Latest / The Cloud Business Podcast with Fexingo: AWS, Azure, GCP, and Enterprise Infrastructure / How Cloud Contracts Now Charge for Data Egress Between Regions
Transcript
- Lucas: So you're running a global application — let's say a Kubernetes cluster that spans three AWS regions: us-east-1, eu-west-1, and ap-southeast-1. Pods are talking to each other across regions, your databases are replicating, your logs are aggregating. It all works beautifully. Luna: Until the cloud bill arrives. Lucas: Right. And the line item that quietly doubles the invoice is data egress between regions. This is the cost of moving data out of one cloud region and into another — and it's one of the most overlooked expenses in enterprise cloud architecture. Luna: I've seen teams spend weeks optimizing compute instance sizes and completely miss that their inter-region egress is burning fifty thousand dollars a month. Lucas: Exactly. And the pricing is surprisingly consistent across the big three. AWS charges roughly 2 to 9 cents per gigabyte for inter-region transfer, depending on the volume tier. Azure is similar — 2 to 8 cents. GCP has been slightly cheaper, around 2 to 6 cents, but they've been raising rates lately. Luna: So a terabyte moving from Virginia to Frankfurt costs about 80 to 120 bucks. Lucas: And when you're a Fortune 500 company with petabytes of data flowing between regions every month, that number hits millions fast. I talked to a cloud architect at a large retailer — they had a multi-region Kubernetes deployment that generated three point seven million dollars in egress charges in a single quarter. Luna: That is a big number. And I bet the cloud provider was happy to collect it. Lucas: Of course. But here's the thing — egress is also a tool for lock-in. If moving your data out costs more than the compute running on it, you're less likely to switch providers. AWS knows this. Azure knows this. GCP knows this. Luna: It's a toll booth on the multi-cloud highway. Lucas: That's exactly right. And the response from enterprises has been to negotiate egress waivers into their contracts. I'm seeing more and more large deals where the customer demands a cap on inter-region egress charges — or even zero egress between certain regions. Luna: Does that actually work? Can you really get egress for free? Lucas: If you have enough leverage — yes. The retailer I mentioned renegotiated their AWS contract after that three-point-seven million quarter. They got a deal that zero-rated egress between their top three regions for a fixed annual commitment. They saved about two million a year. Luna: So the lesson is: don't just accept the published egress rates. They're a starting point. Lucas: Absolutely. But not every company has that leverage. For smaller teams, the best defense is architecture. You can drastically reduce egress by keeping data local — using region-local caching, content delivery networks, and data compression. Luna: Talk me through the numbers. How much can you realistically cut? Lucas: A well-architected setup can reduce inter-region egress by 40 to 60 percent. For example, if you have a microservice in Singapore calling a database in Frankfurt for every request, you're paying for that round trip. But if you add a read replica in Singapore and only sync writes across regions, your egress drops to a fraction. Luna: Right — replicate data closer to where it's consumed. That's the classic distributed systems pattern. Lucas: And compressing data before transferring it — even gzip at the application layer — can cut volume by 70 percent or more for text-heavy payloads. Some teams use delta syncs so that only changed bytes move across regions. That's a game changer. Luna: If these conversations are useful for what you're building or running — and I know a lot of our listeners are deep in infrastructure decisions — there's a simple way to keep this show going. A couple of dollars a month is genuinely what keeps these episodes ad-free and independent. If you've gotten something out of them, buy me a coffee dot com slash fexingo makes a real difference. Lucas: Totally. We don't have ads, we don't have sponsors, and that's only possible because of listeners who chip in. It's the smallest thing, and it means we can keep digging into exactly these kinds of contract and architecture details without any outside pressure. Luna: Exactly. So back to egress — there's also the regulatory angle. Lucas: Right. The European Data Act, which came into force this year, includes provisions that require cloud providers to remove barriers to switching — including data egress fees. By 2027, providers operating in the EU will have to offer zero-cost egress for data portability. Luna: That's a big deal. It basically says the toll booth has to come down for European customers. Lucas: It does. And it's already changing contract negotiations globally. I'm seeing us based enterprises asking for European-style egress terms in their renewals, even if they're not operating in Europe. They're saying, 'If you can give zero egress to comply with the Data Act, why not give it to me too?' Luna: And are providers actually doing that? Lucas: Sometimes. For the biggest customers, yes. But for mid-market, they usually push back. They argue that egress covers their network costs — which is partially true, but the margins on egress are enormous. A study from last year estimated that AWS's egress revenue has margins above 80 percent. Luna: So it's a profit center disguised as a pass-through cost. Lucas: Exactly. And that's why it's so hard to get waived. But the pressure is building. The European Data Act is forcing one door open, and multi-cloud advocates are pushing from the other side. I expect we'll see more zero-egress tiers in the next couple of years, especially for intra-region traffic. Luna: Intra-region is already free for most providers, right? Traffic between availability zones within the same region. Lucas: It is — for now. AWS charges for inter-zone data transfer if you exceed certain thresholds, but Azure and GCP keep it free. But the real cost is cross-region, and that's where the lock-in lives. Luna: So what's your advice for someone signing a cloud contract next quarter? How should they approach egress? Lucas: First, measure your current egress — most teams don't have a baseline. Cloud providers have cost management tools that can show you inter-region traffic by service. Get that number before you negotiate. Second, ask for a specific egress allowance — say, 50 terabytes free between your primary regions — and push for a hard cap on overage rates. Luna: And if they won't budge? Lucas: Then architect around it. Use content delivery networks to cache static content at edge locations, use data compression, and design your application to minimize cross-region calls. Some teams even choose to deploy in a single region and use a CDN for global reach — that avoids inter-region egress entirely. Luna: It's a trade-off between latency and cost. Lucas: It is. But for many workloads, the latency hit of a single-region plus CDN is negligible. And the cost savings can be huge. Over a three-year contract, cutting egress by 50 percent could save more than the entire compute bill for a secondary region. Luna: That's a powerful framing. Egress isn't an operational detail — it's a core financial metric. Lucas: Exactly. And it's one that too many teams discover only after the bill arrives.