Render's April 23, 2026 Repricing: Flat Fees Land, but Included Egress Drops From 100GB to 5GB

On April 23, 2026, Render tore up its per-seat workspace plans and replaced them with flat monthly fees — Hobby $0, Pro $25, Scale $499. That part reads like a win: no more paying $19 or $29 per teammate just to open the dashboard. The part that doesn't make the headline is what happened to the bandwidth that used to come bundled with those seats. Included outbound egress dropped from 100GB to 5GB on Hobby and from 500GB to 25GB on Pro — a 95% cut on both tiers — with every byte past that billed at $0.15/GB instead of the old $30-per-100GB block rate.

Legacy workspaces have until August 1, 2026 before Render auto-migrates them onto the new plan, whether or not anyone opts in. So the question worth answering before that date isn't "is the new pricing better or worse" in the abstract — it's what your actual monthly bill looks like at your actual traffic level, old plan versus new plan versus just running the same stack on your own box. Below is that recompute, line by line.

What Actually Changed on April 23​

Two structural things are happening at once, and they pull in opposite directions. First, the workspace fee stopped scaling with headcount — a five-person team on the old Professional plan paid $95/month just to exist before touching compute; on the new Pro plan that's a flat $25 no matter how many teammates have dashboard access. Second, the egress allowance that used to be quietly subsidized by those seat fees got cut to a fraction of its old size, and the overage meter got finer-grained ($0.15/GB exact metering instead of $30 blocks you round up into).

Render's own framing is that per-GB billing means "you pay for exactly what you use" — true, but it's true in both directions. It also means there's no more free rounding buffer: burn 6GB on Hobby and you're billed for 1GB over, full stop, where the old plan wouldn't charge you until you cleared 100GB.

The Line-by-Line Recompute​

Here's the part the announcement doesn't spell out: what does this actually cost at the traffic levels a real indie app sees? Take a typical small SaaS — a web service, a background worker, and a Postgres instance — and run the numbers at four usage tiers, from a quiet hobby project up to a genuinely busy one.

Formula: workspace fee + max(0, usage − included) × overage rate, where legacy overage rounds up to the next 100GB block at $30/block, and new overage is metered exactly at $0.15/GB.

Three things fall out of that table that the "included egress dropped 95%" headline alone doesn't tell you:

• Hobby users at real traffic levels usually come out ahead on the new plan. A 5GB free tier sounds punishing, but exact $0.15/GB metering beats $30-per-100GB-block billing once you're actually paying overage — the old plan rounded you up to the next full block, the new one doesn't. At every tier above 50GB, New Hobby is cheaper than Legacy Hobby.
• Pro is the tier that actually got more expensive, and it's not close: a solo developer who was paying $19/month on legacy Professional is now paying $28.75–$171.25/month depending on traffic — 1.5x to 9x more. The flat fee replaced a low per-seat cost, but the collapse from 500GB to 25GB included egress outweighs that savings for anyone below a handful of seats.
• Scale's $499 flat fee only wins at genuinely large traffic. At 1TB/month it's already 3x New Pro's bill, and its advantage only shows up well past that — it's not a tier a typical indie app should be reaching for on bandwidth grounds alone.

The team that benefits cleanly is the multi-seat one: a five-person crew on legacy Professional was paying $95/month in seat fees before a single byte of bandwidth. That group drops straight to $25/month flat. Everyone else is now paying for egress that used to be bundled in.

What the Same Stack Costs on a Box You Own​

There's one more comparison point worth running: the same web+worker+Postgres stack on a Hetzner box, where — unlike a managed PaaS — bandwidth isn't a metered line item at all for workloads this size.

Three Hetzner CX22 instances (2 vCPU / 4GB RAM / 40GB disk each — enough to separate the web process, the worker, and Postgres onto their own machines) run €3.79/month each in EU regions, or about €11.37/month total (~$12.30 USD). Each instance includes 20TB of outbound traffic — 60TB combined across the three-node stack — with overage priced at €1/TB (~$0.001/GB) if you somehow blow past that. For the usage tiers in the table above, bandwidth cost on Hetzner is $0 at every single one; 1,000GB doesn't even register against a 20TB-per-node allowance.

That gap widens with traffic because Render's egress is metered against a fixed base fee while Hetzner's is effectively unmetered at this scale — the more bandwidth an app uses, the more the comparison favors owning the box. The trade you're making isn't compute-for-compute — Render's fee buys you a managed control plane, zero-downtime deploys, and no server to patch — it's specifically the egress line item that inverts once traffic climbs past a hobby project.

Who Actually Gets Hurt (and Who Doesn't)​

• Hurt most: Pro-tier apps that push media — image galleries, video, large API payloads, anything client-heavy. These are exactly the workloads that used to lean on the 500GB Professional allowance and now hit the 25GB Pro cap in days, not weeks.
• Barely notice: quiet hobby projects and internal tools. The dollar amounts at 50-150GB/month are small in absolute terms even if the percentage increase looks alarming on paper.
• Come out ahead: teams of three or more that were paying full per-seat Professional or Organization fees just for dashboard access. The flat fee is a real win for headcount, just not for bytes.

What to Do Before August 1​

Legacy workspaces auto-migrate on August 1, 2026 regardless of whether anyone clicks "opt in" — so this isn't a decision that can be deferred indefinitely. Three practical paths, not mutually exclusive:

1. Put a CDN or object storage in front of anything media-heavy. Serving images/video from R2 or S3-compatible storage instead of your Render service's own egress moves that traffic out of Render's metered path entirely — often the single biggest lever on the new Pro cap.
2. Move the compute, keep Render for what it's still good at. Bandwidth-heavy workers or static-asset-serving services are the first candidates to relocate to a box with real included traffic; keep the pieces where Render's managed deploys and zero-downtime rollouts are worth the premium.
3. Self-host the whole stack. If the calculus above tips far enough — and for a media-heavy Pro app it tips fast — running your own infrastructure removes the egress line item altogether, at the cost of owning the ops that a PaaS used to absorb.

That third path is where a git-push PaaS that runs on machines you own, not a vendor's metered fleet, earns its keep — you keep the deploy experience without the per-GB bill scaling against you as traffic grows.

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 recompute every time a pricing page changes. Star the repo on GitHub or deploy your first app today.

Sources:

• Render: New Workspace Plans
• Render: Outbound Bandwidth
• Render Changelog: Updated plans for Render workspaces
• Render Blog: Better pricing for fast-growing teams
• Hetzner Cloud Pricing
• Hetzner Docs: Traffic