Vercel's flat rate CDN caps a viral spike at a fixed price: the math solo builders should actually run
On September 8, Vercel took Flat Rate CDN out of beta and made it generally available for every Pro team. The pitch is simple: instead of paying per CDN request and per gigabyte of Fast Data Transfer, you pay a fixed monthly price, and a Hacker News front page hit or a viral tweet cannot push your bill past that number. If you have ever shipped something on Vercel and felt a flash of dread when a link started spreading faster than you expected, this announcement is aimed directly at you. It landed exactly a week after Vercel introduced Fluid Compute on September 1, a shared runtime for functions, builds, servers, and sandboxes meant to handle the same kind of irregular traffic. Together they read like Vercel finally admitting that "unpredictable spike" is the normal shape of a solo operator's traffic, not the exception.
What actually shipped
Flat Rate CDN replaces metered CDN Requests and Fast Data Transfer billing with a capacity tier you choose up front. Pro accounts get the base tier included at no extra cost: 1 million CDN requests and 1 TB of data transfer a month, which Vercel's own docs work out to roughly 33,000 requests a day. Above that, the paid tiers are $20 a month for 10 million requests and 50 TB of transfer, $100 a month for 50 million requests and 50 TB, and $300 a month for 150 million requests and 50 TB. Each tier also absorbs Blob Data Transfer and the observability events your CDN requests generate, so it's not just page views, it's most of what shows up on a typical usage bill today.
The part that actually matters for the "viral moment" scenario is how spikes get handled. Your tier stays fixed for the entire billing period. A single bad, or good, day of traffic doesn't bump you to a higher tier or generate an overage charge. Vercel only reassesses your tier at the start of the next billing cycle, based on sustained usage, not a one-off burst. In the blog post announcing GA, Vercel described a team whose normal traffic fit comfortably in the $20 tier, until a viral product launch sent it far past that. Under the old usage-based pricing, they estimated the bill would have reached the tens of thousands of dollars. On Flat Rate CDN, it stayed at $20.
An insurance product, not a discount
Here's the framing I think gets lost in the launch post: this is not Vercel giving you cheaper CDN. It's Vercel selling you insurance, and insurance is a bet you're structurally likely to lose in any given month. You pay $20 whether or not anything unusual happens, in exchange for a guarantee that the one month something does happen, you're not staring at a bill with four digits you didn't budget for. That trade can be a genuinely good one. It can also be money you hand over every month for a risk that, realistically, was never going to materialize for your particular project.
I run a handful of small sites on Vercel, and the honest answer for most of them is that nothing has ever gone viral, and nothing is likely to. A recipe blog with steady, boring traffic from search doesn't need spike insurance any more than a house with no history of flooding needs flood insurance. The math only tips in your favor when either your typical usage already justifies the tier price on its own, or your probability of an actual spike is high enough that the expected cost of an uninsured bad month outweighs the guaranteed monthly premium.
Pull your own numbers before you decide anything
Before opting in, go look at your actual CDN spend. Vercel's usage dashboard breaks out CDN Requests (labeled Edge Requests on your bill) and Fast Data Transfer separately, and you can filter by project and by month. Pull the last two or three months and add those two line items together, ignoring Fast Origin Transfer and Functions costs for now since those aren't part of this comparison. That number, averaged, is your real baseline. Most side projects and early-stage solo products I've seen fall well under $20 a month in combined CDN Requests and Fast Data Transfer, sometimes into single digits.
Running the actual math
Say your last three months averaged $6 a month in CDN Requests and Fast Data Transfer combined, on a blog or a small SaaS with steady, non-viral traffic. Opting into the $20 flat tier means paying roughly $14 a month, or about $168 a year, purely for the possibility that one day you spike hard enough to matter. If your product has never been posted anywhere with real reach and isn't likely to be, that's $168 a year for a scenario you can reasonably estimate at well under 5% likelihood in any given year. That's a bad bet, made repeatedly.
Now flip it. Say you're about to launch on Product Hunt, or you've built something with genuine share potential, a tool, a calculator, something that plays well on X or gets picked up by a newsletter with real distribution. Your baseline CDN spend might still be low today, but the tail risk is real and the potential downside is a bill that could wipe out a month, or several months, of revenue. In that case, $20 a month for the flat tier, or even $100 if you expect bigger swings, is cheap protection against a scenario with a much higher probability and a much worse outcome. The flat tier also buys you something the pricing page doesn't quantify: you stop babysitting your usage dashboard during a launch, which for a solo operator juggling everything themselves is worth something on its own.
What flat rate doesn't cover
One nuance worth knowing before you flip the toggle: Flat Rate CDN covers CDN Requests, Fast Data Transfer, Blob Data Transfer, and CDN-related observability events. It does not touch Fast Origin Transfer, the data moving between the CDN and your Vercel Functions, which stays usage-based regardless of which CDN plan you're on. If your app leans heavily on server-side rendering, API routes, or AI model calls that hit a function on every request, a traffic spike can still generate a real bill on that side even with Flat Rate CDN turned on. Vercel's fair use guidelines also exclude bulk file distribution and large-scale media delivery from the flat tiers entirely, so if your product is mostly serving big files or video, this isn't built for you, and Vercel says it will move you to its separate Flex CDN tier if your usage looks like that.
The honest take
If your traffic is flat, predictable, and your last few months of CDN Requests and Fast Data Transfer already sit comfortably under $20 combined, skip Flat Rate CDN. You're paying a recurring premium for a scenario you can reasonably judge unlikely, and that money is better spent elsewhere in a solo operation where every dollar has an opportunity cost. Turn it on when you have a specific, real reason to expect a spike, a launch, a press mention, a feature that's genuinely shareable, not because a viral moment is theoretically possible for any website on the internet. And if you do turn it on for a launch, there's nothing stopping you from opting back out to pay-as-you-go once the traffic settles back down and your baseline usage no longer justifies the tier. Treat it the way you'd treat any insurance policy: buy it when the risk profile actually changes, not by default.
Author
Lukas
@lukcombinator