Fly.io vs AWS for Solo Developers

Quick Comparison

Fly.io Overview

Fly.io deploys Docker containers as Firecracker micro VMs on edge servers across the globe. The entire model is "deploy once, run everywhere." Your app can be live in multiple regions with a single config file and a fly deploy command. The CLI handles everything from provisioning to scaling to secret management.

For solo developers, Fly.io removes the operational overhead of managing infrastructure while keeping you close to the metal. You get persistent volumes, managed Postgres, private networking between services, and LiteFS for distributed SQLite. The free tier includes 3 shared VMs, which handles hobby projects nicely.

The learning curve is moderate. You need Docker knowledge and terminal comfort, but the concepts are straightforward. Deploy, scale, done.

AWS Overview

AWS is the everything cloud. Over 200 services covering compute, storage, databases, AI, IoT, and pretty much anything else you can think of. For app hosting specifically, the relevant services are EC2 (VMs), ECS/Fargate (containers), Lambda (serverless), Lightsail (simplified VPS), Elastic Beanstalk (PaaS), and App Runner (managed containers).

The problem for solo developers is obvious. AWS was built for enterprises with dedicated DevOps teams. The console is overwhelming. The pricing is a spreadsheet. The documentation is exhaustive but dense. I've seen experienced developers accidentally run up $500 bills because they forgot to turn off a test instance.

That said, AWS is the most capable cloud platform on the planet. If you need something specific, AWS probably has a managed service for it. And the 12-month free tier is genuinely generous, including a t2.micro EC2 instance, RDS database, S3 storage, and Lambda invocations.

Key Differences

Simplicity. Fly.io takes minutes to go from zero to deployed. AWS can take hours just to configure IAM roles, security groups, and VPCs before you deploy anything. For a solo developer, this complexity tax is real. Every hour spent on AWS configuration is an hour not spent building your product.

Service breadth. AWS wins this by an absurd margin. Fly.io does one thing well: deploy containers globally. AWS does everything. Need a CDN? CloudFront. Queue? SQS. Email? SES. Search? OpenSearch. Fly.io will never match this breadth, and it doesn't try to.

Global deployment. Both support multi-region deployment. The difference is effort. Fly.io makes global deployment a one-line config change. AWS multi-region requires configuring load balancers, Route 53 routing policies, and cross-region replication. The end result might be similar, but the effort is vastly different.

Pricing clarity. Fly.io's pricing is simple. Per VM, per GB bandwidth, per GB storage. AWS pricing requires a calculator and a finance degree. EC2 has on-demand, reserved, spot, and savings plans. Each service has its own pricing model. I genuinely cannot predict my AWS bill month to month, and I don't think most solo developers can either.

Lock-in risk. Using AWS-specific services (DynamoDB, SQS, SNS, Lambda) creates significant lock-in. Migrating away from AWS means rewriting parts of your application. Fly.io deploys Docker containers, so moving to another platform that runs Docker is relatively painless.

Career value. Here's the honest truth: AWS experience is valuable on a resume. If you're a solo developer who also freelances or might return to employment, knowing AWS well is a career advantage. Fly.io won't show up on job requirements in the same way.

By the Numbers (2026)

The vibe of each platform is one thing. The receipts are another. Here is what each one actually charges and offers, checked on 2026-05-28.

Fly.io compute. Billing is per-second and only while a Machine runs. The smallest always-on option, a shared-cpu-1x with 256MB RAM, lists at about $0.0028 per hour, which works out to roughly $2.02 per month if you leave it on around the clock. A Performance machine with a dedicated CPU and 2GB RAM lists at about $32.19 per month.

Fly.io storage and bandwidth. Persistent volumes are $0.15 per GB per month of provisioned capacity, charged whether the volume is full or empty. Outbound bandwidth is $0.02 per GB in North America and Europe, $0.04 per GB across Asia Pacific, Oceania, and South America, and $0.12 per GB for Africa and India. Inbound traffic is free. Legacy free-tier organizations still get up to three shared-cpu-1x 256MB VMs and 3GB of total volume storage at no cost.

AWS compute. An EC2 t3.micro (2 vCPU, 1GiB) runs at about $0.0104 per hour on-demand in US East, which is roughly $7.59 per month always-on. RDS, S3, Lambda, and every other service carries its own separate meter.

AWS bandwidth. The first 100GB per month of data transfer out to the internet is free across all services and regions. After that it is about $0.09 per GB for the first 10TB, dropping to $0.085 per GB for the next 40TB.

