Amazon SageMaker AI has introduced usage-based metering for bidirectional streaming, enabling model providers to charge customers based on actual usage. This change affects how the AWS Marketplace software charge is calculated for real-time AI models, aligning billing with consumption patterns.
What happened
Amazon SageMaker AI has introduced usage-based metering for bidirectional streaming, allowing model providers to charge customers based on actual usage. This feature is designed to address the billing challenges associated with real-time speech AI. Customers deploying streaming models on Amazon SageMaker AI bidirectional streaming will now pay for the endpoint infrastructure that hosts the model, while model providers can charge for the software based on actual usage. Deepgram, a real-time speech AI company, is the launch partner for this capability, offering usage-based Speech-to-Text (STT) and Text-to-Speech (TTS) billing on AWS Marketplace. This innovation builds on previous developments between Deepgram and AWS, including bidirectional streaming support and enhanced security and observability features. With usage-based metering, customers can now have a more transparent and accurate billing model for real-time AI workloads.
What to weigh
- Customers pay for software based on actual usage.
- Infrastructure charges remain separate.
FAQ
How does usage-based metering work?
Usage-based metering allows model providers to charge customers based on actual usage, providing a more accurate billing model for real-time AI workloads.
What is the launch partner?
Deepgram is the launch partner for this capability, offering usage-based Speech-to-Text (STT) and Text-to-Speech (TTS) billing on AWS Marketplace.
Are there separate infrastructure charges?
Yes, Amazon SageMaker AI endpoint infrastructure charges still apply separately from the software charge.
Source: Amazon Web Services
