Render Killed the $19 Seat and Started Charging $0.15 per Gigabyte: The Team-Size Crossover Table

On April 23, 2026, Render announced the biggest repricing in its history: the $19-per-member Professional plan was gone, replaced by a $25-per-month flat Pro workspace with unlimited members. On August 1, every remaining legacy workspace was force-migrated. If you run a 30-person team on Render, your seat bill just fell from $570 a month to $25 — and if you push serious traffic, your bandwidth bill quietly became the largest line item you have.

This post works the crossover math all the way through: a 2-, 10-, and 30-person team at 100GB, 1TB, and 5TB of monthly egress, under the old pricing, the new pricing, and a fixed-price Hetzner reference fleet. The headline result up front: above roughly 425GB of monthly egress, Render's workspace-plus-bandwidth line alone costs more than an entire three-node Hetzner fleet — before a single dollar of Render compute.

What Render actually changed (April 23 → August 1, 2026)​

The migration replaced every per-seat plan with a flat workspace subscription. The mapping, from Render's own changelog:

Two more meters changed shape: extra custom domains are $0.25 each per month (15 included on Pro, 2 on Hobby), and extra build pipeline minutes are $5 per 1,000. Compute pricing did not change at all — which is exactly why this comparison can isolate the workspace-plus-bandwidth line honestly.

Render's own explanation is admirably candid: the new Hobby and Pro plans include less bandwidth than their legacy counterparts, Render wrote, "because legacy plans subsidized bandwidth usage with seat fees." Say the quiet part out loud: a 30-person team paying $570 a month in seats was funding everyone else's gigabytes. Remove the seats and the gigabytes have to pay for themselves. That is the whole story, and everything below is just arithmetic.

The crossover table: 2, 10, and 30-person teams at 100GB, 1TB, and 5TB​

Here is the concrete thing the title promises. Each cell is the monthly workspace fee plus bandwidth overage — no compute, both sides, so the delta is apples to apples. The Hetzner column is a reference fleet of three CCX23 cloud servers (4 dedicated vCPU, 16GB RAM each, about $26.60 apiece) plus a load balancer: roughly $85 a month, with 60TB of pooled traffic allowance, which swallows every egress level in this table without an overage cent.

Read it three ways. First, down each column: the new plan beats the old plan in every single cell, because both the seat fee and the effective per-gigabyte rate fell at once ($30 per 100GB block was $0.30/GB; the new meter is half that). Nobody on legacy Professional should mourn it.

Second, across each row: the new Pro rows are identical regardless of team size, which is precisely the point of flat pricing — and precisely why the interesting comparison is no longer old-versus-new but Render-versus-fixed-fleet. At 100GB, new Pro ($36.25) undercuts the $85 fleet comfortably. At 1TB ($171.25), the fleet is already half the price. At 5TB ($771.25), Render's bandwidth line alone costs nine times the entire fleet.

Third, the small-team surprise: a 2-person team at 100GB saves $1.75 a month from this whole migration ($38.00 → $36.25). The headline "seat fees are dead" is worth almost nothing to the team it sounds like it helps most. All of the winnings went to large teams and high-traffic workspaces — and the high-traffic winnings are still dwarfed by what fixed infrastructure costs.

How to read it: seat savings are capped, bandwidth is not​

The one-line rule: every member beyond the second saves you exactly $19 a month, forever capped — while every gigabyte past the 25 included costs $0.15, forever uncapped. Flat pricing turned the team-size variable into a constant and left the traffic variable running.

That asymmetry pins down the only crossover that still matters: the egress level where Render's workspace-plus-bandwidth line exceeds the fixed fleet. Solve $25 + $0.15 × (GB − 25) = $85 and you get roughly 425GB a month. Below that, Pro is the cheaper way to exist; above it, you are paying a premium for managed hosting that grows $150 for every terabyte — before compute, which both sides still need in some form.

Two plugged-in examples make it tangible. A 10-person SaaS team pushing 600GB a month (a busy marketing site plus an API with mid-size payloads) pays $25 + $86.25 = $111.25 on new Pro versus ~$85 for the fleet — close enough that managed deploys, autoscaling, and zero ops hiring win the argument. The same team at 3TB (add user uploads, PDF exports, or a chatty mobile client) pays $471.25 — five and a half fleets — and the "but we don't want to run servers" argument now has a $4,600-a-year price tag attached to it. That is the number to take into the build-versus-buy meeting.

