News on Medial

Chip startup Cerebras unveils new AI processor; Here's what you need to know

IndianStartupNewsIndianStartupNews · 1y ago
Chip startup Cerebras unveils new AI processor; Here's what you need to know
Medial

Cerebras Systems, an AI startup, has released a new version of its AI processor with improved performance at the same cost. Unlike traditional chip designs, Cerebras uses a single large processor rather than multiple smaller chips. The CEO, Andrew Feldman, highlighted their rapid technological advancements, moving to a five-nanometer part from a seven-nanometer process within just three years. Cerebras' third-generation chip, the Wafer-Scale Engine 3, addresses the challenge of power consumption in AI processing and boasts 4 trillion transistors and the capacity to perform 125 petaflops of computing. The company plans to sell these systems alongside Qualcomm AI 100 Ultra chips to enhance AI application performance.

Related News

AI is a particularly well-suited tech trajectory for India: Cerebras' Feldman

LivemintLivemint · 1y ago
AI is a particularly well-suited tech trajectory for India: Cerebras' Feldman
Medial

Cerebras Systems CEO Andrew Feldman believes that AI can solve many of the world's problems, although it is not a "silver bullet." Cerebras Systems recently released the CS-3, which is touted as the world's largest and fastest AI chip. The chip enables the training of AI models at faster rates than traditional GPUs by distributing a single model over a larger surface area. Various companies across sectors, including GlaxoSmithKline Pharmaceuticals and TotalEnergies, are leveraging the CS-3 for tasks such as genomic research, drug design, hospital administration, and AI work in oil exploration. Cerebras Systems has a strategic partnership with Abu Dhabi-based Group 42 Holding for building supercomputers and Arabic natural language models (LLMs). Feldman also highlighted the potential of India's university systems in the AI field and discussed the advantages of their large chip over GPUs. He emphasized that while many customers focus on fine-tuning language models and inference tasks, there is still a need for significant compute power and accuracy. Feldman believes that AI will continue to evolve and have a substantial impact on various fields, but it is not the answer to every problem. He also expressed cautious optimism about the development of artificial general intelligence (AGI), stating that while machines may possess thoughtful reasoning capabilities, he does not believe we are close to achieving self-awareness. CXOs looking to implement AI projects should consider their advantage in algorithms, data, and computing power and form partnerships accordingly for successful implementation.

Download the medial app to read full posts, comements and news.