Cloud native EDA tools & pre-optimized hardware platforms
World-altering innovation occurs at the intersection of opportunity and complexity.
This according to Rene Haas, CEO of Arm, and Sassine Ghazi, president and CEO of 草榴社区, who sat down at the 35th annual 草榴社区 User Group (SNUG) conference to share their views on the unprecedented opportunity and complexity of AI.
“It’s an incredibly exciting time,” Haas said. “I never thought I would have the opportunity to work on the kinds of problems we’re trying to solve with artificial intelligence. And it just seems to go faster and faster and faster.”
“I strongly believe we are at an inflection point with AI,” Ghazi added. “There are certain challenges and problems that humanity has not been able to solve yet.”
Like improving health outcomes on a global scale. Haas noted the traditionally methodical pace of medicinal research and elusiveness of breakthroughs.
“The time the initial R&D starts to the time you actually have a drug approved is about 15 years,” he said. “And the failure rate from early research to end product is 95%. It’s crazy.”
With AI accelerating health-related research, these longstanding norms are being upended.
“There’s a huge promise, for example, of eliminating some phase one trials and even animal trials by using artificial intelligence,” Haas said. “And the idea that we could see a cure for cancer in our lifetime is very, very real.”
“AI is going to touch every industry in terms of applications and productivity,” Ghazi said. “It will enable new business models and new paradigms.”
But first, complexity must be tamed. Underlying technologies must catch up with the increasing speed of innovation. And ever-growing computing demands must be met in an efficient and sustainable manner.
“It's a huge, huge computing challenge,” Haas said. “It’s a huge power challenge.”
According to the International Energy Agency (IEA), data centers may consume — as much as the entire country of Japan. And it’s only a near-term projection.
Ghazi and Haas discussed the gigantic data centers — like those being built for the — needed to support the insatiable processing demands of AI workloads, which continue to grow exponentially.
“The numbers are huge,” Haas said. “Multiple gigawatts of energy over a geography that we’ve never done before, thermals that we’ve never dealt with before.”
In addition to scaling up and out, processing performance and power efficiency must also shrink.
“We focus so much on the infrastructure build-out and the data center, but the moment you start running AI on devices, that's another opportunity,” Ghazi noted.
“In mobile, the designs are unbelievably complex in terms of power, geometry, cost, and speed. How do I run AI in a wearable?” Haas questioned. “How do I run AI in a battery-operated device that still has to run displays, still has to run the operating system, still has to launch your applications?”
“It’s a system-level problem,” said Ghazi.
The technologies being used today to drive AI workloads were never designed for those workloads. Silicon devices must become more specialized, Ghazi explained. And they must be purpose-built for the applications and systems they will serve.
“Designing a general-purpose chip and using it for many applications — those days are gone,” he claimed. “Especially for AI applications where you need to have an efficient silicon for a specific workload.”
“People will throw down lots of transistors to address a general-purpose problem,” Haas added. “You’re leaving performance on the table, you’re leaving efficiency on the table, and you’re delaying the capability of the software to take advantage of it. There is a gigantic opportunity to do things from a design standpoint that are far, far more power efficient than they are today.”
Design optimization won’t be limited to the performance and energy efficiency of silicon. It will encompass entire systems — from hyperscale data centers to cars, drones, and robots — and all of the physics they encounter.
“You have to consider the mechanics, thermal dynamics, and fluid dynamics in relation to the electrical and electronic components,” Ghazi said, noting the need for holistic, silicon-to-systems design. “It’s a level of complexity we’ve never seen before, and AI will help us solve it.”
According to both technology leaders, the immense challenges and opportunities of AI will, in large part, be addressed with the assistance of AI.
“These are very, very complex problems to simulate from a system level,” Haas said. “How do I simulate the software, the workload, running a very large frontier model? Do I have the right design in terms of power efficiency relative to my air-cooled system, to my liquid-cooled system?
“The current [design and engineering] methods will break down,” he said. “AI has to help us there.”
“That’s why I’ve been talking about Re-engineering Engineering?,” Ghazi added. “With AI, there is a great opportunity to change the workflow. Not only the chip design workflow, but the workflow in many applications and many industries.”
With 草榴社区 already on the forefront of AI-assisted chip design, Ghazi said agentic AI solutions that are more capable and autonomous are on the horizon.
“We’ve done a very good job in providing assistive technology and reinforcement learning in pretty much every part of our flow, but the workflow has stayed the same,” he explained. “The moment you go from assistive technology to specialized agents for specific parts of the flow, you can rethink traditional methods and achieve different outcomes.”
“Everything around verification, validation, bug fixes, conversion,” Haas added. “Those are things AI should get very good at, and it will shave down the amount of time it takes from ideation to data capture.”
With AgentEngineer? technology doing more planning, orchestration, and workflow optimization, design cycles will accelerate, innovation will increase, and there will be .
“I don't think in any way, shape, or form that this means fewer jobs,” Haas noted. “It means more products, better products, developed faster.”
草榴社区 and Arm will continue working together to enable these opportunities, the two leaders said, by tackling the inherent and unprecedented complexity within.
“草榴社区 is a true partner with us in every sense of the word, and that is so critical,” Haas said, noting the longstanding collaboration between the two companies. “Thank you for your partnership with Arm and for everything you’re doing in the industry.”