Microsoft Doubles Down on DeepSeek’s R1 Model Despite IP Drama

While OpenAI is crying foul over DeepSeek’s alleged IP theft, Microsoft is rolling out the red carpet for DeepSeek’s R1 reasoning model on its Azure AI Foundry platform. This isn’t just a partnership—it’s a POWER MOVE. Microsoft is betting big on R1, even as whispers of controversy swirl around its development.

“As we continue expanding the model catalog in Azure AI Foundry, we’re excited to see how developers and enterprises leverage R1 to tackle real-world challenges and deliver transformative experiences.”

Microsoft

Why R1 is a Game-Changer for Azure AI

Microsoft isn’t just adding R1 to its cloud—it’s making it BETTER. The version of R1 on Azure AI Foundry has undergone:

• Rigorous red teaming to test its limits
• Automated behavior assessments to ensure safety
• Extensive security reviews to mitigate risks

And that’s not all. Soon, developers will be able to run “distilled” versions of R1 locally on Copilot+ PCs, Microsoft’s AI-ready hardware. This isn’t just innovation—it’s a REVOLUTION.

The Elephant in the Room: DeepSeek’s Controversial Past

Let’s not sugarcoat it: DeepSeek’s R1 comes with baggage. Microsoft reportedly launched a probe into DeepSeek’s alleged misuse of OpenAI’s API in 2024, accusing it of exfiltrating massive amounts of data. Despite this, Microsoft—OpenAI’s largest shareholder—is pushing forward with R1. Why? Because R1 is HOT, and Microsoft knows it.

R1’s Achilles’ Heel: Accuracy and Censorship

But here’s the catch: R1 isn’t perfect. According to NewsGuard, R1 delivers inaccurate or non-answers 83% of the time on news-related topics. And when it comes to China? R1 clams up, refusing to answer 85% of prompts. This isn’t just a bug—it’s a feature of the censorship-heavy environment in which R1 was developed.

“R1’s limitations are a stark reminder of the challenges facing AI models in navigating complex, real-world data.”

TechCrunch

What’s Next for R1 and Azure AI?

Microsoft’s decision to bring R1 into its cloud fold is a bold one. Will they tweak the model to improve accuracy and reduce censorship? Only time will tell. But one thing’s for sure: Microsoft is playing to WIN, and R1 is their latest weapon in the AI arms race.

Want more AI insights? Sign up for TechCrunch’s AI newsletter and stay ahead of the curve.