Microsoft confirmed the acquisition of Fungible, a promising startup that manufactures data processing units (DPUs).
In December, reports emerged that Microsoft had acquired Fungible, a startup fabricating datacenter hardware known as data processing units (DPUs). The deal was estimated to be worth about $190 million. Microsoft confirmed the reports on Monday. It did not comment on the price tag.
The tech giant plans to use the startup’s technology and experts to deliver multiple DPU solutions, as well as network and hardware system enhancements.
“Fungible’s technologies help enable high-performance, scalable, disaggregated, scaled-out datacenter infrastructure with reliability and security”, Girish Bablani, CVP of Microsoft’s Azure Core Division, said in a blog post.
Microsoft and DPUs
Bablani added that the announcement “further signals Microsoft’s commitment to long-term differentiated investments in our datacenter infrastructure, which enhances our broad range of technologies and offerings, including offloading, improving latency, increasing datacenter server density, optimizing energy efficiency, and reducing costs.”
DPUs are hardware units specifically designed to handle particular data processing activities, such as network traffic routing and data security. The method aims to lessen the burden placed on CPUs and GPUs when performing core computing activities associated with a particular workload.
Fungible was founded in 2016 by Pradeep Sindhu, co-founder of Juniper Networks, and Bertrand Serlet, former senior vice president of software engineering at Apple.
Fungible offers DPUs that manage flash storage volumes using a MIPS microprocessor architecture and two operating systems, one proprietary and the other open source.
Microsoft’s acquisition of Fungible may help the tech giant strengthen Azure by offering new DPU technology in subscription services and tier-based block storage solutions. The deal follows Microsoft’s December 2022 acquisition of Lumenisity, a startup that produces high-speed fiber solutions for datacenters.