HSBC and Google Cloud announced an expanded multi-year AI partnership on 17 June 2026, extending a relationship that now spans more than 600 of the bank's applications running on Google's infrastructure. The agreement gives HSBC access to Gemini models and the Gemini Enterprise Agent Platform, with forward-deployed engineering teams from Google Cloud and Google DeepMind working directly alongside HSBC staff. The programme covers three initial delivery areas: hyper-personalised wealth management support, financial crime risk management and AI tools for frontline staff and relationship managers. HSBC expects it to enable more than 200 new AI use cases over two years. HSBC has committed to a group return on tangible equity (RoTE) of at least 17% through 2028 and is investing in AI as one of the primary levers to reach it. Only use cases estimated to return more than $100 million, in revenue or efficiency, will be prioritised for delivery. The $100 million bar HSBC has set for itself At the fourth-quarter 2025 earnings call, Group Chief Executive Georges Elhedery identified generative AI as the bank's single largest new technology investment area. In its first-quarter (Q1) 2026 results, HSBC reported operating expenses of $8.7 billion for the quarter, up 8% year-on-year, with technology spend among the cited drivers. The bank has committed to cost growth of approximately 1% on a target basis for full-year 2026 and is retiring 3,000 of its more than 9,000 applications by 2028. HSBC's RoTE of 18.7% in Q1 2026, excluding notable items, is its strongest reading in nearly two decades. Under its most severe modelled stress scenarios, RoTE could fall below the 17% medium-term target without mitigation. HSBC's AI vendor strategy: Google Cloud and Mistral AI Six months before the Google Cloud announcement, in December 2025, HSBC finalised a separate multi-year partnership with Mistral AI. That agreement gives the bank access to Mistral's commercial models on a self-hosted basis, running on HSBC's own internal systems, for use cases including financial analysis, multilingual reasoning, translation and prototyping. HSBC noted at the time that it is "continually assessing a wide range of large language models as part of its technology strategy." The Google Cloud deal operates on different foundations: Google's cloud infrastructure, Gemini models and the Gemini Enterprise Agent Platform, with forward-deployed engineering teams from Google Cloud and Google DeepMind working directly alongside HSBC staff. Self-hosted models on internal infrastructure and cloud-native agentic platforms carry different implications for data governance, model risk oversight and regulatory reporting, particularly for an institution operating across 56 countries and territories. The bank appointed David Rice as its first Chief AI Officer in March 2026, effective 1 April, and concurrently expanded Chief Technology Officer Mario Shamtani's remit to include building a central AI platform for employee access to models. Both appointments were made in March 2026, after the Mistral partnership was finalised in December 2025. Why HSBC is prioritising AI in wealth management Wealth fee income grew at a double-digit rate for at least five consecutive quarters through early 2025 and remained a primary driver of Q1 2026 revenue. HSBC serves wealth clients predominantly in Asia, where relationship manager capacity and the ability to serve clients across borders are competitive differentiators. HSBC said the AI-powered decision assistant already in use for frontline staff is reducing prep time "from hours to minutes" for thousands of users, building on something already in production. The bank is also working to codify regulatory procedures into an AI structure to provide relationship managers with structured options and analysis for decision-making. HSBC AI governance: the challenge of scaling across two platforms HSBC's AI programme now encompasses a self-hosted European LLM partnership, a cloud-native agentic platform with the world's largest cloud provider, 100 active generative AI use cases recorded in its 2025 Strategic Report and AI productivity tools available to 85% of its employees. How models are selected, validated, monitored and retired across jurisdictions remains outside the scope of either announcement. TABInsights research across 29 G-SIBs has found that leading institutions sequence platform and data infrastructure ahead of individual AI use cases, embedding governance into that architecture from the outset rather than retrofitting it. HSBC has the use case ambition and the vendor relationships. Whether the governance architecture matches that ambition will become clearer as the 200-use-case programme progresses through 2026 and 2027.
