Topic: compute power
-
Ex-Apple Star Aims to Build the Perfect GPU Software
Chris Lattner co-founded Modular in 2022 to create a universal software layer that enables AI applications to run efficiently across diverse processors without extensive rewrites. The company's strategy involves a unifying software stack and a new Python-based programming language to maximize har...
Read More » -
Huawei Unveils New AI Infrastructure Amid Nvidia's China Exit
Huawei has launched a new AI infrastructure platform, including its SuperPoD Interconnect technology, to link up to 15,000 GPUs and enhance computing capabilities. The platform clusters Huawei's Ascend AI chips to compensate for their lower individual performance, enabling competitive AI training...
Read More » -
OpenAI and Oracle Seal Historic Cloud Partnership
Oracle's stock surged after announcing multi-billion-dollar cloud deals, including a historic $300 billion agreement with OpenAI set to begin in 2027. OpenAI is diversifying its cloud infrastructure beyond Microsoft Azure, partnering with Oracle and Google to meet its massive computational demand...
Read More » -
How to Build an AI-Powered Enterprise for the Future
Companies face immense pressure to adopt AI, with 98% reporting increased urgency and 85% believing they have under 18 months to deploy strategies to avoid adverse effects. Most organizations are unprepared for AI integration, with only 13% feeling fully ready and 68% citing inadequate IT infrast...
Read More » -
Tesla's Dojo: A Complete Timeline and Breakdown
Tesla developed the Dojo supercomputer to advance its Full Self-Driving (FSD) system by training neural networks with massive computational power and data. Despite initial progress and investment, Tesla shifted strategy in 2024 to Cortex, a new supercomputing cluster based on Nvidia GPUs, moving ...
Read More »
