There was a tremendous amount of news delivered at this year's IBM THINK 2019, which I personally attended along with my senior storage analyst colleague Steve McDowell. One thing that impressed me, which didn’t get much press, was the number of rich conversations from IBM customers on how they were using Power Systems to help their AI, ML and DL workflows. It's hard enough to get customers to talk about much of anything on stage with detail, and I was impressed at the depth and breadth of those discussions. Let’s take a closer look.
JP Morgan Chase & Co
JPMorgan Chase & Co was one of the companies on site discussing its experience with IBM’s Power Systems. JPMC manages $2.6 Trillion in assets and operates in 100 countries. Executive Director Elenita Elinon spent time on stage talking about how the bank utilizes Machine Learning on IBM’s Power Systems, saying that it chose IBM’s systems because they work better on larger datasets like security, fraud, regulatory, and rogue trades. Power’s scale-up capabilities are unparalleled and work the best right now with accelerators given its flexibility and memory coherency capabilities.
Morgan Stanley
Morgan Stanley was another customer that showcased its work with IBM Power Systems at the event. Morgan Stanley executive director Marcelo Labre speaking with IBM's Sumit Gupta says that IBM Power Systems’ computing power and AI-readiness is enabling the organization to explore new AI/ML use cases in finance, with the overall goal of increased efficiency and alignment with customer needs. For example, Morgan Stanley’s Labre elaborated at THINK 2019 on how his organization is utilizing AI to challenge outdated risk models. Using AI to improve risk models is a common theme I hear over and over in the industry. You truly need big data to do this well and Power fits the bill. IBM's Thomas Harrer captured the action well here in a tweet.
Fannie Mae
Also in attendance were representatives from mortgage loan company Fannie Mae. The company says it relies on high-performance computing for the risk management of its massive mortgage portfolio (which totals roughly $3.2 trillion). Fannie Mae shared at THINK 2019 that it has built a heterogeneous computing framework on IBM’s Power Systems that purportedly reduces its infrastructure costs by as much as 90%. Heterogeneous computing is what Power was made for, and I call it the “Swiss Army knife of acceleration” given its highest bandwidth and I/O flexibility with coherent and non-coherent options.
NASA Frontier Development Lab
Another cool customer story came from NASA’s Frontier Development Lab, who shared that it is utilizing IBM POWER9 to develop AI solutions (DL and DNN) to understand and predict how solar radiation interacts with and damages satellites, GPS systems, and power grids. Space radiation can interfere, and some cases take completely offline GPS and RF, the basis for all communications and positioning. Think cars and cellphones, important stuff. Space weather forecasting—the future is here, folks.
Wrapping up
As I said earlier, it was great to hear IBM’s customers open up a bit about exactly how Power Systems is being utilized within their organizations. It's hard to get customers to open up about anything like this. These were just four Power Systems customers doing solving real problems, but I hear there are nearly one hundred like this. I’m excited to hear more.
Taken altogether, these stories show me that Power is being adopted by major enterprises for their AI workloads and that it is no longer in the tire-kicking stage. AI is real and here to stay, and IBM’s Power Systems have a crucial role to play. IBM says it best when it says, “AI is not magic.” It’s here today.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.
Disclosure: Moor Insights & Strategy, like all research and analyst firms, provides or has provided paid research, analysis, advising, or consulting to many high-tech companies in the industry, including Advanced Micro Devices, Apstra, ARM Holdings, Bitfusion, Cisco Systems, Dell EMC, Diablo Technologies, Echelon, Ericcson, Frame, Gen Z Consortium, Glue Networks, GlobalFoundries, Google (Nest), HP Inc. HewlettPackard Enterprise, Huawei Technologies, IBM, Jabil Circuit, Intel, Interdigital, Konica Minolta, Lenovo, Linux Foundation, MACOM (Applied Micro), MapBox, Mavenir, Mesosphere, Microsoft, National Instruments, NOKIA (Alcatel Lucent), Nortek, NVIDIA, ONUG, OpenStack Foundation, Peraso, Portworx, Protequus, Pure Storage, Qualcomm, Rackspace, Rambus, Red Hat, Samsung Technologies, Silver Peak, SONY, Springpath, Sprint, Stratus Technologies, TensTorrent, Tobii Technology, Synaptics, Verizon Communications, Vidyo, Wellsmith, Xilinx, Zebra, which may be cited in this article.