One more cell worth computing, because someone will ask: should a high-traffic team buy Scale for its 1TB of included bandwidth? At 5TB, Scale costs $499 + $0.15 × 4,000 = $1,099 — worse than Pro's $771.25. Scale buys SSO, HIPAA eligibility, and multi-workspace management; it has never bought cheap gigabytes. Don't upgrade for bandwidth.

The other side of the table: what the fixed-price fleet actually costs​

The Hetzner column deserves its own audit, because "$85 flat" invites suspicion. The reference build is three CCX23 instances at €24.49 each (about $26.60 at recent rates) plus a small load balancer, landing near $85 a month. Each server includes 20TB of traffic, pooled across the project — 60TB total before any meter runs. Overage beyond that is about €1 per TB, roughly $0.0011 per GB: around one one-hundred-twentieth of Render's $0.15.

That ratio is the entire economics of this post in one number. Moving a terabyte costs about $150 on Render and about $1.10 on Hetzner overage — and $0 if you're inside the pooled allowance, which at 5TB a month you are twelve times over. The fleet column stays flat through this table and keeps staying flat through 10TB, 20TB, and beyond; Render's column climbs $150 per terabyte as far as the eye can see.

Dedicated hardware sharpens the point further: an AX42-class box (8-core Ryzen, 64GB DDR5 ECC) rents for roughly €46–77 a month depending on configuration, also with a 20TB allowance. Two of those replace the three cloud VMs for most PaaS-shaped workloads at a similar or lower total. Either way, the shape is the same: fixed money, traffic essentially free, versus metered money where traffic is the bill.

What the table doesn't include (honest caveats)​

A crossover table that hides its exclusions is a sales deck, so here are all of them. First, Render compute is excluded from both sides — deliberately, since it didn't change — but it is not small: a Standard web service (2GB RAM, 1 CPU) is $25 a month per instance, and a real production topology of web plus worker plus Postgres plus Redis adds hundreds before bandwidth enters the chat. The fleet side needs its own compute too, but it's already inside the $85.

Second, ops labor is unpriced. The $85 fleet assumes someone — you, a platform engineer, or a Cluster API-driven PaaS layer — provisions nodes, applies upgrades, and answers the 3 a.m. page. For a team with no one willing to own that, Render's premium is genuinely buying something. Price it honestly: even a quarter of an engineer's time dwarfs the bandwidth delta at 1TB, and stops mattering around 5TB.

Third, geography and timing. Hetzner's 20TB allowances describe its EU regions; US cloud allowances are smaller, so an all-US deployment should re-check the pooling math. Hetzner also raised prices on some servers and cloud plans in its June 15, 2026 adjustment (new orders only), so treat every euro figure here as a snapshot to re-verify, not a quote.

Fourth, there is a standard mitigation this table ignores: putting Cloudflare (or any CDN with generous free egress) in front of Render absorbs the cacheable share of traffic before Render meters it. Teams whose gigabytes are mostly static assets and repeat image loads can cut metered egress dramatically without moving anything. The table's egress figures are origin egress as Render bills it — if a CDN takes 80% off the top, re-run the numbers with the remainder.

Conclusion: who the new pricing is actually for​

The migration created three clear winners and one clear signal. Large teams with modest traffic won the most: a 30-person team at 100GB went from $570 to $36.25, a 94% cut. High-traffic teams on legacy plans won somewhat: per-gigabyte overage halved, even as included bandwidth shrank. Small, low-traffic teams got a token improvement and a simpler bill.

The signal is for everyone above ~425GB a month: your hosting decision is now, more than ever, a bandwidth decision. Seat count no longer moves the number; traffic does, at $150 a terabyte, against infrastructure where traffic is functionally free. Render made the right call for fast-growing teams — flat fees scale with headcount the way per-seat never could. But in the same stroke it made the meter visible, and once you can see $150-per-terabyte next to $1-per-terabyte, you can't unsee it.

Running the numbers and finding your egress line costs more than the fleet underneath it? Bex.co is the open-source, AI-native Render alternative — push a git repo, get a running HTTPS service on machines you own, where the bandwidth allowance is measured in terabytes, not gigabytes. Star the repo on GitHub or deploy your first app today.