Posted: December 20, 2023 4:09 AM Updated: December 20, 2023 4:09 AM
Edit and fact check date: December 20, 2023, 4:09 AM
briefly
Preferred Networks increases its investments in custom AI chips, aiming to secure access to hardware amid surging global AI investments.
Japan-based startup Preferred Networks is investing in custom artificial intelligence (AI) chips, aiming to secure access to critical hardware amid a surge in global investment driven by advances in generative AI. It is increasing.
Preferred Networks specializes in chips optimized for a variety of AI tasks, boasting lower power consumption and improved computing performance, achieved in part by moving functions typically handled in hardware to software.
The startup has finalized the design of its second-generation AI chip, which is currently in production at TSMC and will power upcoming supercomputers.
“We recognized the need to optimize energy consumption and minimize procurement risk, both critical factors in sustaining our business.”
said Toru Nishikawa, Co-Founder and CEO of Preferred Networks.
The company began developing the first generation of AI chips to power supercomputers in 2016 with investors including automaker Toyota and robot maker Fanuc.
Preferred Networks plans to deliver state-of-the-art technologies for large-scale language model (LLM) creation and drug discovery next year and deliver pure computing power to customers by 2027.
Companies are accelerating the production of AI chips in-house.
There is a shortage of GPUs as the demand for generative AI, which typically uses GPUs for training and operations, increases. To reduce reliance on GPUs, deep-pocketed companies, especially big tech companies, are making and sometimes offering custom chips designed for AI model development and deployment to their customers.
Amazon and Microsoft are among the companies developing chips internally.
In recent developments, AWS unveiled AWS Graviton4 and AWS Trainium2, two families of chips designed in-house for foundational model training and inference.
Microsoft also launched Azure Maia 100 and Cobalt 100, custom AI chips designed specifically for the company’s cloud infrastructure. These silicon chips are intended to power Microsoft’s Azure data centers and are scheduled to be released in 2024.
The aggressive pursuit of custom AI chips by companies around the world reflects a strategic response to the global demand for technology hardware.
disclaimer
In accordance with the Trust Project Guidelines, the information provided on these pages is not intended and should not be construed as legal, tax, investment, financial or any other form of advice. It is important to invest only what you can afford to lose and, when in doubt, seek independent financial advice. We recommend that you refer to the Terms of Use and help and support pages provided by the publisher or advertiser for more information. Although MetaversePost is committed to accurate and unbiased reporting, market conditions may change without notice.
About the author
Alisa is a reporter for Metaverse Post. She focuses on everything related to investing, AI, metaverse, and Web3. Alisa holds a degree in Art Business and her expertise lies in the fields of art and technology. She developed a passion for journalism through writing about VCs, notable cryptocurrency projects, and participating in science writing.
more articles
alice davidson
Alisa is a reporter for Metaverse Post. She focuses on everything related to investing, AI, metaverse, and Web3. Alisa holds a degree in Art Business and her expertise lies in the fields of art and technology. She developed a passion for journalism through writing about VCs, notable cryptocurrency projects, and participating in science writing.
more articles