Render Cut Free-Tier Bandwidth 20x in April 2026 — Here's What the Same App Costs on a $4.59 Hetzner Box

On April 23, 2026, Render tore up its pricing model. Per-seat workspace fees — $19/user on Professional, $29/user on Organization — are gone, replaced by flat monthly fees. The number that should actually worry a bandwidth-heavy hobby project: included egress on the Hobby plan dropped from a flat 100GB a month to 5GB. Pro drops from 100GB to 25GB. A personal blog that never paid a cent for bandwidth under the old plan can now owe a real bill under the new one.

That's the headline. It's also only half the story, and the other half cuts the other way: Render simultaneously cut its overage rate from a blocky $30-per-100GB penalty to a granular $0.15/GB. For an account that was already blowing past its free allowance, that rate cut can outweigh the smaller allowance entirely. Below is the line-by-line math for both cases, and what the same workload costs run on owned hardware instead — where bandwidth isn't a line item a vendor can reprice on four months' notice.

What Actually Changed, Line By Line​

Render's own numbers, from its pricing changes blog post and new workspace plans docs:

Existing workspaces can opt in immediately; anything still on a legacy plan converts automatically on August 1, 2026. Compute pricing — Render's per-instance fees for web services and workers, $7/month Starter up to $450/month Pro Ultra — is unchanged; this repricing is entirely about the workspace subscription and the bandwidth meter sitting inside it.

The Hobbyist Tax: From $0 to a Real Bill​

Here's where the "100GB to 5GB" number actually bites. A personal blog or portfolio site with a modest image gallery — nothing dramatic, just a typical WordPress-style traffic pattern — can easily push 60GB of outbound bandwidth in a month. Under the old Hobby plan, that usage sat comfortably inside the 100GB free allowance: $0. Under the new plan, only 5GB is free:

At low usage, the story is unambiguous: a free tier that quietly billed nothing now bills something, every month, forever, with no code change and no traffic growth required to trigger it. That's the real content of "Render slashed Hobby bandwidth 20x" — it isn't a hike on top of an existing bill, it's the introduction of a bill where there wasn't one.

The Team Case: Where the Repricing Actually Helps​

Now take a 3-person team running a media-heavy app on Pro — product screenshots, user uploads, the kind of workload that pushes 400GB of egress a month. Compute: one Standard instance at $25/month.

Old pricing:

• Seats: 3 × $19 = $57
• Compute: $25
• Bandwidth: 400GB usage, 100GB free, 300GB over → 3 blocks of 100GB × $30 = $90
• Total: $172/month

New pricing:

• Workspace: $25 flat, unlimited seats
• Compute: $25
• Bandwidth: 400GB usage, 25GB free, 375GB over × $0.15 = $56.25
• Total: $106.25/month

That's a 38% drop, and it isn't close — driven almost entirely by two mechanisms pulling in the same direction once a team has more than one seat and real bandwidth volume: flat-fee seats beat $19/head as soon as a workspace has two or more people, and the granular $0.15/GB rate beats the old $30-per-100GB-block penalty for anyone who was routinely spilling into overage.

The seat math alone is worth isolating, because it's the cleanest way to see where the crossover sits:

A solo developer is the one case where the new flat fee is a strict loss on seats alone — $25 instead of $19 — before bandwidth is even in the picture. Every seat beyond the first is where the old per-seat model starts costing more than the new flat $25, and by 10 seats the old model's seat fee alone is more than seven times the entire new workspace subscription. That's the shape of the 75%-of-customers-pay-less-or-the-same claim: it's true for essentially any team of two or more, and it's the solo, low-bandwidth accounts — the same population hit by the shrunk free allowance in the previous section — who end up on the losing side twice.

So which is it — a price cut or a price hike? Both, and the split runs along a specific line: the shrunk free allowance is what hurts low-usage accounts; the halved overage rate is what helps high-usage ones. A hobby project or small team sitting comfortably under the old 100GB ceiling never touched the overage rate at all — for them, the only thing that changed is the size of the free bucket, and it just got smaller. A team already paying overage fees barely notices the smaller bucket because they blew past it in both models; what they feel is the per-GB rate.

Render's own claim — 75% of paying customers see costs decrease or stay flat — is consistent with both halves of this: most paying accounts have more than one seat and enough volume for the rate cut to dominate. The 25% left holding a bigger bill are disproportionately small teams and free-tier hobbyists whose usage sat in that now-gone 5GB(or 25GB)-to-100GB gap.

Both Numbers vs Owning the Box​

Take the two Render numbers above — $8.25/month for the hobby blog, $106.25/month for the 3-seat media app — and put them next to what the same workload costs self-hosted on a single Hetzner CX22: 2 dedicated AMD EPYC vCPUs, 4GB RAM, 40GB NVMe, and 20TB of outbound transfer included, for $4.59/month. That's roughly double the CPU and RAM of Render's $25/month Standard compute instance, at a fifth of the price, before bandwidth even enters the comparison.

The media app case is the sharper one: $106.25/month on Render's new, cheaper pricing is still north of a 23x multiple over a single Hetzner box that has 50x the bandwidth headroom the workload actually needs. Traffic would have to grow roughly 50x — to 20TB/month — before that box's included transfer runs out at all, and even then overage on Hetzner runs about $1/TB, not $0.15/GB ($150/TB).

Both Render numbers in this table are bandwidth-metered by design; the Hetzner number isn't metered against this workload at all, because 400GB doesn't register as usage worth billing on a 20TB allowance.

What Doesn't Change No Matter How Render Prices Bandwidth​

The mechanism behind April's repricing is worth naming plainly, because it isn't specific to Render: bandwidth on a multi-tenant PaaS is a metered resource, and a metered resource is a dial the vendor controls. Render moved that dial in a direction that happens to help most of its paying customers this time — the seat-fee removal and rate cut are real, and teams should take them. But the dial exists regardless of which way it turns, and the August 1, 2026 forced-migration deadline is the part worth sitting with: workspaces that do nothing get switched to the new terms automatically. That's not a criticism unique to Render — it's what "the vendor owns the pricing model" means in practice, for any hosted platform, any time margin pressure or a strategy shift makes a change worth making.

On owned hardware, that dial doesn't exist. A Hetzner box's 20TB allowance isn't a promotional tier that can be revised at the next pricing review — it's a property of the machine a team already paid for and already controls. Egress on owned infrastructure is a capacity question (do we have enough of the 20TB this month), not a line item that can be repriced out from under a workload that hasn't changed at all.

Bex.co is the open-source, AI-native Render alternative — push a git repo, get a running HTTPS service on machines you own, with no metered egress line item to renegotiate every time a vendor reprices. Star the repo on GitHub or deploy your first app today.

Sources:

• Better pricing for fast-growing teams — Render Blog
• New Workspace Plans — Render Docs
• Updated plans for Render workspaces — Render Changelog
• Outbound Bandwidth — Render Docs
• Render Cloud Pricing Calculator — costgoat.com
• Hetzner Price Adjustment 15 June 2026 — Hetzner Docs