The untold secret of artificial intelligence (AI) is its huge computational requirements.
The cost to feed them is also nothing to smell. As new AI applications arrive, peak demand will increase. Power requirements will increase. IT requirements will also increase.
SoftBank recently announced that it is building data centers, in partnership with Nvidia, capable of hosting both AI generative and 5G wireless applications on a common multi-tenant server platform designed to both reduce costs and be more energy efficient. .
This is because when 5G networks come online around the world, they will create a literal network effect that will fuel the rapid worldwide deployment of generative AI applications and services.
By using 5G data centers synchronously for AI applications, enterprises can achieve a return on investment and computing and processing power.
As we enter an era where society coexists with artificial intelligence, the demand for data processing and electricity requirements will increase rapidly, said Junichi Miyakawa, president and CEO of SoftBank, in a statement. announcing the collaboration.
The demand for accelerated computing and generative AI is driving a fundamental shift in data center architecture, added Jensen Huang, founder and CEO of Nvidia. Nvidia Grace Hopper is a revolutionary computing platform designed to process and scale generative AI services.
The platform will initially be deployed in new AI data centers deployed in Japan that can host both generative AI and wireless apps on a shared server platform.
The growing number of data centers to securely store and manage data has increased the demand for AI-powered storage.
But despite the promising potential of AI, our current computing infrastructure isn’t built to handle the workloads that AI will throw at it.
This is because while, over the past half-century, most computing architectures tended to be CPU-centric, the future of generative AI will require higher performance and more energy-efficient computing capabilities.
The size of AI networks has grown 10x annually over the past five years, and by 2027, observers predict that 1 in 5 Ethernet switch ports in data centers will be dedicated to AI, machine learning, and to accelerated processing.
To escape this spiral, we need to rethink information architecture from the ground up. One solution is to completely disaggregate compute platforms, eliminating interdependencies between CPUs, GPUs, DPUs, memory, storage, networking, and so on.
We are seeing significant underutilization of networks under construction, and the return on investment (ROI) on 5G has been relatively low, said Ronnie Vasishta Sr., vice president of telecommunications at Nvidia, during an analyst briefing.
That’s why Nvidia is making its 5G infrastructure not only virtualized, but also fully software-defined, so you can run an efficient, high-performance 5G network alongside AI applications, all within the same data center.
This contemporary moment within the future of technical infrastructure is marked by the convergence of two key phase change drivers in the technology sector. The first is a shift in computer architectures and the second is the emergence and hyper-rapid commercialization of generative AI.
Generative AI, now a household term, requires architectures that scale, which is driving huge demand for networked AI factories or data centers like the ones Nvidia and SoftBank partnered to build.
The challenges of AI are daunting, but they are all within the bounds of our imagination. The return on effort will be compounded as innovations made for AI migrate to all other forms of computing.
#SoftBank #Nvidia #redesign #architecture #workloads