AWS free tier, the 2025 reset. This is the single biggest thing that has changed since this post first went up, and it matters for solo devs. For accounts created after July 15, 2025, the old 12-month deal (a t2.micro running 24/7, an RDS instance, 5GB of S3) is gone. New accounts now get up to $200 in credits ($100 at sign-up plus up to $100 more for completing setup tasks) across a free plan that lasts up to six months, whichever runs out first. Accounts created before that date keep the legacy 12-month allowances. The "Always Free" services survived the reset: Lambda still gives 1 million requests and 400,000 GB-seconds of compute per month, and DynamoDB still gives 25GB of storage plus 25 read and 25 write capacity units, neither of which expires.

Footprint. Fly.io lists 18 regions worldwide. AWS spans 123 Availability Zones across 39 geographic Regions and advertises more than 240 services. That gap is the whole story of this comparison in two numbers.

Real Cost at Solo-Dev Scale

Numbers in isolation lie. Here is a concrete, realistic solo-dev workload priced on both platforms, so you can see where the money actually goes. The assumptions are stated so you can swap in your own.

The workload. One always-on web app, one Postgres-class database, a 10GB persistent volume, and 200GB of outbound traffic per month. Modest, but more than a hello-world.

On Fly.io. Two shared-cpu-1x 256MB Machines for the app at about $2.02 each gives $4.04. A comparable shared Postgres setup lands in the $5 to $15 range depending on size, call it $10. The 10GB volume at $0.15 per GB is $1.50. Bandwidth at 200GB minus nothing free, all in North America at $0.02 per GB, is $4.00. That is roughly $19.54 per month, and it is predictable because every line item is a flat per-unit rate.

On AWS, new account. A t3.micro at $0.0104 per hour always-on is about $7.59. An RDS db.t3.micro adds roughly $12 to $15 depending on engine and storage, call it $13. EBS storage for 10GB is about $0.80. Bandwidth at 200GB, with the first 100GB free and the next 100GB at $0.09 per GB, is $9.00. That is roughly $30.39 per month once your $200 of credits and six-month window are spent. During the credit window it can be effectively free, which is the real point of the AWS free plan now.

The honest read. At this scale Fly.io is cheaper and far easier to predict, by maybe ten to twelve dollars a month. AWS only pulls ahead on raw cost if you exploit the Always Free tier hard (running on Lambda and DynamoDB instead of an always-on server and RDS) or if you are still inside the six-month credit window. For a solo dev who just wants a server, a database, and a deploy command, the Fly.io bill is both lower and legible. For a solo dev willing to architect around serverless, AWS can run close to free for longer than any Fly.io setup.

When to Choose Fly.io

• You want to deploy globally without an infrastructure team
• Simple, fast deployment is your priority
• You don't need the breadth of 200+ cloud services
• Predictable pricing matters more than maximum flexibility
• You want to avoid the AWS complexity tax entirely

When to Choose AWS

• You need specific AWS services (SQS, DynamoDB, Lambda, etc.)
• Your project will grow to require enterprise-grade infrastructure
• You want the new-account credits (up to $200 over six months) to validate an idea at near-zero cost
• AWS experience matters for your career goals
• You need compliance certifications (HIPAA, SOC2, etc.)

The Verdict

For solo developers, Fly.io is the better hosting platform in almost every practical sense. It's faster to set up, simpler to manage, and cheaper for typical workloads. The global deployment model gives you capabilities that would take days to configure on AWS.

AWS makes sense in specific scenarios. If you're building something that requires SQS, or needs Lambda for event-driven processing, or requires S3 presigned URLs for file uploads, then AWS is the only option. It also makes sense if you're using the new-account credits strategically to run a project at near-zero cost during the six-month window (the old 12-month free tier was retired for accounts created after July 15, 2025).

But let me be direct. Most solo developers who choose AWS are over-engineering. They don't need 240 services. They need a server, a database, and a deploy command. Fly.io gives you exactly that, globally, in minutes. Save AWS for when you actually need what only AWS can provide.

Sources

All figures checked on 2026-05-28.

• Fly.io Resource Pricing (compute per-second and monthly rates, volume $0.15/GB, outbound bandwidth tiers, legacy free org allowance)
• Fly.io Regions (region count)
• AWS Free Tier overview (up to $200 credits, six-month free plan for new accounts)
• AWS Free Tier FAQs (credit structure, Always Free services)
• AWS Lambda Pricing (1 million requests and 400,000 GB-seconds Always Free)
• Amazon DynamoDB Pricing (25GB plus 25 RCU and 25 WCU Always Free)
• AWS EC2 On-Demand Pricing (on-demand model) and t3.micro spec and price (about $0.0104/hr, 2 vCPU, 1GiB)
• AWS Global Infrastructure (123 Availability Zones across 39 Regions)