How DeepSeek and Next-generation aI Agents might Erode Value Of Language Models

In the web transformation, we’re moving from as the primary service to really building internet-native companies – so, the Airbnb of AI, the Stripe of AI,” he added. “They are not about the design. They are about the system and how you make the model beneficial for tasks.”

The arrival of DeepSeek’s R1 design recently is seen by some tech CEOs as a further sign LLMs are becoming significantly commoditized.

R1 includes some crucial elements that differentiate it from other designs on the marketplace. For example, the model incorporates a “mixed precision” structure that utilizes a combination of full-precision 32-bit floating point (FP32) numbers and low-precision 8-bit drifting point (FP8) numbers.

The latter is much faster to process but can be less accurate. However, rather than rely on one or the other, DeepSeek uses FP8 for most estimations and switches to FP32 for certain tasks where a higher degree of accuracy is required.

“As AI gets more efficient and accessible, we will see its usage skyrocket, turning it into a commodity we just can’t get enough of,” Microsoft CEO Satya Nadella wrote on social media platform X, in an evident recommendation to DeepSeek.

Meanwhile, Matt Calkins, CEO of U.S. software company Appian, told CNBC that DeepSeek’s success merely shows that AI designs are going to become more of a product in the future.

“In my opinion, we’re visiting a commoditization of AI. Many business will achieve competitive AI, and a lack of differentiation will be bad for high-spending first-movers,” Calkins said via email.