Serverless Is More Than Just a Technology
Most engineers think of serverless as a way to simplify operations and improve scalability. While those benefits are real, serverless can also become a powerful business advantage. In this article, I share how we built Structor almost entirely on serverless AWS services, reduced our infrastructure costs to nearly zero, and gained the flexibility to compete on pricing against established players. Sometimes, the biggest benefit of serverless isn't technical, it's strategic.
When people talk about serverless, they usually focus on scalability, reduced operational overhead, or developer productivity.
Those benefits are real.
But after building Structor , I discovered something even more important:
Serverless isn't just a technology decision. It's a business decision.
In some markets, it can become a competitive advantage powerful enough to influence your entire pricing strategy.
And that's exactly what happened to us.
The Product Was Better. But That Wasn't Enough.
Early on, we realized that delivering more value than competitors was only part of the equation.
Many Infrastructure as Code platforms focus heavily on Terraform expertise and infrastructure knowledge.
We wanted to make Infrastructure as Code accessible to everyone. That differentiation was valuable, but a better product alone doesn't guarantee success.
Pricing matters, and that's where serverless changed everything.
How Serverless Changed Our Economics
One of our main competitors starts its first premium plan at approximately 99€/user/month.
Our architecture, however, was built almost entirely on serverless services. As a result, our operational costs remain extremely low.
For long periods of time, the infrastructure cost of serving users is effectively close to $0 per month.
That creates an interesting situation:
We are not forced into a pricing model by infrastructure expenses.
We can choose pricing based on business strategy rather than operational constraints.
If needed, we could charge half the price. Or even less.
Not because we are sacrificing margins, but because our architecture allows it.
This is where serverless becomes much more than a technical architecture pattern.
It becomes a competitive weapon.
Our Architecture
The following diagram illustrates the serverless architecture powering Structor.
Where We Actually Spend Money
The interesting part is not how many AWS services we use. The interesting part is which services generate meaningful costs.
1. AWS WAF
By far our largest fixed infrastructure expense. Unlike many serverless services that scale down to nearly zero, security has a baseline cost. A production-grade application exposed to the internet requires protection.
For us, AWS WAF typically represents the largest recurring infrastructure expense, often around $10–15 per month.
This is one of the few costs that cannot easily be optimized away.
2. AI Token Consumption
Our second major cost category is AI usage. And honestly, this is exactly where we want to spend money. If users consume AI assistance, they are receiving value from the platform.
This means our costs scale with customer usage rather than idle infrastructure. That's a healthy business model.
3. Lambda SnapStart
User experience matters. Cold starts can negatively impact responsiveness. To provide a faster experience, we leverage Lambda SnapStart where appropriate.
This introduces additional costs, but the trade-off is often worth it when measured against improved latency and user satisfaction.
Everything Else Costs Almost Nothing
Most of our platform runs on fully managed serverless services:
- Amazon Cognito
- Amazon SNS
- Amazon SQS
- Amazon DynamoDB
- Amazon S3
- Amazon API Gateway
- AWS Certificate Manager (ACM)
- AWS Systems Manager Parameter Store (SSM)
- Amazon CloudFront
- Amazon SES
- Amazon Bedrock AgentCore
Many of these services benefit from free tiers, pay-per-use pricing, or costs that remain below a few dollars per month during normal operation.
Even workloads that require compute execution are heavily optimized.
For example, we use ECS-based runners inside a VPC, but these are launched only when users invoke specific actions. Because they are not continuously running, their cost remains negligible compared to traditional always-on infrastructure.
Traditionnal architecture comparison
To understand the impact of our architectural choices, let's compare this with a more traditional deployment model.
To deliver a similar experience, we would likely need:
- Two continuously running application servers for high availability
- A managed database instance
- A load balancer
- Monitoring, logging, backups, and storage
- Additional networking components
On AWS, such an architecture would typically cost between $100 and $180 per month, even if nobody used the platform.
For example:
| Component | Approximate Monthly Cost |
|---|---|
| 2 × EC2 instances | $55 |
| Managed database | $15–30 |
| Application Load Balancer | $25–35 |
| Storage, logs, backups | $10–30 |
| Total | $105–150+ |
And that's before accounting for engineering time spent patching servers, managing operating systems, scaling infrastructure, or handling availability concerns.
By contrast, our serverless architecture remains close to $0/month when there is little or no usage.
The difference is simple
- With traditional infrastructure, you pay for capacity.
- With serverless, you pay for value creation.
This means that instead of spending over $1,200–2,000 per year just to keep infrastructure online, we can invest those resources into improving the product, adding new features, and reducing prices for our users.
The Hidden Lesson About Serverless
The biggest misconception about serverless is that it exists primarily to make developers' lives easier.
While operational simplicity is valuable, the larger impact often appears on the business side.
Serverless allows companies to:
- Reduce fixed infrastructure costs
- Align expenses with actual usage
- Launch products with minimal operational overhead
- Experiment with pricing models
- Stay competitive without sacrificing margins
In our case, the architecture directly influences our ability to compete in the market. That is a much larger outcome than simply avoiding server management.
Conclusion
Building Structor reinforced a lesson I didn't fully appreciate when I started working with serverless technologies:
Serverless is not just an architectural pattern.
It changes the economics of a product.
When most of your infrastructure costs scale only when customers generate value, you gain flexibility that traditional architectures rarely provide.
- Better scalability.
- Less operational overhead.
- Lower fixed costs.
And most importantly, the freedom to compete on your own terms.
Sometimes the biggest benefit of serverless isn't technical at all. It's strategic.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article