This 'Outdated' IBM Technology Just Did Something It Hasn't Done in 20 Years

In the property of unreality computing, it's casual to hide that IBM (NYSE: IBM) inactive sells its hulking mainframe systems. Following IBM's fourth-quarter net report, it's wide that the mainframe concern is not lone surviving but besides thriving.

Here's an unthinkable information IBM CFO Jim Kavanaugh disclosed during the net call: IBM's mainframe concern recorded its champion 4th fourth gross successful much than 20 years. Revenue soared 61% twelvemonth implicit year, adjusted for currency, driving a 17% summation successful the infrastructure segment.

Where to put $1,000 close now? Our expert squad conscionable revealed what they judge are the 10 champion stocks to bargain close now, erstwhile you articulation Stock Advisor. See the stocks »

IBM's mainframe systems are inactive wide used, peculiarly successful industries that request utmost reliability and security. 71% of Fortune 500 companies usage mainframes, according to Market Reports World. About 92% of ample fiscal institutions and 63% of authorities agencies usage mainframe systems successful their operations, with IBM accounting for much than 90% of installed mainframe systems. Amazingly, implicit 87% of planetary recognition paper transactions are processed connected mainframe systems.

How person mainframes remained applicable for truthful long? IBM has repeatedly iterated, upgraded, and improved its systems to conscionable lawsuit needs. The latest z17 mainframe systems, which are driving grounds results, are tailor-made for the AI era.

IBM's z17 mainframe strategy is designed to grip much than 250 AI usage cases, including indebtedness hazard mitigation, chatbot services, aesculapian representation analysis, and retail transgression prevention. The z17 tin grip up to 450 cardinal AI inferencing operations per day, up 50% from its predecessor. Importantly, those AI operations person mean effect times of conscionable 1 millisecond, making the z17 perfect for real-time operations wherever latency isn't acceptable.

While the z17 has plentifulness of on-board AI processing power, IBM besides sells the Spyre AI accelerator. The Spyre accelerators tin beryllium installed successful z17 mainframe systems, enabling customers to tally much almighty AI models, including IBM's Granite models.

IBM has a wide thought of wherever the AI inferencing marketplace is heading. While astir AI workloads are moving successful the unreality today, IBM CEO Arvind Krishna expects that to alteration drastically. "...if I look retired 3 to 5 years, 50% of the endeavor usage of AI going to beryllium successful either a backstage unreality oregon is going to beryllium successful their ain information centers. And the different 50% is gonna beryllium usage of nationalist models," Krishna said during the fourth-quarter net call.