Venture capital is tapping into the artificial intelligence transformation that enterprise software has not been able to deliver. Rather than selling AI tools to companies, startups are acquiring legacy companies outright and rebuilding them from within around AI.
This bet will put VCs on the offensive and leave traditional private equity, which spent the last cycle buying enterprise software at the highest prices, on the defensive.
In Silicon Valley, this strategy is known as an AI rollup. In the past six months, the company has entered the public market twice. General Catalyst and Trian took Janus Henderson (JHG) private in December for $7.6 billion, and Long Lake Management agreed to take American Express Global Business Travel (GBTG) private in May for $6.3 billion for a 65% premium.
General Cataly Managing Director Madhu Nambri calls this “software-as-a-service.” This is the idea of Software-as-a-Service (SaaS), which increases profitability for software companies because growth does not require increased costs. AI Rollup applies the same logic to service businesses.
The venture has been implementing this strategy since 2023, primarily within the private market. General Catalyst, which supports Long Lake alongside Alpha Wave, co-developed about a dozen of these roll-up vehicles.
Joshua Kushner’s Thrive Capital runs Thrive Holdings with the same model and more than $1 billion in capital. We recently leveraged that funding to support an AI rollup for a regional accounting firm. Lightspeed and Andreessen Horowitz are also on the list, albeit prematurely. Targets have common characteristics. They are in industries that have been slow to adopt software, such as healthcare, accounting, insurance, customer service, property management, and construction.
It also changes who can make these transactions. Traditional private equity is built around the financial engineering of taking fixed cash flows, leveraging them, and squeezing profits. AI Rollup is built around growth. AI expands customer-facing teams, and reinvested cash funds more acquisitions. This is a twist on the venture model, essentially applying a growth mindset to an existing company. Long Lake plans to own the company permanently, a strategy the company has adopted. Berkshire Hathaway.
Long Lake is the clearest example of what a bridge to effective AI adoption looks like. The three-year-old company has acquired more than 30 businesses ranging from HOA management, construction, and now corporate travel. It runs a proprietary AI platform called Nexus, which is tailored to the specific workflows of each industry that Frontier Labs did not initially target.
Long Lake CEO Alex Taubman says the company’s internal evaluations show Nexus performs five times better than general-purpose models such as Claude and ChatGPT. Beyond technology, the bet is that by owning a company and embedding engineers in-house for years, change will become permanent. Most of Long Lake’s engineers come from Ramp; Palantirengineers work in the field with customers for months at a time.
Traditional private equity made the opposite bet. The company spent the early 2020s buying enterprise software at the highest multiples, based on the argument that SaaS recurring revenue was the most defensible cash flow in the business. Transactions include Vista’s acquisition of Citrix, Thoma Bravo’s acquisition of Anaplan and Coupa, and Silver Lake’s acquisition of Qualtrics. Three years later, these companies are the ones most exposed to AI disruption.
The recently announced Anthropic and black stoneHellman & Friedman, and goldman sachs — and a parallel venture with OpenAI, backed by Apollo and General Atlantic — is the answer. They are incorporating the Frontier model into their portfolios that are already on their books. But it looks like a consultant’s attempt at a deployment problem. Someone’s AI introduced into someone’s company by someone who neither owns.
Two problems can arise with the VC model. The first is returns. Industrial companies have historically delivered returns of 100% to 200% on long-term holdings, but not the 10x that math venture funds often promise. Pension funds and endowments that write checks for venture exposure may end up looking more like private equity. The second is execution. Vista and Thoma Bravo spent decades building an operating team to run the company they took private. Venture companies write checks to startups. Mr. Taubman defends himself. “Three years in AI actually equals 30 years before AI.”
The next privatization cycle has already begun, but it’s not in the software. Below that are boring non-tech companies.
