DigitalOcean pricing: predictable monthly caps, droplets starting at $4/month

If you've ever opened a hyperscaler invoice and tried to figure out why the number at the bottom doesn't match the number in your head, you already understand the appeal of DigitalOcean. Their pitch is the opposite of mystery math: flat monthly caps, per-second billing, and price lists that fit on a single page. This is a walkthrough of how that pricing actually works in 2026 — what each plan costs, where the small fees hide, and which tier makes sense for the kind of work you're probably doing.

The short version of how DigitalOcean pricing works

DigitalOcean bills on a per-second basis for compute, with a 60-second minimum (or $0.01, whichever is higher). Effective January 1, 2026, that per-second model applies to Droplets too, which is genuinely useful for short-lived workloads like batch jobs, CI test runners, and ephemeral preview environments — you stop paying the second you power the thing down.

The other thing worth knowing up front: every Droplet ships with a generous chunk of free outbound transfer (starting at 500 GiB/month on the smallest plan), inbound bandwidth is always free, and overage on outbound is a flat $0.01 per GiB. There are no per-request fees on the storage side, no surprise "data processing" line items, and no premium for using multiple regions.

If you want to poke at the live numbers yourself before reading further, 👉 check the current DigitalOcean pricing and sign up — new accounts are typically eligible for a $200 credit valid for 60 days, which is enough to run a mid-tier Droplet for the full trial window without touching your card.

Droplets: the five plan families, explained

Droplets are DigitalOcean's name for virtual machines, and they're split into five families based on what resource ratio they optimize for. Prices below are the verified monthly caps as of the January 2026 per-second billing update.

Basic Droplets — the $4 entry point

Basic Droplets are the cheapest path in and the plan most people start with. CPU here is shared, which makes them best for bursty workloads that don't need dedicated threads — small web apps, personal projects, dev boxes, a low-traffic WordPress site.

512 MiB / 1 vCPU / 500 GiB transfer / 10 GiB SSD — $4/month
1 GiB / 1 vCPU / 1,000 GiB transfer / 25 GiB SSD — $6/month
2 GiB / 1 vCPU / 2,000 GiB transfer / 50 GiB SSD — $12/month
2 GiB / 2 vCPUs / 3,000 GiB transfer / 60 GiB SSD — $18/month
4 GiB / 2 vCPUs / 4,000 GiB transfer / 80 GiB SSD — $24/month
8 GiB / 4 vCPUs / 5,000 GiB transfer / 160 GiB SSD — $48/month
16 GiB / 8 vCPUs / 6,000 GiB transfer / 320 GiB SSD — $96/month

If you're not sure which to pick, the $6 plan is the sweet spot for "I just need a Linux box on the internet." The $24 plan (4 GiB / 2 vCPU) is where you land for a small production web app with a database on the same box.

👉 Start with a Basic Droplet from $4/month

General Purpose Droplets — dedicated CPU, balanced ratio

General Purpose plans give you dedicated vCPUs with a balanced 4:1 memory-to-CPU ratio. This is the workhorse tier for production workloads that need predictable performance — an app server under steady load, a mid-traffic API, a CI runner that actually has to finish on time.

8 GiB / 2 vCPUs / 4,000 GiB transfer / 25 GiB SSD — $63/month
16 GiB / 4 vCPUs / 5,000 GiB transfer / 50 GiB SSD — $126/month
32 GiB / 8 vCPUs / 6,000 GiB transfer / 100 GiB SSD — $252/month
64 GiB / 16 vCPUs / 7,000 GiB transfer / 200 GiB SSD — $504/month
128 GiB / 32 vCPUs / 8,000 GiB transfer / 400 GiB SSD — $1,008/month
160 GiB / 40 vCPUs / 9,000 GiB transfer / 500 GiB SSD — $1,260/month

The Premium variant of General Purpose Droplets adds up to 10 Gbps outbound networking and NVMe SSDs.

CPU-Optimized Droplets — 2:1 memory-to-CPU, fast dedicated cores

These are tuned for workloads that lean hard on the processor: media transcoding, game servers, real-time analytics, CI build fleets. The dedicated vCPUs run at 2.6 GHz or higher.

