RESEARCH TRIANGLE PARK – IBM unveiled details about the company’s upcoming Telum Processor, which includes “on-chip accelerated AI” according to a company spokesperson, at the annual Hot Chips conference, and in a statement issued this week.

The chip, designed by an IBM team headquartered in Poughkeepsie, New York in partnership with the IBM Research AI Hardware Center in Albanywhich created the underlying technology, may be able to identify, quickly, fraudulent activity when consumer credit cards are misused or abused, the company described in a blog post.

“With Telum, financial institutions will be able to move from a fraud detection to a fraud prevention posture, catching instances of fraud while the transaction is still ongoing,” the company stated in its blog post.

The chip will be manufactured by Samsung, a company spokesperson told WRAL TechWire in an interview, and the research and development process took about three years.  The chip will be a part of what IBM called in a press release a “Telum-based system,” available at some point during 2022, according to the company.

Other use cases for the chip technology, contained within a Powerpoint presentation shared with WRAL TechWire by IBM, include potential customer behavior prediction, supply chain optimization, anomaly detection to enhace security, and workload placement.

IBM-Red Hat deal with Palantir is big boost for its artificial intelligence, cloud strategy

The Telum processor, IBM’s first that contains on-chip acceleration for artificial intelligence interfacing, adds new neural network processing assist instruction, according to the IBM presentation shared with WRAL TechWire.

The company stated that the Telum processor can enable tens of thousands of AI-infused transactions per second, noting its features in its construction: “eight processor cores, running with more than 5GHz clock frequency, optimized for the demands of enterprise-class workloads. The completely redesigned cache and chip-interconnection infrastructure provides 32MB cache per core. The chip also contains 22 billion transistors and 19 miles of wire on 17 metal layers.”



The company said in a statement that the result of the AI-accelerator is that it will “enable applications to run efficiently where the data resides, helping to overcome traditional enterprise AI approaches that tend to require significant memory and data movement capabilities to handle inferencing.”

The company lists a variety of potential industry applications in its presentation, including financial transactions and banking, retail, healthcare, hospitality, insurance, and transportation.

“Telum follows IBM’s long heritage of innovative design and engineering that includes hardware and software co-creation and integration that spans silicon, system, firmware, operating systems and leading software frameworks,” a company spokesperson told WRAL TechWire.

“We see Telum as the next major step on a path for our processor technology, like the inventions of the mainframe and servers before,” IBM stated in its blog post.  “The challenges facing businesses around the world are growing in complexity with every passing day, and Telum will be a building block for solving these problems for years to come.”

IBM’s Red Hat cloud play is ‘doing fine,’ but analyst sees bigger challenges