Generative AI Economic Value in Banking Statistics
McKinsey estimates generative AI could add $200 to $340 billion in annual value to global banking; Citi puts the 2028 profit uplift at roughly 9%, or $170 billion. Both are projections, not audited outcomes. Each figure below names its source and year; none was generated by this site. These are analyst estimates built on adoption and productivity assumptions: read them as a ceiling under those assumptions rather than a committed forecast.
On This Page
Statistics
The numbers worth quoting
Generative AI could add the equivalent of 200 to 340 billion US dollars in annual value to the global banking sector
The estimate corresponds to 9 to 15 percent of operating profit, with the largest projected gains in software engineering, customer operations, and marketing. It is a potential-value estimate under McKinsey's assumptions.
Citi estimated AI could boost the banking industry's 2028 projected profits by about 9 percent, or 170 billion US dollars, from just over 1.8 trillion to close to 2 trillion
Citi's figure draws on a Treasury and Trade Solutions survey of 90 respondents (banks, insurers, asset managers) in April 2024. It is a projection of profit-pool uplift, not a measured result.
65% of respondents said their organizations regularly used generative AI in at least one business function, up from about one third a year earlier
This is the cross-industry adoption baseline for regular generative-AI use; McKinsey's survey included financial-services respondents among 14 sectors. The near-doubling year over year frames how fast value capture is being attempted.
37% of surveyed financial firms cited report generation, synthesis, and investment research as a leading generative-AI use case, and 34% cited customer experience and engagement
These were the top generative-AI use cases by respondent interest, consistent with McKinsey's finding that value concentrates in document-heavy and customer-facing functions.
55% of surveyed financial firms reported actively seeking generative-AI workflows for their companies
The figure reflects demand intent among NVIDIA's surveyed financial-services respondents, a sample skewed toward firms already investing in AI.
Key Takeaways
Methodology
Figures are drawn from named research firms (McKinsey, Citi) and a vendor survey (NVIDIA), each reported with its source and year. Value figures are forward-looking projections under each source's assumptions, not audited outcomes, and are labelled as such. No statistic on this page is derived from data collected by this site.
Try These Tools
Run the numbers next
Earnings-Call Summarization Cost Calculator
LLM cost per stock per quarter to summarize earnings transcripts across Sonnet, Opus, GPT-4o, Gemini 2.5 Pro/Flash. Cache-hit-rate aware. Snapshot pricing.
Model Selector for Finance
Input task, latency budget, cost budget, context size, and quality sensitivity; get ranked model recommendations with rationale — grounded in published.
Agent Cost Envelope Calculator
Model an LLM research loop end-to-end — steps, tool calls, convergence checks, markets per day — and see per-loop, daily, and monthly cost with cost-cap.
Sources & References
- The economic potential of generative AI: The next productivity frontier — McKinsey & Company (2023)
- Citi Publishes New Report on AI in Finance — Citi GPS (2024)
- The state of AI in early 2024: Gen AI adoption spikes and starts to generate value — McKinsey & Company (2024)
- AI Takes Center Stage: Survey Reveals Financial Industry's Top Trends for 2024 — NVIDIA (2024)
Related Content
Keep the topic connected
AI in Finance Adoption Statistics
Sourced statistics on AI and generative-AI adoption across financial services: deployment rates, use cases, governance gaps, and where the money is going.
Generative AI Spend in Financial Services Statistics
Generative-AI spend statistics for financial services: AI budget share, ROI self-estimates, bank tech-budget figures, and production use-case counts.
LLM Cost Trends for Finance Workloads Statistics
LLM cost trends for finance workloads: sourced per-token price declines, fold reductions over time, and the drivers behind falling inference costs.