Latest / The Cloud Business Podcast with Fexingo: AWS, Azure, GCP, and Enterprise Infrastructure / How Cloud Contracts Now Charge for Data Egress to On-Premises
Transcript
- Lucas: So you've migrated your workloads to the cloud — databases, analytics, maybe even your main ERP — and you're feeling pretty good about the agility gains. But then your cloud bill shows a line item you barely noticed before: data egress. And it's suddenly a lot more expensive than you expected. Luna: Right — everyone talks about ingress being free, but egress is where the real money is. And I think a lot of enterprises are only now waking up to how much they're paying to get data back out. Lucas: Exactly. And it's not just a few cents per gigabyte anymore. Over the past year, the three major providers — AWS, Azure, and Google Cloud — have all refined their egress pricing structures. In some cases, they've introduced tiered rates that escalate once you cross certain thresholds. Luna: So it's not exactly a price hike — it's more like a repricing that catches heavy users off guard. Got any specific numbers? Lucas: I do. I talked to a mid-sized retailer — about 2 billion in revenue — that runs its inventory and order management on AWS but keeps its legacy financial system on-prem. They sync data nightly, about 50 terabytes of egress per month. Under their old contract, egress was billed at a flat five cents per gigabyte. That's about $2,500 a month. Luna: Which is noticeable but not painful. Lucas: Right. But when they renewed in Q2 this year, AWS moved them to a new tiered structure. The first 10 terabytes are still five cents. The next 40 terabytes are eight cents. And anything above 50 — which they hit occasionally during peak seasons — jumps to twelve cents. Their monthly egress bill went from $2,500 to about $4,800 on average. That's a 92 percent increase. Luna: Wow. And that's just egress. I bet the finance team didn't see that coming when they signed the new contract. Lucas: They didn't. And the thing is, this isn't unique to AWS. Azure has a similar tiered model — they call it 'zone to zone' data transfer pricing. And Google Cloud, which historically positioned itself as the low-cost option, has also started introducing per gb charges for network egress beyond the first 100 gigabytes each month. Luna: So the narrative is shifting. Cloud providers want you to keep data inside their ecosystem. Egress fees are essentially a toll to prevent you from leaving — or even syncing back to your own data center. Lucas: That's exactly how a lot of enterprise architects see it. And it's causing some companies to rethink their hybrid architectures. Instead of syncing entire datasets nightly, they're implementing caching layers on-prem that only pull deltas. Some are even using third-party data transfer appliances to pre-filter what gets sent back. Luna: Which is a lot of engineering effort just to avoid a line item. But when you're talking about six-figure annual egress bills, it starts to make sense. Lucas: Absolutely. And there's a broader implication here: data portability. Regulators in Europe have been looking at cloud egress fees as a potential barrier to switching providers. The UK's Competition and Markets Authority actually opened an investigation earlier this year into whether these fees lock customers in. Luna: Right — and that's not just about on-prem. It's also about multi-cloud. If you want to run some workloads on Azure and others on GCP, moving data between them gets expensive fast. Lucas: Exactly. And that's a conversation we should dig into next episode — the multi-cloud egress trap. But for now, I want to focus on one practical takeaway: if your organization has a hybrid setup, you need to audit your egress patterns before your next contract renewal. Because the pricing models are changing faster than most procurement teams realize. Luna: And maybe start negotiating for a cap or a flat rate. Some of the larger enterprises are getting specialized agreements that limit egress charges to a fixed monthly fee. Lucas: That's a smart point. And honestly, if this episode was worth a coffee to you — maybe because it just saved you a few thousand dollars on your next bill — that's all we ask. You can find us at buy me a coffee dot com slash fexingo. No pressure, no perks, just a way to keep these conversations ad-free. Luna: Yeah, it really helps us keep the show going. And we really appreciate everyone who's chipped in. Lucas: So back to data egress — one more angle I want to cover: the rise of direct-connect services. Companies are now leasing dedicated network links to bypass the public internet — and those links often come with separate egress pricing that can be lower than the standard rates. Luna: AWS Direct Connect, Azure ExpressRoute, Google Cloud Interconnect — they all offer committed bandwidth at a discount. But the setup costs can be significant. Lucas: Right. The retailer I mentioned looked into Direct Connect. They would have needed a one-gigabit connection, which costs around $2,000 per month for the port, plus the data transfer charges. But the egress rate on that connection would drop to two cents per gigabyte. For 50 terabytes, that's about $1,000 in egress, plus the port fee — total around $3,000. Still more than their original $2,500, but way less than $4,800. Luna: And they get lower latency and more predictable performance on top of that. So there's a business case for it beyond just cost savings. Lucas: Exactly. But the key is that these decisions require planning. You can't just decide to move to Direct Connect two weeks before your contract ends. The lead time for port provisioning can be 30 to 60 days. Luna: So the message is: start your egress audit now. And if you're a smaller company without a dedicated cloud finance team, maybe look into third-party cost management tools that track egress and alert you before the bill spikes. Lucas: Good advice. There are tools from vendors like CloudHealth, ProsperOps, and even native services from the cloud providers themselves. But the first step is awareness. So many teams I talk to don't even have a baseline of their monthly egress volume. Luna: And it's not just about money — it's about architecture. If you design your systems to be egress-efficient from the start, you avoid the painful refactoring later. Lucas: That's a great closing thought. Thanks for listening, and we'll catch you next time on The Cloud Business Podcast.