ChatGPT, an OpenAI-backed Artificial Intelligence (AI) language model, could cost up to $700,000 a day to operate, according to Dylan Patel, the Chief Analyst at semiconductor research firm SemiAnalysis. This is undoubtedly a costly endeavor due to the extensive computing power needed to process user prompts and generate responses. In order to reduce this exorbitant cost, Microsoft is developing a new AI chip, called Athena, as an alternative to OpenAI’s pricey infrastructure.
Microsoft began working on Athena in 2019, and now more than 300 employees work on it. It is projected to replace high-powered GPUs like NVIDIA’s and become available for internal use as early as next year. The chip would have both cost and performance advantages and will be essential for OpenAI’s language model operations.
Nick Walton, CEO at Latitude, a startup whose AI dungeon game relies on OpenAI’s language model, said that running the model costs them $200,000 a month in Amazon Web Service servers. To cut costs in half, Latitude decided to switch to an AI software provider backed by AI21 Labs. This experience emphasizes just how prohibitively expensive these AI models are to operate.
Microsoft is taking this into account with their new Athena AI chip. If Microsoft is successful, their AI chip could dramatically reduce operational costs for OpenAI and other companies using language models. This could prove to be a game changer for OpenAI’s financial success, as the cost of software often plays a key role in a company’s overall expenditure.
Microsoft is a tech company whose mission is to empower every person and every organization on the planet to achieve more. Founded in 1975 by Bill Gates and Paul Allen, it has grown to become the world’s largest software maker. Microsoft’s cloud and artificial intelligence platforms, such as Azure and Dynamics 365, are used by businesses and organizations around the world.
Dylan Patel is the Chief Analyst at SemiAnalysis, a semiconductor research firm. Patel is the author of “Semiconductor Analysis: Market Characteristics, Investment Profiles, and Projected Road Map,” a breakthrough publication that provides insight into the strategies and investments of semiconductor companies. He is a graduate of the esteemed University of California, Berkeley School of Engineering. He is a respected voice in the semiconductor industry and has offered insights on AI chips, cost and budgeting for AI projects, and more.