Scale-to-Zero Laravel: The Economics of $0 Idle Cost
Every traditional hosting platform charges you money while your app sleeps. A VPS costs $5-20/month whether it serves 1 million requests or zero. Laravel Cloud charges ~$20/month minimum. Even "serverless" platforms like Railway have minimum compute charges.
Delt is different. When your Laravel app receives no traffic, you pay nothing. Zero. This article explains how — and why it matters more than you think.
The Problem: Paying for Idle Compute
Consider a typical developer's hosting portfolio:
- Production app: 50K requests/day — needs to be always-on
- Staging environment: 100 requests/day — used during development hours only
- Side project #1: 500 requests/day — weekend traffic only
- Side project #2: 10 requests/week — basically abandoned but still live
- Portfolio site: 50 requests/day — rarely updated
On traditional hosting, each of these costs $5-20/month regardless of usage. That's $25-100/month for apps that are idle 90%+ of the time.
What You're Actually Paying For
| Platform | Minimum Monthly Cost | What You Get When Idle |
|---|---|---|
| DigitalOcean VPS | $6/mo | A server doing nothing |
| Laravel Forge + DO | $12/mo + $6/mo | Two things doing nothing |
| Laravel Cloud | ~$20/mo | A container doing nothing |
| Railway | ~$5/mo | A container doing nothing |
| Delt Hobby | $0/mo | Nothing running = nothing charged |
How Lambda Scale-to-Zero Works
AWS Lambda uses a fundamentally different billing model than containers or VMs:
Traditional Compute (VPS/Container)
Cost = (Server Size) × (Hours Running) × (Price Per Hour)
= Always > $0, even with zero traffic
Lambda (Pay-Per-Invocation)
Cost = (Number of Requests) × (Duration Per Request) × (Memory Allocated) × (Price Per GB-Second)
= $0 when requests = 0
Lambda literally doesn't exist when not handling a request. There's no server, no container, no process waiting for traffic. When a request arrives, Lambda creates an execution environment in ~500ms, handles the request, and then freezes. After ~15 minutes of inactivity, the environment is destroyed entirely.
No requests = no execution environments = no cost.
Delt's Cost Model: Per-Tenant Economics
Hobby Tier — $0/month (Free)
A Hobby tenant with typical side-project traffic (1,000 requests/day):
| Component | Monthly Cost |
|---|---|
| Lambda invocations (30K/mo × 500ms × 512MB) | $0.03 |
| API Gateway (30K requests) | $0.01 |
| S3 storage (100MB) | $0.002 |
| CloudWatch logs | $0.01 |
| Total AWS cost to Delt | $0.05/mo |
For a truly idle app (0 requests): $0.00/mo
Delt offers this free because the cost is negligible. Even 1,000 Hobby tenants with minimal traffic cost Delt only ~$50/month in AWS bills.
Growth Tier — $29/month
A Growth tenant with production traffic (50,000 requests/day):
| Component | Monthly Cost |
|---|---|
| Lambda invocations (1.5M/mo × 300ms × 1GB) | $0.85 |
| API Gateway (1.5M requests) | $0.53 |
| SQS queue processing (~100K jobs/mo) | $0.40 |
| RDS PostgreSQL (shared, allocated portion) | $1.00 |
| S3 storage (5GB) | $0.12 |
| EventBridge (scheduler, 43K invocations) | $0.04 |
| CloudWatch | $0.05 |
| Total AWS cost to Delt | $2.99/mo |
Revenue: $29/mo. Cost: $2.99/mo. Gross margin: 89.7%
Business Tier — $59/month
A Business tenant with high traffic (200,000 requests/day) and provisioned concurrency:
| Component | Monthly Cost |
|---|---|
| Lambda + provisioned concurrency (5 units) | $4.50 |
| API Gateway (6M requests) | $2.10 |
| SQS queue processing (~500K jobs/mo) | $0.80 |
| RDS PostgreSQL (shared, larger allocation) | $2.00 |
| S3 storage (20GB) | $0.46 |
| EventBridge | $0.04 |
| CloudWatch | $0.10 |
| Total AWS cost to Delt | $10.00/mo |
Revenue: $59/mo. Cost: $10.00/mo. Gross margin: 83.1%
Comparison: Delt vs Competitors at Various Traffic Levels
Zero Traffic (Idle App)
| Platform | Monthly Cost |
|---|---|
| Delt Hobby | $0 |
| Delt Growth | $29 (flat) |
| Laravel Cloud | ~$20 |
| Vapor + AWS | ~$15-25 |
| Forge + VPS | ~$12-18 |
10K Requests/Day (Side Project)
| Platform | Monthly Cost |
|---|---|
| Delt Hobby | $0 |
| Delt Growth | $29 (flat) |
| Laravel Cloud | ~$25-35 |
| Vapor + AWS | ~$30-45 |
| Forge + VPS | ~$12-18 |
100K Requests/Day (Production App)
| Platform | Monthly Cost |
|---|---|
| Delt Growth | $29 (flat) |
| Delt Business | $59 (flat, zero cold starts) |
| Laravel Cloud | ~$50-80 |
| Vapor + AWS | ~$60-100 |
| Forge + VPS | ~$24-48 |
The key insight: Delt's flat pricing means costs are predictable regardless of traffic spikes. A viral moment doesn't generate a surprise $500 AWS bill.
