OpenAI is on the brink of a tech revolution, and it’s not just about clever chatbots. According to Reuters, OpenAI is gearing up to launch its very own AI chip, potentially finalising the design this year. This move is a game-changer, allowing OpenAI to ease off its reliance on Nvidia’s chips to train and run AI models.
The plan is to send the design to Taiwan Semiconductor Manufacturing Co (TSMC) for fabrication, using cutting-edge 3-nanometer technology. This means more efficient chips with high-bandwidth memory and top-notch networking capabilities. Fancy, right?
Initially, OpenAI will roll out these chips on a limited scale, mainly for running AI models. But don’t think they’re stopping there! Future versions are expected to boast even more advanced processors and capabilities.
Want to hear more? Join Mal on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
This development follows whispers from last year about OpenAI collaborating with Broadcom on a custom chip. With former Google TPU engineer Richard Ho at the helm, the team has doubled in size, now with 40 bright minds driving this ambitious project forward.
While OpenAI and other tech giants splash out on AI infrastructure, the debate continues. Do companies really need thousands of chips to power their systems? DeepSeek seems to think not. But as the AI arms race heats up, it’s clear that OpenAI is determined to stay ahead of the curve. So, keep your eyes peeled, folks—this is one tech journey you won’t want to miss!
Want to hear more? Join Mal on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai
Comments