Is Microsoft CEO Satya Nadella Encouraging His Team to Create a Comparable AI in Computing?
Recently, Satya Nadella, the CEO of Microsoft, praised Deep Seek, a China-based AI company, for its innovative approach to AI computing architecture. During a company-wide meeting, Nadella highlighted Deep Seek as a benchmark for Microsoft’s own artificial intelligence initiatives, pointing out the impressive results achieved by a relatively small team. Reports from The Verge indicate that Nadella emphasized the significance of focused innovation while commending Deep Seek’s groundbreaking contributions to the field of AI.
“What’s most impressive about Deep Seek is that it’s a great reminder of what 200 people can do when they come together with one thought and one play,” Nadella remarked while speaking to Microsoft employees.
Understanding Nadella’s Appreciation
At the forefront of Nadella’s praise was Deep Seek’s R1 model, which has recently gained popularity, ranking among the top applications in the US Apple Store. Nadella attributed this achievement to the clear vision and teamwork exhibited by Deep Seek’s 200-member team. He noted the transition of Deep Seek from a research initiative to a widely-used consumer product, showcasing the effectiveness of its advanced computing architecture in making significant strides in AI technology.
Jay Parikh, who leads Microsoft’s CoreAI engineering division, echoed Nadella’s sentiments, emphasizing how Deep Seek’s accomplishments underscore the vital role of teamwork and swift innovation in the competitive AI landscape. This success not only draws admiration from Nadella but also establishes a new objective for Microsoft in its AI strategy, serving as a powerful motivator to further invest in AI technologies.
Nadella was particularly impressed by Deep Seek’s system optimization capabilities, especially its performance under Nvidia’s CUDA layer. He recognized this as a prime example of cutting-edge technology that holds the potential to shape future advancements in computing.
Nvidia’s CUDA Technology Explained
Nvidia’s CUDA technology accelerates computing by harnessing the power of graphics processing units (GPUs), enabling tasks to be processed at significantly faster rates. By dividing complex workloads into smaller segments, CUDA allows these tasks to run concurrently across multiple GPU cores. This multitasking capability makes processes, like AI and deep learning, much more efficient. Instead of solely relying on traditional central processing units (CPUs), systems can leverage the expansive power of GPUs, enabling them to manage large workloads swiftly.
Comparing ChatGPT and Deep Seek
Deep Seek was launched on January 20, 2025, with its development executed by a team of 200 individuals at a total cost of under $6 million. This budget is considerably lower than that of GPT-4, which exceeded $100 million in development expenses. The creators of Deep Seek managed to navigate US chip export restrictions by stockpiling Nvidia A100 chips while complementing them with more cost-effective alternatives. This innovative strategy, along with optimized memory usage, allows Deep Seek to deliver more efficient and budget-conscious outcomes than ChatGPT.
Deep Seek’s practical approach showcases how a focused team can revolutionize an industry through strategic decisions and advanced technological implementation. The achievements of Deep Seek serve as an inspiration for other tech companies, pushing the boundaries of what is possible with artificial intelligence.