Addressing Cold Start Concerns
The trade-off for $0 idle cost is a cold start on the first request after inactivity. Let's be honest about this:
Cold Start Timeline
- 0-500ms: Lambda creates execution environment, loads PHP runtime
- 500ms-1s: Laravel bootstraps (service providers, config loading)
- 1-1.6s: Worst case for large applications with many service providers
When Cold Starts Happen
- After ~15 minutes of zero traffic (Lambda freezes the environment)
- After a new deployment (new code = new environment)
- During traffic spikes that exceed current warm capacity
Why 500ms Is Acceptable
For context:
- Average web page load time: 2.5 seconds
- Average API response time: 200-500ms
- User-perceptible delay threshold: 100ms
- User-annoying delay threshold: 1 second
A 500ms cold start is within normal web response times. Users won't notice it. And it only happens on the first request — subsequent requests are served in 5-15ms (faster than any container).
Eliminating Cold Starts (Business Plan)
For applications that cannot tolerate any cold start (payment processing, real-time APIs), Delt's Business plan includes provisioned concurrency. This keeps Lambda environments warm 24/7:
- 0ms cold start — environments are always ready
- Cost: included in the $59/mo flat fee
- Configurable: 1-10 concurrent warm environments
The Auto-Hibernate Model (Hobby Plan)
Hobby plan projects auto-hibernate after 30 minutes of inactivity:
- Active: App receives traffic normally (Lambda warm)
- Idle 30 min: Delt marks project as hibernated
- Next request: 500ms-1.6s cold start, then normal operation
- Resume: Project is active again until next 30-min idle period
This is perfect for:
- Side projects you want to keep live without paying
- Staging environments used only during work hours
- Demo apps that need to be accessible but rarely used
- Portfolio sites with sporadic traffic
Why Flat Pricing + Scale-to-Zero Is the Best Model
For Developers
- Predictable costs: $29/mo or $59/mo, period. No surprise bills.
- Free side projects: Hobby tier costs nothing when idle
- No capacity planning: Don't guess how much traffic you'll get
For Delt (the Business)
- High margins: 83-90% gross margin at scale
- Low floor: Near-zero cost for inactive tenants
- Aligned incentives: We want your app to succeed (more traffic = same price for you, same cost for us)
vs Usage-Based Pricing (Laravel Cloud, Vapor)
Usage-based sounds fair but creates anxiety:
- "Will this traffic spike bankrupt me?"
- "Should I add caching to reduce Lambda invocations?"
- "Is my staging environment costing too much?"
Flat pricing eliminates this cognitive overhead. Build your app, don't worry about the bill.
FAQ
What happens during a traffic spike?
Lambda auto-scales to handle concurrent requests. Your flat monthly price doesn't change. Delt absorbs the additional AWS cost — that's what the 83-90% margin is for.
Is there a request limit on flat pricing?
Growth plan includes up to 5 million requests/month. Business includes up to 20 million. These limits are generous enough that 99% of applications never hit them. If you do, we'll reach out to discuss Enterprise pricing.
Can I run multiple apps on the free tier?
Yes. Hobby tier allows unlimited projects. Each project auto-hibernates independently. You could have 10 side projects all at $0/month — they only cost when they receive traffic.
What about database costs on the free tier?
Hobby tier uses SQLite (file-based, stored in Lambda's ephemeral storage). It's perfect for side projects and demos. Growth+ plans include managed PostgreSQL or MySQL.
Conclusion
Scale-to-zero isn't just a cost optimization — it's a fundamental shift in how developers think about hosting:
- Old model: "I need a server running 24/7, even for my side project"
- New model: "My app exists when needed, disappears when not, costs nothing when idle"
Delt brings this model to Laravel with zero configuration. Push your code, pay nothing when idle, pay a flat fee when active. No surprises, no anxiety, no wasted money.
Try scale-to-zero Laravel hosting — Deploy free, pay only when you're ready →