AWS, Google Cloud, Microsoft Azure: Who’s leading the game in Generative AI?
2 min readAug 28, 2023
Generative Artificial Intelligence (GAI) has become a crucial field, sparking intense competition among cloud giants Amazon Web Services (AWS), Google Cloud and Microsoft Azure. These companies are deploying considerable resources to dominate this fast-growing sector, putting forward innovative offerings ranging from fundamental models to sophisticated development environments. The choice of cloud platforms for generative AI is strategic, as it requires high computing power and massive datasets. As a result, these providers are engaged in a race to seduce developers and enterprises with cutting-edge services. This article provides an overview of current generative AI services offered by AWS, Google Cloud and Microsoft Azure.
I) Amazon Web Services :
- Amazon SageMaker JumpStart: Provides an environment for accessing, customizing, and deploying machine learning models, with recent support for foundation models.
- Partnerships: Collaborations with Hugging Face and other third parties streamline the fine-tuning and inference processes, making it easier for customers to integrate generative AI capabilities.
- Amazon Bedrock: Still in private preview, this promises a serverless platform to consume foundation models via API, potentially competing with Azure OpenAI.
- Strategic Alliances: Partnerships with GenAI startups like AI21Labs, Anthropic, and Stability.ai enhance AWS’s offerings in text and image-based foundation models.
- Amazon Titan: A proprietary collection of foundation models developed in-house, aimed at boosting various Amazon services like Alexa and Rekognition.
- Future Developments: AWS is planning to introduce commercial foundation models and a dedicated vector database service, potentially as part of Amazon RDS or Aurora.
Google Cloud :
- Google I/O 2023: Significant announcements were made around generative AI, signifying its importance across various Google divisions.
- Vertex AI: Provides access to Google’s four foundation models (Codey, Chirp, PaLM, Imagen) and allows for customization using customer datasets.
- GenAI Studio & Gen App Builder: A playground and no-code tooling to create generative AI-based applications.
- Duet AI: A new assistant feature aimed at accelerating operations on Google Cloud, integrated into various Google Cloud services.
- Vector Database Gap: Currently lacks a native vector database, although extensions like pgvector in Cloud SQL can serve similar purposes.
Microsoft Azure :
- Exclusive Partnership: Azure OpenAI grants exclusive access to most of OpenAI’s foundation models, providing Azure with a competitive edge.
- Azure ML Integration: Allows the use of familiar tools to consume and fine-tune foundation models, tightly integrated with Azure’s managed ML platform.
- Semantic Kernel: An open-source project aimed at simplifying LLM orchestration, offering prompt engineering and augmentation capabilities.
- Vector Database Support: Azure Cosmos DB and Azure Cache for Redis Enterprise have been enhanced to support semantic search, facilitating efficient querying based on semantic meanings.
