Sharon Zhou, the CEO and co-founder of Lamini AI, has taken a different approach compared to many others in the AI industry when it comes to hardware choices. While most tech CEOs have been rushing to acquire Nvidia’s GPUs, Sharon Zhou has opted for AMD’s GPUs to drive her startup forward.
Last year, there was a frenzy among tech CEOs seeking Nvidia chips, particularly for their GPUs, which were in high demand due to the surge in generative AI applications. Notable figures like Mark Zuckerberg and Sam Altman were eager to secure these GPUs to power apps like ChatGPT. However, Sharon Zhou stood out by not relying on Nvidia’s GPUs and instead choosing AMD’s GPUs for Lamini AI.
Sharon Zhou, a 30-year-old with an impressive background, holds a unique position in the AI landscape. She pursued a dual major in classics and computer science at Harvard, obtained a Ph.D. in generative AI from Stanford under Andrew Ng, and even served as an adjunct professor at the university. Additionally, she was part of the early team at Anthropic, a competitor of OpenAI that recently raised significant funding from Amazon.
Despite her remarkable journey, Sharon Zhou is now focused on leading Lamini AI, a startup she co-founded with Greg Diamos in Palo Alto. The company emerged from stealth mode in April last year with a mission to provide a platform that simplifies the training and customization of large language models for enterprises using just a few lines of code. Their goal is to enable enterprises to fine-tune models like GPT from OpenAI with their own data, essentially offering a localized version of OpenAI’s infrastructure.
In a surprising move, Sharon Zhou disclosed in September that Lamini AI had been exclusively utilizing AMD’s GPUs to build customized large language models, diverging from the trend of Nvidia’s dominance in the space. This decision was not taken lightly, considering the widespread preference for Nvidia’s GPUs in the generative AI community. However, with insights from her co-founder Greg Diamos, a former Nvidia software architect, they realized the potential of AMD’s GPUs and the significance of software compatibility in maximizing AI model performance.
Sharon Zhou’s strategic choice to leverage AMD’s GPUs was also driven by the need to address the growing demand for large language models without being hindered by supply constraints associated with Nvidia’s GPUs. By offering vendor-agnostic compute options that deliver comparable performance on both Nvidia and AMD GPUs, Lamini AI aims to meet the rising demand for large language models effectively.
Sharon Zhou’s confidence in AMD’s technology is evident as Lamini AI continues to invest in AMD’s latest offerings. In January, she showcased AMD’s new MI300X chip, touted by CEO Lisa Su as the “highest performing accelerator in the world,” actively in production at Lamini. This move signifies Lamini AI’s commitment to AMD as a key partner in their journey to provide high-performance AI solutions.
As Sharon Zhou and Lamini AI pave their path forward with AMD’s GPUs, they present a formidable challenge to Nvidia’s dominance in the AI hardware landscape. With a focus on scalability, performance, and accessibility, Lamini AI is poised to make a significant impact in the AI industry, backed by Sharon Zhou’s visionary leadership and strategic decision-making.
