Necessary Always Active
Necessary cookies are required to enable the basic features of this site, such as providing secure log-in or adjusting your consent preferences. These cookies do not store any personally identifiable data.
|
||||||
|
||||||
|
||||||
|
Published on
July 12, 2024
8 Min read
Amazon Unveils Non-AI Graviton4 Chip Amidst Tough Competition
Amazon Web Services (AWS) has unveiled a new processor called Graviton. Yahoo Finance reported that the Amazon Graviton4 chip offers superior efficiency and performance. Compared to its predecessor, the new chip comes with higher computing power and memory.
Commenting about the chip, Amazon’s Compute and AI Director of Product Management, Rahul Kulkarni, said, “Collectively it’s delivering more price performance, which means for every dollar spent, you get a lot more performance.”
Demand for Chips
Demand for chips is on the rise as semiconductors power most items and play a critical part in the global economy. Driven by the demand for AI, the chip industry is currently valued at $544 billion. This value is expected to surpass the $1 trillion mark by the year 2033.
Nvidia currently dominates the AI chip market. Its market share for graphics processing units (GPU) stands at 80%. However, the current demand for chips is sufficient to support multiple manufacturers. Increasingly, hyperscalers like Alphabet, Apple, Microsoft, and Amazon are developing custom chips to reduce cost, meet their unique needs, and provide affordable options for their customers.
Patrick Moorhead, CEO and Chief Analyst at Moor Insights AND Strategy said, “All of these companies are spending a lot of money on developing chips. They won’t talk about how much they’re investing, but they have these giant R&D budgets.”
Not an AI Chip
The Amazon fourth-generation Graviton4 chip is not an AI chip. It’s designed to support AWS’s Trainium and Inferentia chips, which focus on AI technology. Trainium competes with Nvidia AI chips, which are viewed as highly powerful and fast in the market.
AWS isn’t looking to unseat Nvidia. Rather, it is focused on providing customers who are focused on price performance with a viable option. The cloudservice provider hopes that by doing so, it’ll take a profitable slice of the fast growing AI-chip market.
“Right now, if a customer is more focused on time to market, Nvidia-based products that we offer are a great option. There are customers for whom cost becomes a very prohibitive aspect of running their business. If they want to do more cost-optimized AI workloads like training or inference, then our Inferentia and Trainium products become a great alternative,” Kulkarni said.
Minimizing Costs
Amazon has adopted different strategies to minimize chip development costs. Besides sourcing chips from manufacturers like AMD, Intel, and Nvidia, Amazon is developing chips in-house. At the moment, AWS offers two types of chips – general-purpose chips like the Graviton4 chip and AI-purpose chips.
The company has yet to make the price of Amazon Graviton4 chip public, but it leases processors at $0.02845 per hour of compute power. This price-to-performance model allows AWS to leverage its proprietary chips to power its servers and cloud infrastructure. This enables it to half its IT costs. Amazon’s influence in the semiconductor industry is also growing.
“AWS has a lot of credibility in the semiconductor space. Ten years ago, I had questions about how a company like this could do chips when you have companies investing multiple billions of dollars to do that. But they are very good at it,” Moorhead said.
The company expects its earnings from semiconductors to surpass those of Alphabet or Microsoft.
