Contabo Alternatives: When Cheap RAM Stops Being Enough

Contabo’s pitch is not marketing. Around $7/month buys 4 vCPU and 8GB of RAM, and nothing else in the mainstream market comes close on that axis.

That’s a genuine gap, and anyone telling you Contabo is a scam hasn’t looked at it. The real question is what pays for the difference, and whether your workload is the kind that finds out.

What we can and can’t tell you

I want to be direct about the evidence here, because this keyword is full of confident performance claims and most of them are unsourced.

We have not benchmarked Contabo. We don’t have accounts across these providers, we haven’t run sysbench or fio, and we won’t publish numbers we didn’t measure — particularly in an article that stands to profit from steering you away from a provider we don’t earn from.

What we can point at is published third-party benchmarking. VPSBenchmarks runs sysbench and fio suites across providers, and the pattern reported there is that Hetzner’s storage measures faster and, more importantly, more consistently under mixed workloads — with Contabo’s NVMe described as hitting high IOPS at high queue depth but dropping sharply when the workload is mixed rather than uniform.

Treat that as a signal from someone else’s measurements, not ours.

The architectural facts, which don’t need a benchmark

Two things are published product properties rather than performance claims, and they’re more useful than any IOPS number.

Contabo’s vCPUs are shared. So are most providers’ at this price — Hostinger’s, DigitalOcean’s Basic droplets, Hetzner’s CX and CPX lines. Shared means contended: you get a portion of a core that varies with what your neighbours are doing. The alternative is a dedicated-vCPU product, which Hetzner sells as its CCX line and which costs considerably more precisely because nothing is shared.

If your workload runs continuously, contention compounds. If it’s idle most of the time, you may never notice.

Port speed is specified, and it’s lower. Contabo’s VPS 10 tier publishes a 200 Mbit/s port, against 1 Gbps as the norm elsewhere. That’s a fivefold difference in ceiling, published on the plan rather than inferred. For a backup target or a low-traffic app it’s irrelevant. For anything serving media it’s the constraint you’ll hit first, and no amount of RAM helps.

Which workloads expose it

The honest framing isn’t “Contabo is slow.” It’s that cheap shared infrastructure is priced for a duty cycle, and some workloads have the wrong one.

Keep Contabo for:

• Storage and backup targets. Enormous disk for the money, and the workload is bursty by nature.
• Development and staging boxes. Idle most of the day, nobody’s revenue depends on latency.
• Batch jobs where wall-clock doesn’t matter. If it finishes overnight, it finishes.
• Anything memory-hungry and compute-light. This is Contabo’s genuine sweet spot — you’re buying the RAM, and the RAM is real.

Move for:

• Continuously-running automation. n8n on schedules or an agent polling channels runs constantly, which is exactly the pattern that turns contention into latency.
• Browser automation. Headless Chromium is both CPU-hungry and bursty, a bad combination on a contended core.
• Database-backed apps with real users. Mixed read/write under concurrency is the workload third-party benchmarks report as least consistent.
• Anything serving media. The 200 Mbit/s port decides this before any other consideration.

What to move to

Ordered by how far you’re moving from Contabo’s position rather than by preference.

Hostinger is the smallest step and the best value on this list — 8GB at $14.99 renewal, roughly twice Contabo’s per-gigabyte rate rather than five or seven times. You keep most of the price advantage, gain a control panel and one-click templates, and its promotional rate of $8.79 narrows the gap further if you’ll commit to the term.

UltaHost sits nearby with more cores per plan than most at this price, and unusually its 2-year discounted rate is also its renewal rate rather than a step-up — $15.87 for 4GB with 3 vCPU.

IONOS is the value pick on cores specifically: four vCPU on a 4GB plan at $11, the best price per vCPU in our VPS index. If contention is your complaint, more cores for the money is a direct answer. Watch the three-month promotional window, which is the shortest in the market.

Cloudzy is the option if you want to test the theory without committing — month-to-month, no term, nothing stranded if you go back.

Hetzner is where you go if you’ve concluded the problem is contention itself. Its CCX line is explicitly dedicated vCPU, which is the actual fix rather than a faster shared core. Two caveats before planning around it: it repriced its CPX line upward by as much as 2.75× in June 2026, and its cheap CX tier has been showing as unavailable.

