A new buzzword for many VCs today is “rollups.”
This year, a slew of VC firms have turned to the PE playbook of buying and merging multiple similar small companies in the hope of building industry leaders. But unlike its big brother, VC’s secret sauce is technology, specifically AI.
Put simply, the strategy is to buy companies in a fragmented sector, embed proprietary AI technology to boost margins and scale faster than incumbents. In essence, instead of selling tech to companies, VC-backed firms are just buying them.
VCs believe so thoroughly in the power of AI that several are raising capital just to pursue this strategy.
General Catalyst committed $1.5 billion from its recent $8 billion raise to rollups, while Thrive Capital is raising a $1 billion-plus permanent capital vehicle for this purpose. And it’s not just the big guns getting into the PE game; Finnish VC firm OpenOcean is said to have raised €100 million (about $114 million) for rollups.
While the rollup model has captured investors’ imagination, questions remain about whether this strategy can deliver the outsized VC returns that they have come to expect.
Why now?
Rollups are not new; they’ve been widely used in PE since the 1980s. They briefly appeared in VC during the 2021 boom era with Amazon aggregators, which ultimately played out poorly for investors once the downturn hit.
The timing of their arrival in VC today is no accident. Several forces have converged to make the model appealing to the asset class, perhaps most importantly, being AI’s increasing maturity.
“We’re at a unique moment in time when there’s AI out there that can theoretically grow margins,” Euclid Ventures general partner Nic Poulos said.
According to a McKinsey report, AI could help companies unlock around $4.4 trillion in long-term productivity growth. Furthermore, a Salesforce survey found that 91% of small and medium-sized businesses using AI said the technology boosted their revenue.
Generative AI can automate a multitude of white-collar work, from bookkeeping to legal document reviews. The only problem is a lack of adoption.
I think there’s a lot of hype, and I can see a lot of VCs getting hurt
Many industries that would be targets for AI rollups are fragmented, populated by thousands of small operations that are slow to adapt to new technology. One such industry is property management, which was the target of Berlin-based Buena in its GV-led $58 million raising last month.
“The traditional average property management company is typically run by someone close to retiring and wants someone [else] to take over the business,” Buena CEO Din Bisevac said. “It’s really hard to sell technology to them, but the efficiency gains are huge. When we started, the idea wasn’t to do rollups, but we saw that we were more efficient and profitable than traditional incumbents through our technology.”
Beyond the technology, another compelling reason for large-scale VCs to pursue the model is simply that there isn’t enough traditional venture for them to do. Returning a $10 billion or $20 billion fund is a mighty feat, and investors with huge AUMs must branch out.
“As firms grow larger or they seek to grow past a certain point, you generally have to launch new products. [AI rollups] allow VCs to manage more capital, and it’s a natural progression,” Bisevac said.
Structural challenges
For all its appeal, the AI rollup model is far from straightforward, and investors question how the strategy makes sense for a typical VC.
“I don’t get it as a VC model,” Foundamental investor Fabio Bronzin said. “I think there’s a lot of hype, and I can see a lot of VCs getting hurt.”
To make rollups work, you need a lot of capital to fund acquisitions and technology development with cash, at least in the short term, presumably coming from VCs. According to Bronzin, “the maths doesn’t make sense.”
Bronzin gives a hypothetical example of a VC investing $20 million for a 20% stake in a startup at a $100 million valuation. If the entirety of the capital is spent on acquisitions, the VC owns 20% of all the businesses and yet supplies 100% of the cash. They are effectively paying 5x the book value of the companies being acquired, and even if the businesses together grew 5x, the VC’s return would only be 1x.
A natural mismatch between capital deployed and ownership for investors doesn’t exist in PE, where capital providers own a controlling stake and can capture most of the value uplift from acquisitions.
Related: GPs seek to add value with AI-powered portfolios
Venture debt has proven to be a popular alternative to equity in funding acquisitions, but that also has challenges. High interest payments eat into EBITDA, especially in businesses still absorbing R&D costs or that haven’t achieved scale. It also increases fragility, which can limit flexibility and growth.
Massive firms such as General Catalyst could, in theory, act more like PE players, but the issue comes for smaller investors whose profiles don’t align with the model.
PE firms are also unlikely to sit on the sidelines while VC swoops in. According to Euclid’s Poulos, any initial AI-driven edge that a VC-backed platform enjoys might be fleeting once larger, more experienced consolidators decide to compete for the same acquisition targets. This is particularly true as PE firms are already looking into using AI to improve their portfolio companies.
Even if VC-backed rollups can secure deals ahead of PE, they face the operational burden of running two different businesses simultaneously. Founders must build defensible AI while operating a service-heavy business spread across multiple acquisitions. Integration is rarely straightforward, and layering on tech development only adds to the risk.
According to Foundamental’s Bronzin, even if everything works, there’s still no guarantee of venture-scale returns.
“If you’re the VC underwriting an AI rollup, you’re underwriting on the thesis that the public market will value you on a tech multiple rather than a service multiple,” he said. “This is unproven and unlikely. The reason is that the tech you’re developing is only for yourself, your TAM for that specific tech is only the companies you acquire. You’ll probably get a premium, but it won’t be tech level.”
Fundamentals first
While there are several reasons to doubt the model, founders like Buena’s Bisevac argue that the strategy can deliver venture-scale outcomes when executed with the right fundamentals. He says the key is to treat the method as a product-first play, and not a thin layer of tech wrapped around traditional services.
“It boils down to the economics, and one has to decide, is it PE or VC?” he said. “It’s a naive bet to think that you can disrupt an industry by focusing on operations and not technology. If you’re not going to focus on innovation, then you should go the PE route. If you can truly innovate and move margins from 3% to 60%, and show venture-like growth, the market is going to value that accordingly.”
But even Bisevac shares concerns that the VCs chasing the hype will end up funding companies that execute roll-ups poorly, leading to failures and destroying the perception of the model.
“Whether it’s a lasting trend really depends on how companies like ours play out,” he said. “I can imagine AI rollups being part of portfolio construction, but if we don’t execute well, it could be over very soon.”
Featured image by Adriana Hansen/PitchBook News
Learn about our editorial standards.
