In 1988, eight people squeezed into a single room in New York to start an investment firm around an unusual premise: data and technology could give investors a clearer view of risk.
That firm became BlackRock. Its Aladdin system grew from an internal risk tool into a platform connecting portfolio construction, trading, operations and accounting. BlackRock ended 2025 with $14 trillion under management after attracting $698 billion of net inflows in a single year.
The revealing number, though, is eight.
Today the average SEC-registered investment adviser focused on individual clients also employs eight people and manages $424 million. Small advisers are not disappearing. Their number reached a record 16,544 in 2025, and more than two-thirds manage less than $1 billion.
The giant became vastly larger, while the boutique became easier to build.
The pressure is landing between them. DeVoe & Company counted 322 wealth management transactions in 2025, a record, up from 272 the year before. The more telling detail is who was buying. The industry had 18% more sellers but 19% fewer buyers than the prior year, and first-time buyers accounted for just 8% of deals, the lowest share on record. Acquisitions are concentrating into a shrinking set of private equity-backed platforms.
This is what a technology barbell looks like. The largest platforms spread their data, expertise and infrastructure across more volume. Small specialists rent capabilities they could never afford to build. Firms in the middle carry enough overhead to need scale but lack enough scale to fund a differentiated platform.
Artificial intelligence is likely to accelerate that pattern across the service economy.
AI’s barbell
The debate about AI often asks whether it will concentrate power or democratize it. Both can be true, because scale never made coordination easy. It made coordination affordable.
Only a large company could spread the fixed costs of a revenue management team, a 24-hour service center and an enterprise technology stack across enough volume, and coordinating it all was a tax that made large organizations slow. AI cuts that drag for the largest companies and removes the affordability threshold for the smallest, which can now use capabilities they never could afford to hire. What it does not erase is purchasing power, insurance pooling, capital access, distribution and institutional credibility, which remain real advantages of size.
That leaves the middle with the worst of the arithmetic: the coordination burden of an institution without the volume to spread a real platform across it. Those firms will build one, join one, or become specialized enough that customers pay for the difference.
Hotels show what this looks like in an industry few people associate with AI, because every improvement there shows up in a nightly rate, a labor hour, a guest review or an owner’s cash flow. A disclosure before I go further: my company builds and operates exactly this kind of system, so I have a stake in the answer. The evidence should matter more than the source.
Hotels are next
Walk into a hotel conference and you can feel the split. Brand executives discuss growing fee revenue. Owners talk about insurance, labor, property taxes, renovation mandates and loans coming due. Everyone flew to the same conference hotel. It does not always feel like the same industry.
Hotel brands increasingly operate asset-light, fee-based businesses. Owners carry the real estate, debt, capital expenditures and interest-rate risk. Across the 2,216 hotels in CBRE’s latest annual Trends survey, revenue grew 2.6% in 2025 while total expenses grew 3.1%. Gross operating profit margins declined from 35.1% to 34.8%. Insurance costs remained roughly twice their 2019 level.
This should worry brands and managers as much as owners. Every future fee stream begins with someone deciding to invest millions of dollars to build, acquire, convert or renovate a hotel. When the expected return no longer justifies the risk, the rooms never open.
A 50-room hotel still has to forecast demand, set prices, distribute rooms, answer guests around the clock, schedule housekeeping, coordinate maintenance, reconcile payments and interpret its results. A global chain spreads those capabilities across thousands of properties. An independent owner cannot.
AI can change that equation, but not if hotels treat it as one more subscription. They already run separate systems for reservations, pricing, distribution, guest messaging, housekeeping, maintenance, accounting and reputation management. Adding an AI assistant to each may improve individual tasks. It does not coordinate information and action across the business.
An AI-native operating model starts with a connected operating layer. Pricing, distribution, guest communication, property operations and accounting share data. Routine decisions happen continuously. People focus on exceptions, judgment and hospitality. What the system learns at one property becomes the starting point at the next.
The buildings did not change
In January, Kasa acquired Mint House, an apartment-hotel brand that operated nearly 1,000 units across 22 properties. We integrated the majority of the portfolio onto our platform and closed a small number of locations that did not fit our model or quality bar. Across the hotels that we transitioned, direct bookings increased more than sixfold as a share of business, reducing third-party commissions. Average Google review scores rose from 4.24 to 4.56. Same-store revenue per available room increased 19.2% year over year.
The buildings did not change. The operating layer did.
That layer weighs booking pace, competitor rates, local events and channel costs throughout the day. It handles routine guest questions and knows that extra towels can wait while a report of gas odor cannot. It connects negative reviews to the maintenance condition that caused them, and gives owners an explanation of what changed in revenue, labor and cash flow, plus the few decisions requiring attention. The point is to remove repetitive coordination so people’s time goes to judgment, creativity and genuine care. Guests don’t remember a flawless software integration. They remember the person who solved their problem and made them feel special.
The result could be the same barbell visible in asset management. The largest hotel companies and management platforms will use AI to make their infrastructure more powerful. Independent hotels and small brands will gain access to capabilities once reserved for global chains while preserving their architecture, quirks and local service.
My suspicion is that the greatest pressure will fall on regional managers and smaller brands in the middle, perhaps those operating 15 to 50 hotels. They are large enough to carry corporate overhead but too small to spread the cost of a differentiated technology platform across enough rooms. That squeeze has already started without AI’s help: in a single recent year, Stonebridge acquired Real Hospitality to push its portfolio past 160 hotels, Nautic Partners bought Davidson Hospitality, Griffin merged into Meyer Jabara, and PM Hotel Group absorbed Sightline Hospitality and its 22 properties. Sightline sat squarely in the middle band.
This pattern will travel beyond hospitality. Property management, healthcare services, accounting, insurance and logistics all contain businesses built around fragmented software and repetitive coordination.
The next generation of important AI companies may look less like software vendors and more like operators. They will combine industry expertise, people and technology into one operating model, use it to run the service directly, and make its capabilities available to businesses of every size.
AI will strengthen the biggest platforms and make the smallest specialists more formidable. The businesses caught between them will face the hardest strategic choice.
Roman Pedan is the founder and CEO of Kasa, an AI-native hotel management company and operating platform, and an Adjunct Assistant Professor at Columbia University’s Graduate School of Architecture, Planning and Preservation, where he co-teaches Leveraging Data and AI for Real Estate Development. Kasa manages more than 80 independent and branded hospitality properties using proprietary technology to automate hotel operations.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