DigitalOcean is the far end at $48 for 8GB — 6.9× Contabo’s per-gigabyte rate. What you’re buying is documentation, API predictability and a decade of answers to your problem already written. Its tracked link isn’t working on our end, so we’ve named it without linking.

The test worth running yourself

Since neither we nor most articles on this keyword have measured your workload, the useful move is to measure it rather than read about it.

Watch steal time. On any Linux box, top and vmstat report a st column — the percentage of time your vCPU wanted to run and couldn’t because the physical core was busy elsewhere. Sustained non-trivial steal is contention, measured on your machine with your neighbours, which is the only measurement that describes your situation.

If steal time is near zero under your real workload, the performance argument in this article doesn’t apply to you and you should keep the cheaper server. If it isn’t, you now have a number rather than an opinion — and hourly-billed providers make testing the alternative cheap.

How we checked this

Contabo’s pricing is from secondary sources, not read at source: its site returned a 403 to every request we made, from two different networks and via a browser. The figures we’ve quoted — Cloud VPS from roughly $4.95–6.99/month for 4 vCPU and 8GB RAM, a VPS 10 tier at about $5.50 with a 200 Mbit/s port, and plans scaling to 64 vCPU and 192GB — are consistent across published reviews but we could not verify them ourselves. Check Contabo’s own site before relying on any of them.

Comparison pricing is verified: Hostinger’s KVM 2 at $14.99 renewal, IONOS’s VPS L+ at $21, Hetzner’s CPX32 at €38.87 including its IPv4 charge, and DigitalOcean’s 8GB Basic droplet at $48 — all read at source in August 2026 for our VPS price index.

A correction to this article’s brief. It described Contabo as offering “roughly 12× the RAM per dollar of premium hosts.” At the 8GB tier the actual multiples are 6.9× against DigitalOcean, 5.6× against Hetzner’s CPX line and 2.1× against Hostinger — still a large advantage, and a materially smaller one than claimed. The Hostinger comparison in particular changes the recommendation, because a 2.1× gap is one many people would pay to close.

On benchmarks. The brief called for a table of sysbench CPU, fio random write and steal-time measurements, and noted that real benchmarks were needed for credibility. We have not run them and have not published a substitute. The storage-consistency pattern described above is attributed to VPSBenchmarks’ published testing, and we’d encourage reading their results directly rather than taking our summary of them. Inventing a benchmark table would have been the easy way to write this article and the wrong one — especially given that Contabo is the only provider named here that we have no commercial relationship with.

What we did not do: no measurements of any kind, no accounts on Contabo, no comparison of delivered performance. The recommendations above are reasoning from published product architecture — shared versus dedicated vCPU, and a published port speed — plus third-party benchmarks we’ve attributed rather than replicated.

The provider links above are affiliate links. Contabo is not one of them, which is precisely why this article says to keep using it for the workloads where it’s the right answer.

FAQ

Is Contabo actually cheap or is there a catch?

The price is real — around $7 for 4 vCPU and 8GB, roughly 6.9× the RAM per dollar of DigitalOcean. The trade is shared vCPU contention and, on some tiers, a published 200 Mbit/s port rather than 1 Gbps.

Why is my Contabo VPS slow?

Check steal time first, via the st column in top or vmstat. Sustained non-trivial steal means CPU contention rather than a fault. Also check whether you’re saturating a 200 Mbit/s port if your tier has one.

What is the best Contabo alternative?

Hostinger is the smallest step at $14.99 for 8GB — about twice Contabo’s per-gigabyte rate rather than five or seven times. Hetzner’s CCX line is the answer if contention specifically is your problem, because it’s dedicated vCPU rather than shared.

Contabo vs Hetzner — which is better?

Contabo wins decisively on RAM per dollar. Published third-party benchmarking reports Hetzner’s storage as faster and more consistent under mixed workloads. Which matters depends on whether your workload runs continuously or idles.

Should I move my n8n or AI agent off Contabo?

Those are the workloads most likely to expose contention, because they run continuously rather than in bursts. Measure steal time under your real load before deciding — if it’s near zero, stay.

What is steal time and why does it matter?

It’s the percentage of time your virtual CPU was ready to run but the physical core was busy serving someone else. It’s the direct measurement of shared-vCPU contention, and it’s visible on your own server in top.

Is Contabo good for backups?

It’s one of the best options for it. Backup targets want enormous cheap disk and have a bursty duty cycle, which is exactly the shape Contabo’s pricing suits.