4 GiB / 2 vCPUs / 4,000 GiB transfer / 25 GiB SSD — $42/month
8 GiB / 4 vCPUs / 5,000 GiB transfer / 50 GiB SSD — $84/month
16 GiB / 8 vCPUs / 6,000 GiB transfer / 100 GiB SSD — $168/month
32 GiB / 16 vCPUs / 7,000 GiB transfer / 200 GiB SSD — $336/month
64 GiB / 32 vCPUs / 9,000 GiB transfer / 400 GiB SSD — $672/month
96 GiB / 48 vCPUs / 11,000 GiB transfer / 600 GiB SSD — $1,008/month

Memory-Optimized Droplets — 8 GiB RAM per vCPU

For workloads that would otherwise swap themselves to death: in-memory caches (Redis, Memcached), large relational databases, search indexes, JVM apps with heap-heavy configurations. These all use NVMe SSDs.

16 GiB / 2 vCPUs / 4,000 GiB transfer / 50 GiB SSD — $84/month
32 GiB / 4 vCPUs / 6,000 GiB transfer / 100 GiB SSD — $168/month
64 GiB / 8 vCPUs / 7,000 GiB transfer / 200 GiB SSD — $336/month
128 GiB / 16 vCPUs / 8,000 GiB transfer / 400 GiB SSD — $672/month
192 GiB / 24 vCPUs / 9,000 GiB transfer / 600 GiB SSD — $1,008/month
256 GiB / 32 vCPUs / 10,000 GiB transfer / 800 GiB SSD — $1,344/month

Storage-Optimized Droplets — big NVMe local disks

Built for jobs where local disk throughput is the bottleneck: large databases, analytics on local datasets, Elasticsearch clusters, anything that doesn't want to round-trip through block storage. The local NVMe can be an order of magnitude faster than regular SSDs.

16 GiB / 2 vCPUs / 4,000 GiB transfer / 300 GiB NVMe — $131/month
32 GiB / 4 vCPUs / 6,000 GiB transfer / 600 GiB NVMe — $262/month
64 GiB / 8 vCPUs / 7,000 GiB transfer / 1,170 GiB NVMe — $524/month
128 GiB / 16 vCPUs / 8,000 GiB transfer / 2,340 GiB NVMe — $1,048/month
192 GiB / 24 vCPUs / 9,000 GiB transfer / 3,520 GiB NVMe — $1,572/month
256 GiB / 32 vCPUs / 10,000 GiB transfer / 4,690 GiB NVMe — $2,096/month

👉 Compare all Droplet families and pick a plan

Beyond Droplets: the other products you'll probably use

A real stack rarely stops at a VM. Here's how the rest of the DigitalOcean catalog is priced.

App Platform — PaaS for when you don't want to SSH anywhere

App Platform is the managed PaaS layer: push code, get a URL. There's a free tier for up to three static sites, and component-based pricing for everything else starting at $5/month. You pay for the components you run (web services, workers, databases), and autoscaling means you only pay for the instances actually serving traffic. It's the right choice if "manage a Linux server" isn't a sentence you want in your weekend.

👉 Try App Platform — free for static sites, from $5/month for apps

Managed Kubernetes (DOKS) — $12/month base, $40/month for HA

DOKS gives you a managed Kubernetes control plane for a flat $12/month per cluster. If you want the high-availability control plane (recommended for anything production), it's $40/month. Node pricing is just regular Droplet pricing — a basic 2 vCPU / 4 GiB node runs about $24/month. Compared to the hyperscalers, who typically charge around $73/month per cluster just for the control plane, this is one of the more aggressively priced items in the catalog.

Spaces Object Storage — $5/month flat for 250 GiB + 1 TiB egress

Spaces is the S3-compatible object storage product. It's $5/month for 250 GiB of storage and 1 TiB of outbound transfer, with no per-request fees. Inbound bandwidth is always free. For media hosting, backups, log archives, and static asset delivery, the flat pricing is much easier to model than AWS S3's per-GB-and-per-request math.

Block Storage — $0.10 per GiB/month

Volumes (block storage) attach to Droplets and are billed at $0.10 per GiB per month — so 100 GiB costs $10/month. Useful when your Droplet's local SSD isn't big enough but you don't want to resize the whole machine.

