Hari Gopalkrishnan on AI infrastructure
From Bank of America Chief Technology and Information Officer Hari Gopalkrishnan at Semafor World Economy · · Semafor Events
“First thing is data. I mean, you have to know who your clients are. You have to know everything you know about them. Interactions, transactions on us, off us data set. So data is truly the fuel behind all the work we're doing around AI and along with that compute. These models aren't cheap. They're not inexpensive. They are take a lot of hardware to run. So making sure you have your data and compute strategy. And along with this, understand the economics of what it takes to run these things. It's very easy to spend a lot of money and find out that you got nothing in return.”
On , Hari Gopalkrishnan, Chief Technology & Information Officer at Bank of America, spoke about AI infrastructure during Bank of America Chief Technology and Information Officer Hari Gopalkrishnan at Semafor World Economy on Semafor Events.
About Hari Gopalkrishnan
At the Semafor World Economy event on April 30, 2026, Bank of America Chief Technology and Information Officer Hari Gopalkrishnan discussed the institution's approach to artificial intelligence deployment. He described a shift from proof-of-concept projects focused on small tasks to end-to-end process transformation aimed at revenue, client experience, or expenses. Gopalkrishnan stated that the bank uses a prioritization model combining top-down business strategies with bottom-up ideas from over a hundred listening sessions with employees, generating thousands of suggestions on reducing toil-based activities. He said these are prioritized against $4 billion in new investments. Gopalkrishnan also highlighted the bank's focus on data, compute, security, resiliency, governance, and human oversight as crucial for AI in finance. He noted that Bank of America has established an academy for reskilling and upskilling on AI, covering topics from basic prompt engineering to advanced development, and that 44% of jobs in recent years were filled through internal mobility. He expressed optimism about AI's potential to improve client service through data, personalization, and digitization.