IBM’ strategic partnership with NVIDIA
IBM (IBM) has invested billions of dollars in machine learning and AI (artificial intelligence) through partnerships and acquisitions to develop its Watson platform, which is aimed at the AI space. On the heels of this strategy, IBM has partnered with NVIDIA (NVDA) to promote a computer’s capacity to “think and learn in more human-like ways.”
NVIDIA GPUs (graphics processing unit) are seeing increased adoption by supercomputers and cloud companies for AI tasks.
Under this partnership, IBM aims to roll out a software toolkit, IBM PowerAI for IBM servers, which contains NVIDIA’s Tesla accelerator cards. NVIDIA’s Tesla processors are used in deep learning technology. Recently, Intel (INTC) unveiled the latest version of its Xeon Phi processors to compete with NVIDIA’s Tesla processors.
In September 2016, IBM launched a new server, the Power S822LC, primarily for AI deep learning and other HPC (high-performance computing) workloads. The new server will use IBM’s POWER architecture. IBM’s POWER architecture, a part of its OpenPOWER Foundation, was built and launched specifically for big data analytics, cloud, and HPC workloads.
IBM’s newest server is the only server available in the market that completely integrates NVIDIA’s NVLink technology. NVDA can use its NVLink fabric to accelerate the data transfer at a speed of up to 80 GBps (gigabytes per second) of bandwidth. Apart from NVIDIA’s GPUs, IBM’s Power8 is the only processor available that includes built-in NVLink support.
Forecasts and expectations from AI and the deep learning space
Salesforce (CRM) is set to become a $10 billion dollar company by 2018. Its cofounder and CEO, Mark Benioff, recently stated, “What I see is an AI first world, and for every customer, going to be able to get a whole another generation of productivity out of artificial intelligence, machine learning, and deep learning.”
Tractica, a market intelligence company, expects AI to become a part of every industry, replacing existing business models with new ones that rely on deep learning. Tractica forecasts that global AI revenue could rise from $643.7 million in 2016 to $36.8 billion by 2025.