VENDORiQ: Google’s Sovereign AI push

Why it Matters

For some time, Google Cloud Platform (GCP) has been a capable, if often overlooked, alternative to its main cloud rivals. Its latest announcements, however, suggest a more aggressive strategy centred on differentiated AI capabilities. Google’s doubling down on Agentic AI for outcomes, aligning with a broader industry trend where the focus is shifting from building complex infrastructure to achieving specific business outcomes. 

The move to offer sovereign AI processing within Australia for Gemini 2.5 Flash is a significant development for the local market. This directly addresses sovereign creation and refinement of AI, a major hurdle for government, finance, and healthcare sectors considering global offshore cloud AI services. While competitors have local data centres, this specific on-shore AI model processing capability is a notable differentiator that could compel organisations to reconsider GCP for sensitive workloads. 

IBRS client inquiries confirm that hyperscale clients are increasingly concerned about the dominant cloud AI services that may have model processing, and specifically training and refining activities, stored outside of Australia, even if the client’s data is said to remain locally. This is becoming a concern for regulators like the ACCC.  

These announcements collectively bolster the argument that cloud vendor selection should not be based solely on legacy experience or developer familiarity. While Azure and AWS hold dominant market positions and provide a high degree of domestic AI capability, Google’s deep investment in AI, backed by its vast data assets from Search, and an $85 billion investment commitment this year across Alphabet, suggests its platform is not only able to demonstrate greater value for basic cloud services, but highly sophisticated data and AI services.

Who’s Impacted?

• Software Development and Cloud Architecture Teams: The availability of an expanded agent ecosystem via Vertex AI, and sovereign ML training and processing is compelling. Teams should assess the capabilities for projects where generative media, advanced automation, or strict data residency are requirements.
• Data & Analytics Leaders: The keynote reinforced the tight coupling between Google’s data platforms, like BigQuery and its AI tools. The ability to run AI on local data, as demonstrated by The Iconic processing millions of records to cut query times from minutes to seconds, presents a compelling case for modernising data strategies.
• CFO: Google’s competitive pricing, coupled with its data and AI capabilities, challenges the notion of selecting a cloud provider based on past expenditure or existing skill sets. Decisions should factor in the future value of integrated, sovereign AI capabilities against the potential long-term costs of being locked into a vendor less differentiated in this space.

Next Steps

Organisations should consider the following actions:

• Task architecture teams to evaluate the ‘AI agent’ framework on Vertex AI. Compare the development effort and outcome of a sample workflow (e.g., automated customer service response) against existing solutions.
• Conduct a comparative analysis of Google Cloud Platform against incumbent vendor offerings for a relevant business use case, such as marketing content creation, or when considering ERP migrations.
• Re-evaluate multi-cloud strategies to explicitly consider which platform offers the most mature and compliant AI services for future-state business processes, rather than treating cloud providers as interchangeable infrastructure commodities.