Dear Railway Sales / Support Team,

I hope this message finds you well. I am currently evaluating Railway as a potential infrastructure platform for our organization and have a few questions regarding your billing model and enterprise-level capacity. I would greatly appreciate your clarification on the following topics.

Part 1: Billing Model Clarification — Actual Usage vs. Provisioned Resources

I am running a persistent service on Railway that operates 24/7. Upon reviewing my billing dashboard, I noticed that the total cost is surprisingly low. Here is a summary of my current billing details (screenshot attached):

• Memory Usage: 213.00 minutely GB at $0.000231 / GB / Minute (Cost: $0.0493)
• CPU Usage: 13.76 minutely vCPU at $0.000463 / vCPU / Minute (Cost: $0.0064)
• Network Egress: 1.74 total GB at $0.05 / GB (Cost: $0.0870)

Given that this service runs continuously (approximately 43,200 minutes per month for a 30-day cycle), I would like to confirm the following:

1. Is Railway billing based strictly on actual CPU and memory consumption, rather than the provisioned or configured resource limits? For example, if I configure a service with a 1 vCPU / 1 GB RAM limit, but the service only consumes 5 MB of RAM and 0.0003 vCPU on average while idling, am I billed only for that actual micro-consumption?
2. This billing model is very attractive compared to traditional cloud providers that charge based on provisioned instance size regardless of actual utilization. Could you confirm that this is indeed how Railway operates — i.e., idle or low-utilization services incur minimal cost even when running 24/7?

Part 2: Enterprise Plan — Capacity and Scale for Large-Scale Deployment

Our organization is considering a large-scale deployment scenario and would like to understand the capacity of Railway's Enterprise plan. Specifically:

1. Can the Enterprise plan support 100,000+ (10W+) persistent, always-on services running concurrently? We are evaluating Railway for a use case that requires a very high number of long-running microservices or worker processes.
2. What are the practical limits on the number of services, projects, and environments under an Enterprise account? Your documentation mentions up to 2,400 vCPU and 2.4 TB RAM limits, but we would like to understand the limits in terms of service count and concurrent deployments.
3. For deployments at this scale (100K+ services), would you recommend Railway-managed infrastructure, dedicated VMs, or the Bring Your Own Cloud (BYOC) option? We would appreciate guidance on which hosting model best suits ultra-high-density deployments.
4. Is custom pricing or a committed spend agreement available for this level of usage? We are open to discussing volume discounts or reserved capacity arrangements.

We are very impressed with Railway's developer experience and usage-based pricing model, and we look forward to exploring a deeper partnership. Please let me know if a call or meeting would be more appropriate to discuss these topics in detail.

Thank you for your time and assistance.

Best regards,