Addepar crosses $10T as asset servicing's scale race heats up

The $10T milestone

Addepar announced on October 21 that it has reached $10 trillion in assets on its platform, serving a client base of more than 1,500 firms. Alternatives represent over 40% of assets on the platform. Nearly half of Addepar's workforce is based outside the United States, and the company operates 10 offices in 9 countries. This is a company announcement, so treat the figures as self-reported, but the direction is unambiguous: scale in wealth-tech data is concentrating. m:…1949

The milestone was framed around AI. Addepar launched Addison, an AI partner for investment professionals, in 2026, and Addepar MCP connects portfolio data to AI tools such as Claude. That follows the 2025 launch of Alts Data Management and Private Fund Benchmarks. The pitch: the platform is not just a book of record but the data layer that AI agents will reason over. m:…1949

Northern Trust's quarter

Northern Trust's third quarter, reported the same day, shows what scale looks like on the custody side. Net income was $642.9 million, diluted EPS rose 40% year over year excluding notable items, total revenues grew 11%, and net interest income grew 14% on higher deposit balances and a favorable deposit mix. The firm achieved over 600 basis points of operating leverage — revenue growing well ahead of expense. m:…1446

Assets under custody and administration reached $19.9 trillion, up 9% year over year, and assets under management grew 11% to $1.96 trillion. Asset servicing pretax margin hit 30.3% for the quarter. And the strategic signal: Northern Trust plans to convert six Northern Funds, representing $33 billion in assets, into ETFs in Q1 2027 — a structural shift toward the wrapper investors are actually buying. The firm also noted that AI agents save relationship managers 2 to 3 hours on every client memo, a concrete productivity figure on the earnings call. m:…1446

The build-or-buy call

The third piece of the puzzle is FundGuard's analysis of the AI infrastructure choice facing asset managers. Seema Amble outlined three paths for software buyers integrating AI agents: keep incumbent systems and add agents, build in-house, or buy AI-native platforms. Asset managers are moving beyond AI pilots to integrating agents into investment operations, and AI agent effectiveness depends on the quality, timeliness, and accessibility of data — which is precisely what the Addepar milestone is about. m:…8381

The numbers make the urgency clear. BCG estimates agentic workflows could raise investment operations capacity by 55 to 65% and cut operational costs by around 40%. Yet 57% of asset managers expect digital transformation to take more than five years, regulators require lineage and the ability to reconstruct state at a past point in time, and AI lowers the cost of recreating the first 80% of a system of record — meaning the defensible moat is the data, the lineage, and the last 20%. Notably, 44% of managers see collaboration with third parties as a way to build scale and share risk, which points toward buying rather than building. m:…8381

My read

My read: the October 21 announcements describe the same phenomenon from two sides. Addepar's $10T and Northern Trust's $19.9T in custody assets are both scale stories, and both companies are explicitly pitching AI as the return on that scale — agents that reason over a proprietary data corpus competitors cannot replicate cheaply. m:…1949 m:…1446

For professionals weighing the three paths, the question FundGuard's piece implies is whether your data estate is agent-ready. Batch platforms that settle data overnight limit agent reasoning to stale positions, per the analysis. The firms that win the next phase are the ones whose records are current during the day and reconstructible on demand — which is a data problem before it is a model problem. m:…8381