Deep learning, one of the major economic challenges of the 2030s according to a report

20 avril 2021

According to US asset management firm ARK Invest, deep learning will be one of the major economic challenges in one to two decades. This analysis is part of its “Big Ideas 2021” report, which also covers topics such as Bitcoin and 3D printing. According to the company, deep learning could generate up to 30,000 billion dollars within 15 to 20 years.

The ARK Invest report states that the Internet generates an economy of up to 13 trillion dollars. Deep learning is currently estimated to generate about $2 trillion, but in 15 to 20 years, this figure could rise to$ 30 trillion, or 2.5 times more than the internet currently does. In view of the significant financial windfall that deep learning could potentially bring, many companies and investors want to implement technological innovations that promote deep learning techniques and systems.

This explains the increasing research into the creation of next-gen data centers and new processors that are more efficient, less energy-consuming and less expensive to produce. In 2021, x86 microprocessors will be the most widely used in data center architecture, accounting for 92% of the total, but the arrival of ARM and RISC-V architecture in the high-performance processor market is expected to change the game completely.

According to ARK Invest, ARM and RISC-V processors will be used in 71% of data centers by 2030, relegating x86 processors to only 27% of the market. ARM and RISC-V processors could reach $19 billion in revenue by 2030 and data center spending on AI processors could increase<strong> </strong>more than fourfold over the next five years, from$ 5 billion a year today to $22 billion in four years.