IBM Enables DeepSeek R1 on watsonx.ai

Computing giant IBM recently announced that a distilled version of DeepSeek-R1 can be deployed on watsonx.ai, an enterprise-grade AI development studio that facilitates the creation, training, and deployment of AI models within business applications. Developers can upload and deploy custom foundation models to utilise watsonx.ai’s inference capabilities. 

The company also outlined a detailed stepwise procedure to help users deploy distilled variants of DeepSeek-R1 for inference. 

“The release of DeepSeek-R1, an open-sourced reasoning model on par with OpenAI’s o1 series of models, is a significant step in this direction,” IBM said in the announcement. “We hope that it will inspire other model providers to follow suit and contribute to the growth of the open-source AI ecosystem.”

On Wednesday, IBM also announced its earnings results for the fourth quarter of 2024. The company reported a revenue of $17.6 billion (up 1%), while its software revenue grew by 10%. IBM’s gross profit margin increased by 40 basis points to 59.5%. However, consulting revenue was down 2%, and infrastructure revenue was down 8%. 

“Our generative AI book of business now stands at more than $5 billion inception-to-date, up nearly $2 billion quarter over quarter,” said Arvind Krishna, CEO of IBM. He also revealed that about eight points of growth were organically led by the company’s AI products, including IBM Concert and other AI assistants.  

The company also expects a revenue growth of at least 5% for 2025, along with a free cash flow of around $13.5 billion. 

“We are confident that our investment in partnerships and skills, as well as our early leadership in GenAI, position us to accelerate consulting growth as we move forward,” Krishna added. 

Last November, the company released a new variant of their open-source AI model, the Granite 3.0. Armand Ruiz, VP of product at IBM’s AI platform, publicly disclosed the datasets on which the model is trained. This is a practice IBM has adhered to even in the past with new model releases. 

“This is true transparency. No other LLM provider shares such detailed information about its training datasets,” Ruiz further said.