IBM, Google Cloud build AI infrastructure

IBM and Google Cloud said on April 23 they were expanding their partnership to package enterprise AI and hybrid-cloud tools as a combined enterprise stack rather than as separate infrastructure products. IBM described the effort as a way to connect data, models, infrastructure and operations across environments, while Google Cloud used its Next ’26 conference to pitch what CEO Thomas Kurian called a “fully integrated AI stack.” IBM’s announcement centered on software and control layers that large companies use to run AI in production. The products named as available now include watsonx.data on Google Cloud Marketplace, HashiCorp Terraform Enterprise, Vault and Consul on the marketplace, and Red Hat OpenShift through the Google Cloud Console and marketplace. The collaboration matters because it shows where enterprise AI spending is moving. (newsroom.ibm.com) IBM framed customer demand around hybrid flexibility, predictable operations and “trusted AI,” while Google Cloud’s April 22 keynote argued that customers no longer want “fragmented systems.” Those are vendor claims, but they point to the same sales pitch: enterprises want packaged combinations of data, governance, security and deployment tools, not just raw compute. ### What, exactly, are IBM and Google Cloud putting together? IBM said the current package combines Google Cloud’s AI and data services with IBM’s hybrid-cloud, data-management and automation software. In practice, that means enterprises can buy IBM watsonx.data through Google Cloud Marketplace, use HashiCorp tools for policy-driven provisioning and secrets management, and run Red Hat OpenShift workloads with Google Cloud billing and committed-spend credits. (newsroom.ibm.com) Google Cloud’s side of the story is its broader platform push. Thomas Kurian said at Next ’26 that Google Cloud was offering a vertically integrated stack spanning models, infrastructure, data and security, with products including Gemini Enterprise, an agent platform, custom processors and cross-cloud lakehouse tools. ### Why is watsonx.data one of the first products named? IBM singled out watsonx.data because enterprise AI projects often stall on data access, governance and retrieval rather than on model availability. (newsroom.ibm.com) IBM describes watsonx.data as a hybrid, open data lakehouse for AI and analytics with source-level access controls and hybrid retrieval-augmented generation support. On Google Cloud, IBM says, customers can use existing cloud commitments to buy it through the marketplace. (cloud.google.com) IBM has been building its broader AI pitch around that data layer. At its Think conference in May 2025, IBM said new watsonx.data capabilities could help produce more accurate AI agents and tied the product to governance, observability and deployment across hybrid environments. ### Where does infrastructure fit if this is not mainly about chips? Google Cloud’s infrastructure still sits underneath the partnership. (newsroom.ibm.com) Kurian used Next ’26 to highlight TPU and processor rollouts, storage and networking upgrades, and usage figures including more than 16 billion tokens per minute via direct API use and nearly 75% of Google Cloud customers using its AI products. (newsroom.ibm.com) IBM’s announcement, though, emphasized the layer above hardware. The company listed identity security, orchestration, virtualization, data architectures and developer tooling, including Red Hat Lightspeed Agent for Google Cloud, as the pieces enterprises need to make AI systems usable across on-premises and multiple clouds. ### What are the companies saying comes next? (cloud.google.com) IBM said future work includes integrating Google’s AI offerings, including Gemini models and Gemini Enterprise, with key parts of IBM’s software portfolio. The company did not give launch dates in the April 23 post, but it presented those integrations as the next stage of the partnership. (newsroom.ibm.com) Google Cloud’s public materials point readers to the partner directory, AI partner pages and Next ’26 product announcements for updates, while IBM’s partnership page lists the current joint products and consulting services. For now, the named participants in the next step are IBM software teams, Google Cloud’s Gemini product group and joint enterprise customers buying through Google Cloud Marketplace. (ibm.com) (newsroom.ibm.com)