clause. Luna: That's real money. And it's not just about saving costs — it's about avoiding surprises. Nobody likes a surprise six-figure bill. Lucas: Speaking of avoiding surprises — if these kinds of deep-dive conversations are useful for what you're building or running, this show stays ad-free thanks entirely to listener support. The economics of a niche podcast like this are weird — we don't have scale for advertisers, but we have a very engaged audience. And a bunch of listeners have started chipping in at buy me a coffee dot com slash fexingo. It's a small gesture that makes a big difference in keeping the lights on and the research flowing. Luna: Yeah, I love that we can stay focused on the granular details — like data gravity penalty calculations — without having to pause for ad reads. It makes the show better. Lucas: So back to those calculations — the fintech we mentioned ended up reducing their redundant data by 30 percent after the audit. That alone saved them more than the penalty cost, even before they negotiated the surcharge down. So in a weird way, the penalty clause forced them to become more efficient. Luna: It's almost like the penalty is a feature, not a bug — it drives behavior change. But only if you know about it and act on it. Lucas: Right. And that's the key takeaway for me: the data gravity waste penalty is a signal from the hyperscalers that they're serious about cost governance. They're not just selling storage anymore — they're selling a discipline. And customers who embrace that discipline will come out ahead, while those who ignore it will pay a premium. Luna: So for 2026, the smart play is to audit your data storage before your next contract renewal. Map every petabyte, justify every copy, and negotiate your carve-outs early. Lucas: Exactly. And if you're building a multi-cloud strategy, factor in these penalties from day one. The cost of data gravity waste might just tip the scales toward a single-cloud approach for a lot of organizations. Luna: I think that's a real possibility. And it changes the calculus for cloud architects who thought multi-cloud was always the safer bet. Lucas: It does. And it's another reminder that cloud contracts are becoming more complex, more nuanced, and more consequential. The fine print isn't just legalese anymore — it's a strategic document. Luna: And it's worth reading carefully. Lucas: Worth every minute. Thanks for listening, and we'll be back next week with another clause that might change how you think about your cloud spend.