Citi moves vital infrastructure to Google Cloud as part of broader AI push
In This Article:
Citigroup today announced a multiyear strategic partnership that will see the bank move parts of its vast financial infrastructure to a platform hosted by Alphabet subsidiary Google Cloud. New York–based Citi, the fourth largest bank in America by assets, currently relies heavily on hardware. It expects the integration to a virtual environment will make it easier to customize its financial pipes and respond to consumers' ever-changing demands.
As part of the partnership, Balaji Kumar, Citi’s head of global technology infrastructure, told Fortune, the bank will also be using Google Cloud’s Vertex AI platform to build a number of financial tools for customers using artificial intelligence. Though details about which applications might launch first weren’t disclosed, Kumar says “quite a few” pilots are already in the works.
“It's a massive opportunity for us to engage in technology modernization,” says Kumar. The infrastucture upgrade is part of a larger effort at Citi to better capitalize on the bank’s financial pipes, which power much of the global economy, and comprise half of the bank’s revenue.
Though the partnership is still in its early days, Rohit Bhat, the managing director of Google Cloud’s financial services division, says that Google Cloud’s engineering, product, and sales teams have been working with Citi’s technology and business enablement team for months and much of the technology they’ve codeveloped is quite mature.
Top among the Google Cloud upgrades at Citi is a high-performance-computing (HPC) capability that will run millions of risk calculations a day for Citi’s Markets business. The software is expected to reduce latency—or delays—by preemptively checking transactions to ensure they have enough compute power and storage to execute.
Through the partnership Citi also has access to Vertex AI, the enterprise version of Google's proprietary Gemini AI, and other open source large language models (LLMs), including Gemma, Meta, and Mistral. In the future, the bank anticipates using the AI tool to complete other pilots currently being developed, including for customer service and call centers, document digitization, and customized marketing tools.
Implementing cloud and AI tools give rise to new security considerations, ranging from where the data comes from to who can access it. To address those concerns, Kumar says, "every use case is governed and monitored and approved" before it is rolled out for public use. That includes careful analysis of the kind of data used, where is it is hosted, and whether the data is being stored for future use or being actively used.