Image showing the Cerebras Systems WSE-2 chip size compared to a hockey puck. AI chip start-ups are raising large funds from venture capitalists to challenge Nvidia and Intel. (Image Credit: Cerebras Systems)

 

The world needs more. I'm glad to see some expansion in the area of manufacturing.

 

Silicon Valley venture capitalists are investing in the semiconductor industry with high hopes to produce next-gen AI chips, which could pose a challenge to Intel and Nvidia. The latest surge of semiconductor innovation has made minimal effort to address the chip shortage based on older technology plaguing the auto industry. This shortage comes from varying factors while automakers, which shut down during the pandemic, compete against the consumer electronics industry for chip supplies.

 

Start-up chip makers such as SambaNova Systems, Groq, and Cerebras Systems said they are producing chips capable of performing AI tasks better than Nvidia processors. Cerebras has an innovative approach: developing a chip that hosts 2.6 trillion transistors and is 56 times larger than a postage stamp.

 

Over 50 companies are creating AI application chips. The market could be valued at over $70 billion by 2025, an increase from $23 billion in 2020. Last year, funding for chip start-ups in the U.S. hit $1.8 billion, which is the largest amount in twenty years. An additional $1.4 billion has already been provided so far this year.

 

In April, SamaNova received $676 million for its reconfigurable AI chips. Groq, the developer of an extremely powerful single-core design that’s simple and fast to program, announced it raised $300 million. The funding was significantly larger than anticipated for both companies due to strong investor interest.

 

Software-focused venture capital firms, such as Bessemer Venture Partners, Battery Ventures, and Foundation Capital, have shown interest in these opportunities. These were some of the most active semiconductor investors during the past two years.

 

Nvidia is expected to stay ahead in the AI chip race. The company’s GPUs have advanced over the past seven years, meeting AI software requirements. A few chips were designed to handle AI. New materials and methods to produce chips are propelling innovation in other areas apart from AI, such as chips for quantum computers.

 

All of this is expected to occur while the U.S. government finalizes new subsidy programs aiming to support the chip industry. The geopolitical clash between the U.S. and China could push venture capital firms to invest in both chip design companies and manufacturing. AI chips are making it easier for algorithms to enhance chip manufacturing, cut down on costs, and open manufacturing to start-ups.

 

Have a story tip? Message me at: http://twitter.com/Cabe_Atwell