Boston Consulting Group (BCG) finds that the financial sector delivered an above‑industry performance in 2025 with a 30.2% total shareholder return (TSR), creating an opening for sustained value creation. However, investors have not fully priced in durable growth. The coming years will test whether financial institutions allocate free capital into technology‑driven growth, AI‑based operating models and disciplined acquisitions rather than predominantly to buybacks and dividends.
What happened in 2025 — numbers and regional patterns
Financial institutions achieved an average TSR of 30.2% in 2025, outperforming all other industries. The improvement was driven mainly by stronger P/B (price‑to‑book) ratios, while P/E (price‑to‑earnings) ratios remained largely unchanged — indicating markets remain cautious about sustainable growth prospects.
Performance was uneven: over a three‑year horizon, specialized and digital banks led, while small and mid‑sized institutions and traditional card acquirers lagged. Regionally, Europe, Japan and South Korea stood out, partly because these markets had accumulated significant catch‑up potential.
Capital returns (dividends and buybacks) further boosted realized returns. For the first time in years, a majority of global bank stocks traded above book value, though more than half of bank stocks — including many smaller players — still traded below book value.
Drivers of profitability recovery
BCG stresses that the recovery in profitability is not solely the result of a normalized interest rate environment and wider net interest margins. Disciplined cost management, prudent risk control and improved asset–liability management played at least as important a role. Better liquidity and interest‑rate risk management reduced net interest income volatility and supported steadier operations.
Fee income, however, has not broadly returned to earlier higher levels and remains low on an asset‑adjusted basis in several regions, underlining the strategic challenge of growth beyond interest income.
Several structural supports exist: many banks use structural hedges to limit downside to net interest margins, and regulatory loosening could free lending capacity. BCG notes that planned U.S. capital reforms could unlock up to about $2,600 billion of additional capacity. The sector’s market capitalization has grown by roughly $5,500 billion since the end of 2022.
Nevertheless, on P/E the sector still trades at a material discount — roughly 40% below the broader market — showing investors are not yet fully convinced banks can sustain higher growth.
AI as an operating model, not just a tool
A central BCG conclusion is that AI can drive structural change across banking, rewriting processes, productivity and cost structures rather than merely delivering incremental efficiency gains.
Bank operations remain heavily human‑dependent across data reconciliation, document summarization, customer handling and many back‑office tasks. Traditional digitization has not delivered the anticipated breakthrough: operating costs relative to assets have barely declined, and global sector headcount has grown about 2% per year.
BCG’s survey shows banks plan to invest roughly 2% of revenues in AI by 2026 — approaching tech‑sector levels. The crucial issue is not just spending levels but whether investments are directed to truly transformative programs.
BCG identifies four structural shifts: synthetic voices rapidly becoming competitive with human interaction; AI agents with long‑term memory executing complex workflows autonomously; personal AI assistants delivering radical transparency in consumer financial decisions and pressuring margins; and a surge in AI‑enabled fraud and cybercrime requiring AI‑based defenses.
BCG provides examples of realized impact: an Asian bank used an agent‑based AI architecture to free over 30% of wealth‑adviser capacity, producing a 30% rise in fee income and triple customer engagement; a European bank’s AI agents boosted credit officer productivity by 50%, cut decision times to 24 hours and improved fraud detection by over 30%; a U.S. big bank used AI developer assistants to raise developer productivity by an average of 30%, up to ~60% for top teams.
Capital allocation: return to shareholders vs reinvestment
BCG argues most institutions now earn returns above their cost of capital, forcing a strategic choice: continue buybacks/dividends or redeploy excess capital into growth and AI transformation.
Base‑case projections through 2030 show mixed results for traditional banking: retail and small‑business banking may grow around 4% per year (with AI‑driven pricing transparency potentially reducing this to ~2%); corporate and investment banking about 5%; and commercial banking roughly 6% annually. These are solid rates, but may not suffice to materially improve P/E multiples versus fast‑growing digital challengers.
For example, Revolut grew its customer base by more than 40% in 15 months and exceeded 70 million customers by 2026 while expanding into full banking services.
For institutions trading above book value, sustainable shareholder returns may rest on two pillars: (1) gaining share through AI‑based innovation, and (2) portfolio reshaping via disciplined acquisitions. Technology adoption alone is not a durable advantage because the industry can copy the tech; advantage comes from using AI to structurally change operating models.
