Venture capitalists are very excited about rollups, and it’s easy to see why. On paper, the model is seductive: buy up established services firms, retrofit them with AI, and watch margins grow. The opportunities to write large checks are plentiful. We think this story, however, ignores some hard truths: buying well is difficult, “AI-ifying” is patchy and slow, and rollup cultural dynamics can be painful. The unique opportunity, we believe, isn’t in stitching together yesterday’s services giants – it’s in building tomorrow’s from scratch.

VCs are infatuated with rollups – for good reason

AI-driven VC rollups have become a very hot play. The likes of Thrive Capital, Khosla Ventures, Elad Gil, and more have all called their shots in this space in the past year. The mechanics are simple: buy mature services businesses – accounting, legal, wealth management, property lettings, IT, even healthcare – and refit them with AI to supercharge scalability and margins. The bets are big, even by mega-fund standards: Crete, backed by Thrive and Bessemer, is reportedly eyeing $500M for accounting firm acquisitions, while General Catalyst has allegedly set aside $750M+ for similar deals. And these are just the public announcements – plenty more rollups are quietly in motion.

The thesis is fairly straightforward, underpinned by three core premises:

• (i) AI-based software can transform these services businesses – enabling workflow automation across the board, from front-office to back-office tasks. This primarily reduces costs, but also supports top-line through volume.

• (ii) Software distribution in these industries is particularly challenging – with highly fragmented customer bases (typically without dedicated software budgets) and heterogeneous legacy tech stacks.

• (iii) Service economics could grow in relative attractiveness to software going forward – while this is not a necessary underpinning of the model, many involved in these roll-up strategies worry that pure software margins may erode going forward with AI (through a combination of pure token COGS and increased competition), whereas service margins are likely to grow with AI-driven automation.

As a result, proponents argue, the logical strategy is to buy these businesses for their distribution and pricing power, then leverage AI to reduce costs and increase volume. The result: where previously there were many small or medium-sized firms with mediocre margins, you end up with a very large services firm with highly attractive margins.

The pitfall: buying companies is hard

While, to varying degrees, we agree with the three premises above, we fundamentally disagree with the conclusion that VC rollups are the best way to capitalize on the opportunities they create. Our skepticism boils down to a few factors.

First, buying well is hard. There is a very competitive private equity buyout market for services businesses. PE firms are very skilled at buying – they see all opportunities, broadly know how to price them, and have a big appetite. While some might cling to the hope that PE firms are mispricing the AI opportunity in these assets, it’s hard to believe that they are completely asleep at the wheel. It’s not easy to get bargains, and it’s almost impossible to do so at scale. These assets are highly combed over. The best one can hope to do is “as well” as everyone else.

Second, “AI-ifying” is easier said than done. These businesses come with legacy tech stacks, ossified processes, incongruent data models, personnel who are set in their ways, and so on. Replacing excess people and legacy software with AI is highly complex – buyers can’t just flip a switch and accomplish it wholesale. And the problem doesn’t go away with scale – integrating AI into acquired business N+1 is not much easier than doing so for business N. Their data models, tech stacks, processes, cultures, etc. are likely to be completely different (see premise ii above).

Lastly, a culture of “shedding” has its own costs. This is a fuzzier point but an important one. Venture firms, software engineers, and tech operators – the archetypes tasked with bringing these rollups to life – are used to environments of growth and abundance. Successful tech companies hire continuously and expand their long-term product ambitions over time. The services rollup environment is very different. Reducing costs is mandatory, so continuous personnel cuts are often necessary, while operational nuances and repetition are the key to scale. It remains to be seen how this cultural disconnect will impact the attraction/retention of the talent required to execute.

A better path

The ideal way to capture the opportunities created by AI in services is to build new services businesses from the ground up. This way, there’s no need to worry about competing with PE firms for acquisitions, and there’s no tech or cultural debt to pay down. Instead, operators get to rearchitect the business from scratch with AI at its core, and reap the economic reward the technology creates.

The big caveat here is related to distribution. While buying businesses is complex, it does a perfect job of getting distribution in the form of paying customers. There are certainly some areas in services where this constraint is almost impossible to overcome. Investment banks, for example, have very strong brands. A startup investment bank will have a very tough time taking clients from an established one. In other areas, however, brand is much less powerful, and clients actually care a lot more about the experience or person they are dealing with, and less about the logo of the firm. This is where a startup can succeed – by attracting top level talent that has established customer relationships, the desire to create more by leveraging technology, and a cultural buy-in with the vision.

The opportunity for AI in services businesses is massive. There are tons of processes that can be made more efficient, legacy software tools that can be upgraded, and additional services that can be offered to clients. While AI rollups are certainly a viable strategy, and can absorb large amounts of capital – we think the firms that will capture the most compelling end of the curve here, however, will be built from the ground up with AI and technology at their core, not stitched together through acquisitions and patch work technology.

We’ve been investing behind this thesis at Tower Research Ventures – check out Arvor Insurance. If you are interested in this and/or working on something in this space, please reach out to us at ventures@tower-research.com or pdecio@tower-research.com.

The views expressed herein are solely the views of the author(s), are as of the date they were originally posted, and are not necessarily the views of Tower Research Ventures LLC, or any of its affiliates. They are not intended to provide, and should not be relied upon for, investment advice, nor is any information herein any offer to buy or sell any security or intended as the basis for the purchase or sale of any investment. The information herein has not been and will not be updated or otherwise revised to reflect information that subsequently becomes available, or circumstances existing or changes occurring after the date of preparation. Certain information contained herein is based on published and unpublished sources. The information has not been independently verified by TRV or its representatives, and the accuracy or completeness of such information is not guaranteed. Your linking to or use of any third-party websites is at your own risk. Tower Research Ventures disclaims any responsibility for the products or services offered or the information contained on any third-party websites.