Les marges des gestionnaires d'actifs se contractent face à la hausse des dépenses d'infrastructure, selon PwC

PwC reports that asset managers’ profit per dollar of assets under management has fallen 19% since 2018 and is projected to decline a further 9.1% by 2030, as infrastructure costs rise and fee compression limits investment for 67% of respondents. Legacy systems and fragmented data are cited as barriers to scaling AI, while 29% of managers spend more than 70% of their technology budget maintaining existing systems. The survey finds 44% favor hybrid sourcing models combining in-house capabilities, purchases and partnerships.

Analysis

Infrastructure economics, not AI capability, is the investable issue. The survey points to a bargaining-power shift: asset managers are unlikely to fund a broad technology rebuild while fee pressure persists, so spending should favor vendors that let clients retire or consolidate legacy systems, not simply add AI layers. That creates a potential squeeze for undifferentiated data and workflow providers if buyers renegotiate bundles; it is not evidence that all market-data businesses are impaired, since proprietary data and switching costs may defend pricing.

Second-order winners: hyperscalers can capture incremental workloads, but may also compress specialist infrastructure providers’ economics by bundling compute, storage and AI services. Modular specialists and implementation partners may win near term where integration, data cleanup and migration—not model sophistication—are the bottleneck. Consolidation among asset managers could accelerate as firms unable to amortize fixed technology costs seek scale or outsource more.

Timing and risk: days-to-weeks, this survey alone is a weak earnings catalyst. Over 1–3 months, watch asset-manager guidance and vendor renewals for actual budget shifts, contract repricing or delayed projects. Over 6–18 months, successful legacy rationalization could restore operating leverage and enable AI deployment; failed migrations, cyber incidents or regulatory constraints could instead raise costs. PwC survey responses are directional, not audited spending data.

Contrarian: the market may over-credit hyperscalers and underweight migration friction. Cloud adoption can shift costs rather than reduce them, while data fragmentation may make near-term AI returns disappoint. With no company-level exposure or valuation data supplied, there is no high-conviction outright trade.

Key Decisions for Investors

• No immediate sector-wide position from this survey. Build a watchlist of hyperscalers and infrastructure specialists positioned for migration, but require evidence of signed contracts or revised revenue guidance before adding exposure.
• Consider a conditional relative-value trade: long diversified cloud exposure versus short undifferentiated legacy financial-data/workflow exposure only if renewal commentary confirms pricing concessions or customer churn. Avoid treating proprietary-data franchises as equivalent without company-level evidence.
• Track asset-manager earnings over the next 1–3 months for technology spend, run-rate cost savings, and outsourcing disclosures; these are more actionable than survey sentiment. Falsify the margin-recovery thesis if infrastructure costs keep rising without lower operating expenses or improved investment capacity.
• Verify vendor-specific revenue exposure, contract renewal dates, cloud migration costs, and asset-manager budget data before sizing positions. A shift toward hybrid sourcing does not establish that any one vendor captures the spending.