In retail banking, personalized pricing could lift revenues 2–3% in the near term. Medium‑term AI‑driven cost reductions could open new markets — making wealth management available to upper‑middle income segments and profitable small‑ticket lending in emerging markets. Digital custody and trading of digital assets could become material revenue streams if regulatory clarity emerges.
In commercial banking, AI can broaden access to mid‑market investment‑banking services, simplify trade finance, and scale treasury and liquidity solutions. In corporate and investment banking, the technology is more about freeing capacity than creating brand‑new revenues: bankers’ client‑facing time and back‑office capacity could be redeployed by 25–40% and 20–35% respectively within five years.
Three megatrends and possible futures
BCG identifies three converging megatrends shaping the financial system: AI, growth of non‑bank financial intermediaries, and the spread of digital assets. Their intersection will likely drive the most disruptive change.
Non‑bank financial intermediaries have become structurally embedded: their share of corporate and investment banking revenues rose from 9% in 2019 to 16% in 2024, and the base case sees that rising toward 22% by 2030. In trading, non‑bank liquidity providers already account for roughly 26% of global trading revenues. Private credit’s expansion is less certain today than a few years ago; non‑bank lending currently represents around 11% of credit but faces funding pressures and has not yet been tested through a full downturn cycle.
Digital asset markets — stablecoins and tokenized instruments — are scaling rapidly and regulatory frameworks such as the EU’s MiCA and the U.S. GENIUS Act are shaping the landscape. BCG’s base case expects digital assets to become embedded in the financial system within five years, initially in niche areas such as tokenized fund settlement, cross‑border B2B payments and securities settlement.
The biggest change would come if programmable financial infrastructure (stablecoins, tokenized deposits and real‑asset tokens) links with autonomous AI agents that can initiate, execute and manage payments, loans and asset transfers — enabling richer, faster and lower‑cost lending and broader retail access to private markets via non‑bank platforms.
BCG sketches three scenarios: the “era of specialization” where AI‑driven transparency squeezes margins and category‑specific expertise becomes the advantage; a “hyper‑granularity” world where many digital money forms coexist at scale; and a “consumer‑savings disruption” scenario where non‑banks directly challenge banks’ deposit dominance via tokenized private credit and embedded AI lending.
What makes AI transformation succeed?
From more than 100 AI transformation projects, BCG distills that institutions reaching enterprise‑level impact concentrated resources on a few high‑value initiatives. Project selection should prioritize: value creation (meaning real revenue lift or cost reduction), sustainability of competitive advantage, reusability of infrastructure and models, and feasible time horizons. Typically 30–50 initiatives filter down to six–eight true priorities.
Five readiness factors govern agent‑based AI adoption: a common platform layer with agent orchestration and built‑in guardrails; data quality, machine‑readability and semantic structure over sheer volume; AI‑specific risk‑management and compliance frameworks; an in‑house AI delivery office inside the operating model; and talent management focused on people and processes more than just technology.
Common failure modes include treating AI as an annual IT budget line, IT ownership rather than business ownership, and automating existing processes instead of redesigning them for AI. BCG recommends the CEO own the AI portfolio, central and continuous funding, and linking value metrics to executive compensation. The largest productivity gains come not from making an old process faster but from re‑architecting it.
Conclusion — the window is open, but action is required
BCG’s analysis concludes that the strong 2025 financial performance has created momentum for rethinking operating models and embedding AI strategically. The biggest competitive gaps over the next years will likely form around AI adoption: institutions that proactively reorganize workflows and treat AI as the engine of productivity and growth — not merely a technology tool — can achieve valuation advantages over reactive peers.
The central question is no longer whether banks will use AI, but whether they can shift to AI‑centric operating models where processes, decision‑making, risk management and customer journeys are designed around intelligent systems. Those that improve productivity, expand growth options, manage capital discipline and prepare for digital assets, non‑bank competitors and autonomous AI agents will be best positioned to win. The opportunity exists — but the cost of falling behind is rising fast.
Key data at a glance
- 2025 average financial sector TSR: 30.2%.
- Planned AI spend (2026): ~2% of revenues (BCG survey).
- Potential U.S. capital‑reform unlocked capacity: up to ~$2,600 billion (BCG mention).
- Sector market capitalization increase since end‑2022: ~ $5,500 billion.
- Sector P/E discount vs. broader market: ~ 40%.
(Source: Boston Consulting Group — "The Future of Finance – Time to Shift Gears?" analysis.)
