According to Bloomberg on the 25th, Amazon is advancing an ambitious self-developed chip development plan aimed at challenging Nvidia's current monopoly in the field of artificial intelligence (AI) chips. However, due to Nvidia's continued leading position in technology updates and market share, market participants believe that its competitive advantage is difficult to shake in the short term.
According to reports, Amazon is currently intensifying the development of a new AI chip called Trainium2 at its engineering laboratory in Austin, Texas. Compared to the previous generation product, its performance has improved fourfold, memory capacity has increased threefold, and it has significant advantages in energy efficiency and cost. Amazon hopes to reduce the procurement cost of AI chips and improve overall efficiency in data processing through these optimizations. This chip is Amazon's third-generation product in the AI hardware field, aimed at providing a more efficient and cost-effective solution for machine learning model training. The product plan is to complete testing and delivery before the end of this year.
The article also points out that in order to truly challenge Nvidia's leadership position in the AI hardware market, companies such as Amazon still face enormous challenges.
Previously, semiconductor manufacturer AMD released its latest artificial intelligence chip. According to data provided by AMD, servers running AMD processors have increased video transcoding time by 4 times, and the performance of each core in virtualization infrastructure has increased by 1.6 times. But the market response was average, with a significant drop in stock price on the day the new product was released.
Traditional chip giants such as AMD and Intel hope that artificial intelligence computers can boost computer sales. However, so far, these efforts have not achieved significant results. According to Gartner's data, global computer shipments continued to decline in the third quarter. As the holiday season approaches, the industry has the opportunity to promote the benefits of artificial intelligence computers to consumers, but whether consumers will accept them is another matter.
Benefiting from the rapid development of AI technology applications, Nvidia has seized the industry opportunity. The Hong Kong University of Science and Technology stated that in the past 30 years, Huang Renxun has led NVIDIA to achieve fruitful results in the field of accelerated computing, and the GPU developed by his team in 1999 laid the foundation for the development of the AI era. Companies operating large data centers have been competing to purchase AI processors that Nvidia excels at. Musk's xAI manufactured a supercomputer called Colossus in Memphis in just a few months, equipped with 100000 NVIDIA Hopper AI chips. Meta CEO Mark Zuckerberg stated last month that the company is already training state-of-the-art AI models with a cluster of chips. OpenAI used approximately 10000 Nvidia chips to train the ChatGPT version to be released at the end of 2022.