Oracle announced two additions to its cloud plans on October 27: access to Google’s Tensor Processing Units and a native Snowflake AI Data Cloud service. The agreements cover computing hardware and enterprise data services on Oracle Cloud Infrastructure, or OCI.
Oracle said Google’s AI accelerators would become available through OCI in calendar 2027, with capacity increasing over several years. The announcement did not specify an initial region or exact launch date.
Under that arrangement, Oracle would supply the power, cooling and data-center networking for Google’s chips within its infrastructure. The companies described the processors as options for training AI models and running them after training. Oracle said customers would be able to choose among different computing architectures for their workloads.
The Snowflake arrangement concerns how companies purchase and use analytics and AI data services. According to the announcement, customers would buy and provision Snowflake directly through Oracle, which would handle pricing, discounts and billing. Customers would also be able to apply Oracle Universal Credits toward usage.
Oracle listed supported enterprise-data sources including Fusion Cloud Applications, NetSuite, Autonomous AI Database, MySQL HeatWave and NoSQL Database. It said the integrations are intended to reduce custom data pipelines and movement between clouds.
The releases describe planned services and company expectations. They do not provide independent benchmarks proving lower costs or better performance for a particular customer. A specific availability date for the native Snowflake offering was not supplied in the announcement reviewed.
Sources: Oracle — Google TPU announcement: https://www.prnewswire.com/news-releases/oracle-brings-google-tpus-to-oracle-cloud-infrastructure-302920827.html Oracle — Snowflake announcement, distributed by PR Newswire: https://www.stocktitan.net/news/ORCL/oracle-expands-native-ai-and-data-services-on-oci-with-xiv8roizf01k.html