docs(ins): add faq for gpu billing

pages/instances/faq.mdx

@@ -50,14 +50,27 @@ Instances are billed on a pay-as-you-go basis, and their prices are made up of t
 
 You can change the storage type and flexible IP after the Instance creation, which may affect the monthly billing.
 
+## How are Instances billed?
+
+Instances are billed based on their runtime while powered on:
+- **CPU Instances**: Billed per hour of uptime (including startup and standby time).
+- **GPU Instances**: Billed per minute of uptime (including startup and standby time), except for RENDER-S GPU Instances, which are billed per hour.
+
+<Message type="note">
+  For Kapsule and Kubernetes users: the underlying Instance in a GPU node pool is billed as soon as it is created, even during the setup phase, before the NVIDIA GPU Operator is fully initialized.
+</Message>
+
+Keep in mind that while compute resources follow this usage-based billing model, other resources—such as public IP addresses and storage volumes—are billed for their entire allocation period, regardless of whether the Instance is powered off.
+
+For more details, visit the [Instances pricing page](https://www.scaleway.com/en/pricing/virtual-instances/).
 
 ### What are the prices for Instances?
 
-Refer to the [Pricing page](https://www.scaleway.com/en/pricing/virtual-instances/) to learn about the cost of each Instance type.
+Refer to the [pricing page](https://www.scaleway.com/en/pricing/virtual-instances/) to learn about the cost of each Instance type.
 
 ### What are the storage prices for Instances?
 
-Refer to the [Pricing page](https://www.scaleway.com/en/pricing/storage/) to learn about the storage fees for Instances.
+Refer to the [pricing page](https://www.scaleway.com/en/pricing/storage/) to learn about the storage fees for Instances.
 
 ### How much does a flexible IP cost?