News

Cloud Giants Offer DeepSeek AI, Restricted by Many Orgs, to Devs

Microsoft, Amazon and Google have all embraced new open-source DeepSeek AI technology from a Chinese company, despite its usage being banned by many organizations.

DeepSeek AI is being restricted worldwide because of data security, privacy, compliance, and national security risks.

"DeepSeek, the Chinese app that sparked a $1 trillion US market meltdown this week, is storing its fast-growing troves of US user data in China -- posing many of the same national security risks that led Congress to crack down on TikTok," said a New York Post article this week. It noted the huge popularity of the new offering: "The artificial intelligence chatbot topped the charts in Apple's App Store and Google's Play Store on Tuesday. DeepSeek has been downloaded more than 2 million times since its debut on Jan. 15, with most coming in the last three days, according to AppMagic."

Also, a Bloomberg article reported DeepSeek AI was restricted by "hundreds of companies" within days of its debut.

The new offering shook up the AI space because of its high performance when compared with other models despite being much cheaper to produce. The crushing costs of the computing power needed to train huge AI models is a primary concern in the industry.

And much of that industry, unlike many other organizations, is largely embracing DeepSeek AI, led by the cloud giants.

Microsoft, Amazon and Google have all made the technology available to developers on their Azure, AWS and Google Cloud platforms. Here's a summary.

Microsoft

The company yesterday announced "DeepSeek R1 is now available on Azure AI Foundry and GitHub" and published guidance on "Using DeepSeek models in Microsoft Semantic Kernel" and just today published "DeepSeek-R1 on Azure with LangChain4j Demo" documentation. Windows developers, meanwhile, were yesterday informed about "Running Distilled DeepSeek R1 models locally on Copilot+ PCs, powered by Windows Copilot Runtime."

And Microsoft-owned GitHub yesterday announced "DeepSeek-R1 is now available in GitHub Models (Public Preview)." In fact, it's the very first option listed:

GitHub Models
[Click on image for larger view.] GitHub Models (source: GitHub).

AWS
Yesterday, AWS CEO Matt Garman announced DeepSeek R1 on the platform in a LinkedIn post:

DeepSeek R1 is the latest foundation model to capture the imagination of the industry. We've always been focused on making it easy to get started with emerging and popular models right away, and we're giving customers a lot of ways to test out DeepSeek AI.

Today, customers can run the distilled Llama and Qwen DeepSeek models on Amazon SageMaker AI, use the distilled Llama models on Amazon Bedrock with Custom Model Import, or train DeepSeek models with SageMaker via Hugging Face.

We've always believed that no single model is right for every use case, and customers can expect all kinds of new options to emerge in the future.

That's why Amazon Bedrock and Amazon SageMaker AI make it seamless for customers to test and evaluate the latest models as they emerge and pick the best ones based on their unique needs.

The company's community site also published "Deploying DeepSeek-R1 Distill Llama Models on Amazon Bedrock," and DeepSeek-Coder-33B Instruct: Let the Code Write Itself is now in the AWS Marketplace.

Google
That same DeepSeek R1 model is available in Google's Vertex AI platform, where the use case is described as: "DeepSeek-R1 achieves performance comparable to OpenAI-o1 across math, code, and reasoning tasks."

DeepSeek R1 on GCP
[Click on image for larger view.] DeepSeek R1 on GCP (source: Google).

Other than that, Google seems to have been mum about DeepSeek in terms of announcements or documentation.

Outside of the company, The Sting Development on the Medium online site offers more information in a post titled "Deepseek on Vertex AI endpoints." That is obviously interesting to some developers including one who authored a Reddit post titled "How to host Deepseek R1 on Google Cloud and access it like a traditional API?"

About the Author

David Ramel is an editor and writer at Converge 360.

Featured

Subscribe on YouTube