Backups and snapshots

Backups have two flavors: percentage-based (20% of the Droplet cost for weekly, 30% for daily) or usage-based plans starting at $0.01/GiB per month with flexible frequency up to every 4 hours and retention from 3 days to 6 months. Snapshots are $0.06/GB per month. Neither is free, but both are cheap enough that "set and forget" is the rational default for anything you'd be sad to lose.

The free stuff that's easy to miss

A few line items that show up at $0 on the invoice and quietly save real money over a year:

Free DNS management, including for records pointing elsewhere
Free cloud firewalls, attachable to any Droplet
Free VPCs, with no charge for intra-datacenter traffic
Free reserved IPs as long as they're assigned to a Droplet
Free container registry (1 repository, 500 MiB storage)
DigitalOcean Functions: 90,000 GiB-seconds per month free, no per-invocation charge
Free monitoring and alerting on Droplet metrics

The free outbound transfer on Droplets is the big one, though. On a $24/month plan you get 4,000 GiB of egress included — that's a lot of web traffic before the $0.01/GiB overage kicks in.

New-user credit: what you actually get

DigitalOcean runs a standing promo for new accounts: $200 in credit, valid for 60 days. The credit applies across the catalog — Droplets, App Platform, Spaces, Kubernetes, the lot. Your card is only charged after the credit is exhausted or the 60-day window closes, whichever comes first. If you spend less than $200 in the window, you simply don't pay anything; if you spend more, the credit covers the first $200 and your card handles the rest.

This is the cleanest way to benchmark a real workload before committing — stand up the architecture you'd actually run, point some traffic at it, and see what the bill looks like under per-second billing.

👉 Claim the $200 / 60-day credit and start building

Which plan should you actually pick?

A few rules of thumb, none of them absolute:

Personal blog, static site, tiny side project — Basic Droplet at $4 or $6/month, or App Platform's free static tier if you genuinely never want to log in.
Small production web app (one box does everything) — Basic at $24/month (4 GiB / 2 vCPU), or step up to General Purpose at $63/month once you want dedicated CPU.
API server under steady load — General Purpose at $126/month and up, depending on traffic.
Database server — Memory-Optimized for relational/search workloads (Redis, Postgres, Elasticsearch), Storage-Optimized if the dataset is large and local-disk-bound.
CI / build fleet / batch jobs — CPU-Optimized, sized to your queue depth. Per-second billing matters most here.
Containerized microservices — DOKS at $12/month base + Droplet-priced nodes, or App Platform if you'd rather skip Kubernetes entirely.

The honest answer for most readers is "start one tier cheaper than you think you need, set a billing alert, and resize up when the alert fires." Resizing a Droplet is a one-click operation with a brief power-cycle, and per-second billing means you're not paying for the old size while you decide.

A note on how this compares

It's worth being straight about the comparison landscape. On raw price-per-vCPU, some European providers (Hetzner being the obvious one) come in noticeably cheaper for similar specs, and the Reddit threads on this are not wrong about the math. Where DigitalOcean earns its premium is in the surrounding stuff: the control panel is genuinely simple, the docs are well-maintained, the billing is predictable, the bandwidth allowances are generous, and there's a datacenter in 14 regions including Bangalore, Singapore, Sydney, London, Frankfurt, and Toronto. If your priority is "cheapest possible compute per dollar" and you're comfortable doing all your own ops, the cheaper providers are real options. If your priority is "I want to ship something this afternoon and understand my bill at the end of the month," that's the gap DigitalOcean is pricing into.

Getting started

The setup is roughly: sign up, drop in a card (a $1 pre-auth will appear and immediately reverse), pick a Droplet or an App Platform component, choose a region close to your users, and you're running in under a minute. The $200 credit covers the first 60 days of experimentation, which is more than enough time to settle on a plan size.

👉 Open a DigitalOcean account and claim the new-user credit

The whole pricing model is built around the idea that you should be able to predict next month's bill from this month's traffic, without a spreadsheet. Whether that holds up for your workload is something the 60-day trial is designed to answer.