aws
356 articles on aws — what drives the cost, how it is priced, and where the savings actually are.
EKS vs AKS vs GKE cost comparison: what the three managed clusters actually cost
The three big managed Kubernetes services charge similar control-plane fees but diverge sharply on node pricing, load balancers, and free tiers. A like-for-like cluster can differ by 15 to 20 percent across them. Here is the comparison with real numbers.
Kubernetes node pool sizing for cost: big nodes, small nodes, and the waste between
Node size is one of the highest-leverage cost decisions in Kubernetes and most teams pick by habit. Too small and system overhead eats your capacity, too big and a single pod strands half a node. Here is how to size node pools with real numbers.
H100 vs A100 cost per training hour: which GPU is actually cheaper
An H100 costs roughly three times an A100 per GPU-hour, but it can be two to four times faster on modern training workloads. The cheaper GPU is the one with the lower cost per unit of work, not the lower hourly rate. Here is how to do that math.
L4 vs T4 GPU inference cost: the quiet upgrade that pays for itself
The T4 has been the default cheap inference GPU for years, but the L4 costs about 53 percent more per hour while delivering two to three times the throughput on modern models. For inference fleets, that inversion is worth real money.
pgvector vs a managed vector database: the real cost comparison
Adding pgvector to a Postgres instance you already run looks free. A managed vector service starts at several hundred dollars a month. The honest comparison involves index memory, replica sizing, and how much your team wants to operate.
GPU spot interruption economics: when a 70 percent discount is worth the risk
Spot GPUs cost 50 to 70 percent less and can vanish with two minutes of warning. Whether that trade is good depends on one number: how much work you lose per interruption, which is a function of your checkpoint interval.
AI workload data transfer cost: the line item nobody forecasts
Training pulls terabytes from object storage, distributed jobs push gradients across zones, and inference ships embeddings between services. At $0.01 to $0.09 per GB, AI data movement turns into a five-figure line quickly.
Model checkpoint storage cost: the terabytes nobody deletes
Training checkpoints accumulate silently. A team running dozens of experiments can hold hundreds of terabytes of optimizer states they will never load again, at $0.023 per GB-month it adds up to a real recurring bill.
GPU cost per teraflop compared: normalizing hourly rates into value
Hourly rates hide enormous differences in delivered compute. Dividing price by teraflops and by memory bandwidth turns a confusing price list into a ranking you can actually reason about.
SageMaker vs self-managed EC2: what the managed premium actually buys
SageMaker instance rates carry a premium of roughly 15 to 40 percent over the equivalent EC2 instance. Whether that is good value depends on how much of the surrounding platform you would otherwise build yourself.
GPU reserved capacity vs on-demand: committing to hardware that changes yearly
Committing to GPU capacity saves 30 to 60 percent, but GPU generations turn over every 18 to 24 months. Getting the commitment horizon right matters more here than anywhere else in cloud purchasing.
CloudWatch vs third-party APM: comparing the cost models honestly
One bills per metric, log GB, and alarm. The other bills per host and per seat. They cross over at predictable points, and the crossover depends more on your fleet shape than on your traffic. Here is where each wins.
Synthetic monitoring and alerting cost: what every canary and alarm adds up to
Canaries and alarms are individually cheap enough that nobody budgets for them, which is exactly how a few thousand dollars a month appears with no owner. Here is the per-unit math and how it compounds.
Dashboards and queries: the observability cost that runs while nobody watches
An auto-refreshing dashboard is a scheduled job that bills every thirty seconds forever. Most teams have dozens, half of them unopened for months. Here is what they cost.
Audit log cost: the compliance requirement that quietly scales with your API traffic
Audit logs are mandatory, immutable, and billed by volume that grows with every automated call your systems make. Data events are the line that surprises people. Here is how to size and control it.
How much does it cost to run a SaaS app with 1,000 users?
A small B2B SaaS with 1,000 users, a managed database, a cache, and a CDN lands around 195 dollars a month on AWS. Here is the full component-by-component breakdown, the assumptions behind it, and where the money actually goes.
How much does it cost to run a SaaS app with 100,000 users?
A multi-tenant SaaS at 100,000 users, 250 million API requests a month and a 2 TB database costs roughly 6,540 dollars a month on AWS. Here is the full breakdown, and why observability turns out to cost more than the application tier.
How much does it cost to run an e-commerce store on AWS?
An e-commerce store doing 500,000 monthly visits and 25,000 orders costs about 2,190 dollars a month on AWS, which is 9 cents of infrastructure per order. Here is the full breakdown, including why the CDN is a third of the bill.
How much does it cost to run a video streaming service?
A streaming service with a 10,000 hour library and a million viewing hours a month costs about 33,680 dollars on AWS, and 89 percent of that is CDN egress. Here is the breakdown, and why bitrate is the only number that really matters.
How much does it cost to run a mobile app backend?
A serverless backend for a mobile app with 250,000 monthly active users costs about 1,558 dollars a month, and managed authentication is 71 percent of it. Here is the breakdown and the one line worth challenging.
How much does it cost to run a real-time chat app?
A chat app with 10,000 concurrent WebSocket connections and 20 million messages a month costs about 977 dollars on AWS. Here is the breakdown, plus where managed WebSockets stop being cheaper than running your own.
How much does it cost to run an IoT fleet of 100,000 devices?
A fleet of 100,000 devices reporting telemetry every five minutes costs about 1,886 dollars a month on AWS, or 1.9 cents per device. Here is the breakdown, and why message batching is worth more than every other optimization combined.
How much does it cost to run a multiplayer game backend?
A multiplayer game with 200,000 monthly players and 3,000 concurrent at peak costs about 8,325 dollars a month on AWS. Here is the breakdown, including why patch downloads cost more than the game servers.
How much does it cost to run WordPress at scale on AWS?
A WordPress site doing 2 million page views a month costs about 825 dollars on AWS, and full-page caching cuts that to around 650. Here is the component breakdown and why the database is the fragile part.