Revolutionary A.I. Outpaces the Chip Industry’s Capabilities

Chips are not meeting the computing needs of today’s advanced applications

The chip industry has seen an explosion in value due to A.I. and a bullish forecast from Nvidia. Despite this, the industry is facing a significant challenge in that chips have stopped providing substantial leaps in computing power, particularly as we see a rise in power-hungry applications such as generative A.I. Historically, microprocessors have doubled in performance every two years, known as “Moore’s Law,” but we haven’t seen this level of performance increase in quite some time. It’s now common for a microprocessor to only increase speed by 10-15% per year, which is often too insignificant to notice in software applications. At the same time, the performance needs of tasks such as computer vision, natural language processing, and speech processing have seen a substantial increase that requires far more computing power than chips can provide.

The chip industry needs more innovation, but it is inhibited by the cost and time of experimentation. It can take between $10-30 million just to launch the first product, along with an additional $70-100 million to scale up. With such vast sums required, the sector is not seen as an attractive investment for entrepreneurs and investors. The gestation period for new ideas within the industry is also too long, further discouraging investors and innovators. Furthermore, the industry is relatively consolidated, with just a few large players dominating the market.

The U.S. government can use its CHIPS Act funding to make the industry more welcoming to innovation. One potential solution is to incentivize the industry to adopt agile methodologies, open-source tools, and open standards. The government can also encourage funding recipients to make commonly used hardware components widely accessible, enabling other companies to combine them with new components and create new hardware platforms.

The government can also require academic CHIPS Act beneficiaries to modernize their chip design curriculums to emphasize accessibility and impact. Additionally, the government can allocate funding to support the development of a shared, subsidized infrastructure for design and fabrication with mature, trailing technologies to reduce the cost of producing proof-of-concept hardware. Ultimately, right-to-repair legislation can encourage a culture of tinkering with hardware, boosting innovation and the attractiveness of the industry.

Conclusion

Now is the time for the chip industry to embrace innovation. The gap between computing needs and the computing power of chips can only be bridged by more innovation, making computing expensive and exclusively accessible only to rich and powerful corporations. By using its resources, the government can have a significant impact in helping the industry reduce the cost and time of experimentation while making it more attractive to entrepreneurs and investors, ultimately leading to more disruptive ideas and innovations to meet the computing needs of today and the future.