AWS’ AI Chips ‘Extremely Important’ For Cost And Supply Chain: Partner

'We have seen, especially in that AI space, these extremely high-powered chips on back order because they can’t make them fast enough. AWS own chips are going to be extremely beneficial on that side of it, especially as these AI workloads increase,' says ClearScale CEO Jimmy Chui.

Solution providers are bullish on Amazon Web Services’ plan to invest heavily in building its own chips specifically optimized for AWS infrastructure that will give partners a leg-up over the competition in the AI market due to lower electrical costs and supply chain benefits.

“We’re already seeing the massive uses of AI and companies identifying, ‘Oh, it’s not just expensive for the chips, but oh my gosh, they take up so much energy.’ So electric bills are going through the roof,” said Jimmy Chui, CEO of AWS partner ClearScale.

“So you can imagine in a huge data center where everything’s spinning up, processing at a high rate of speed—you’re going to chew through wattage very, very quickly,” said Chui.

For example, Oracle is looking to power its data centers via nuclear reactors in order to meet the electricity demand from AI.

"The location and the power place we’ve located, they’ve already got building permits for three nuclear reactors,” said Oracle chairman and co-founder Larry Ellison during the company’s quarterly earnings call this week. "These are the small modular nuclear reactors to power the data center. This is how crazy it’s getting.”

AWS CEO Matt Garman said his $105 billion cloud company will continue to build its own AI chipset, which will be optimized alongside its massive infrastructure.

“We don’t have to build these processors to run in a general-purpose environment. They’re going to run exactly in our server, exactly in our data center, exactly with our networking stack and so we can optimize that just for our customers,” said Garman. “We can optimize like crazy around that, plus we have a very good team that’s building the chips.”

Homegrown AWS Chips To Boost Supply Chain

ClearScale’s CEO said with AWS building its own AI processors, customers won’t be impacted as much from supply chain issues.

The IT industry has been plagued with supply chain shortages related to AI processors due to high customer demand from businesses across the world.

“With their ability to purchase and design these chips, that’s also going to secure them on the supply side, which is really smart,” ClearScale’s Chui said.

“We have seen, especially in that AI space, these extremely high-powered chips on backorder because they can’t make them fast enough. AWS’ own chips are going to be extremely beneficial on that side of it, especially as these AI workloads increase,” said Chui.

Ethan Simmons, managing partner of AWS partner PTP, agreed that AWS manufacturing its own AI processors will help customers both with cost and access to AI chips. “Owning that AI hardware is just so important for a variety of reasons and only someone like AWS can pull it off,” he said. “I’m happy that [AWS CEO Garman] is focused on the right things.”

AWS CEO: ‘Own Custom Processors Can Help Customers Save Money’

In recent years, AWS turned its attention to creating AI processors such as AWS Inferentia machine learning chips and Trainium ML processors for training deep leaning models.

AWS’ initial decision to build custom chips in-house was to support its virtualization technology via Nitro and ability to dictate customer cost.

“That means that we don’t have to go buy those from third parties, which allows us to lower our cost,” said Garman. “We think that there are some use cases where our own custom processors can help customers save money.”

AWS is also forming tight partnerships with the world’s most innovative and largest chipmakers like Intel, AMD and Nvidia.

“We firmly believe that AWS is the absolute best place to run Intel, to run AMD, to run Nvidia processors and we think that we can offer some differentiated capabilities by offering our own processors as well,” Garman said.

AWS Ahead Of The Game

ClearScale’s Chui said it’s “extremely important” that AWS designs chips specifically for AWS’ new AI services, similar to when it built chips designed for the AWS platform and cloud services.

AWS is currently ahead of the game when it comes to AI as most other competitors and companies are simply focused on AI experimentation or initially use cases.

“We’re all experimenting with AI or getting into it. The timing for getting your own chip that is extremely efficient and works on their particular infrastructure to keep costs low—it’s the right timing for it. Because as soon as everybody who is right now experimenting with it gets traction and figures out, ‘Yes, this is my implementation’ and it’s live in production, we’re probably going to see that same explosive energy use and cost,” said Chui.

AWS plans to launch its new Trainium2 processors by the end of this year.

“We feel incredibly excited about that platform,” said Garman. “We think that we have the opportunity to aggressively lower cost for customers while increasing performance.”