Etched, a Cupertino-based startup founded by Harvard dropouts, has secured $120 million in venture funding. The funding round, led by prominent investors, underscores Etched’s ambition to develop a competitive AI chip named Sohu, designed specifically to power and optimize AI models built on transformer architecture.
Led by co-founder and CEO Gavin Uberti, Etched aims to capitalize on the growing demand for specialized chips, known as ASICs (Application-Specific Integrated Circuits), which offer higher efficiency in executing AI tasks compared to general-purpose GPUs from Nvidia. The company’s strategy hinges on the widespread adoption and evolution of transformer-based AI models, such as those used in OpenAI’s ChatGPT, which require increasingly powerful and energy-efficient hardware solutions.
In a recent interview, Uberti highlighted the company’s high-risk, high-reward strategy, acknowledging the fierce competition posed by Nvidia and other tech giants. He emphasized that Etched’s Sohu chip is optimized specifically for transformer architectures, positioning the startup uniquely in the AI hardware market.
Despite Nvidia’s formidable market presence and revenue prowess—exceeding $26 billion in recent quarters—Etched remains undeterred, believing in the transformative potential of its specialized AI chips. The $120 million funding, though substantial, represents a fraction of Nvidia’s revenue, illustrating the scale of the challenge ahead for the fledgling company.
The success of Etched’s venture will depend heavily on the trajectory of transformer-based AI technologies. Should these models continue to drive significant advancements in AI applications, Etched could potentially carve out a significant niche in the industry. However, the startup acknowledges the inherent risks in betting against established giants like Nvidia, whose technological and financial resources pose formidable barriers.
As Etched progresses in its development of the Sohu chip, industry analysts and stakeholders will closely monitor its ability to deliver on its promises and compete effectively in the fast-evolving AI hardware landscape. The outcome of this David-versus-Goliath battle in AI chips will undoubtedly shape the future dynamics of the industry, influencing innovations and market strategies for years to come.
