Who's Deploying Billions to AI Roll-Ups—and How Their Strategies Differ
Part 2 of The Founder's Guide to AI-Enabled Roll-Ups
Three years ago, the idea that venture capital firms would spend billions acquiring accounting practices, call centres, and IT service providers would have seemed absurd. VCs back high-growth software, not mature services businesses with single-digit margins.
But margins can change. And the firms betting most aggressively on AI roll-ups have track records that warrant taking them seriously.
This chapter profiles who's deploying capital, how much, and why their strategies aren't interchangeable. If you're evaluating this space as an investor, you need to understand what sets the players apart. If you're a founder who might sell to one of these platforms, you need to know what each buyer actually offers beyond the headline price.
What's Inside
- $4.5B+ allocated to AI-enabled roll-ups: General Catalyst ($1.5B), Thrive Holdings ($3B+ with OpenAI equity), plus Bessemer, Lightspeed, and 8VC
- Strategies differ significantly: General Catalyst builds AI first then acquires, Thrive embeds OpenAI engineers directly into portfolio companies, traditional PE enters later with off-the-shelf tools
- Capital timeline changes everything: Permanent/evergreen vehicles (GC, Thrive) mean no forced exits, while PE funds have 3-5 year clocks that pressure returns and deal structures
- Founders have more negotiating power: Multiple buyers pursuing similar targets means better terms than traditional M&A — Crescendo's $500M valuation validated the model and attracted more capital
- When evaluating buyers: Ask about delivered AI results (not promises), capital structure, brand preservation track record, and talk to founders who've already sold to them
Capital on the Table
The commitments are substantial enough to reshape industries:
- General Catalyst: $1.5B allocated from its $8B fundraise to the "Creation Strategy"
- Thrive Holdings: permanent capital vehicle launched April 2025; it says it has raised more than $3B since, including over $2B at a $12B valuation in August 2026
- Bessemer Venture Partners: Co-invested in Crete (2024) and in Carr, Riggs & Ingram (CRI), a top-25 US accounting firm
- Lightspeed: Roll-up plays in engineering services, healthcare
- 8VC: Sequence Holdings and Arcos
- Slow Ventures: Teamshares, now listed on Nasdaq
Total identified: $4.5B+ to this specific strategy. Adjacent approaches add billions more.
Why is venture capital, and not private equity, driving this?
Traditional PE improves what exists. Cut costs, centralise operations, sell within 3-5 years. The value creation comes from efficiency and multiple arbitrage. Understanding PE as an investment strategy shows why the timeline matters.
AI roll-ups require building something new. Develop AI tools, integrate them into operations, create feedback loops that improve automation over time. That's technology development, which is VC territory. The result is a hybrid: a PE acquisition strategy combined with VC-backed tech development. The firms deploying capital understand both worlds.
General Catalyst: Build First, Buy Second
General Catalyst pioneered this model through its Creation Strategy, which inverts the typical relationship between technology companies and customers.
Instead of building software and selling it to service businesses, GC incubates AI-native companies that then acquire those businesses outright. The acquired companies become customers, distribution channels, and data sources for the AI platform—while generating cash flow to fund more acquisitions.
Marc Bhargava, who leads GC's efforts here, frames the opportunity simply: "Services globally is $16 trillion in revenue a year. In comparison, software is only $1 trillion globally."
Fifteen times larger. That's the prize if AI can bring software-like margins to services.
Portfolio
Crescendo (call centres) is the most-cited example of the thesis. General Catalyst says it is taking gross margins to 60-65% or more as its platform rolls out; no measured margin has been published. Its chief executive puts traditional outsourcers' EBITDA margins at about 15%, a different measure. It was valued at $500M in its October 2024 Series C and says it passed $100M of annual recurring revenue in April 2026. When it acquired PartnerHero (200+ customers), it absorbed a company that should have been an acquirer in its own right.
Long Lake (HOA/multi-vertical) reached $100M EBITDA in under two years, according to General Catalyst's chief executive, a pace that makes traditional PE roll-ups look glacial.
- Raised: about $670M of equity by mid-2024; PitchBook now lists $3.78B in total, including $2.5B of acquisition debt
- Acquisitions: close to 40, by its chief executive's count (September 2026)
- Productivity gains: 25-30%
In September 2026, Long Lake completed its $6.3 billion acquisition of American Express Global Business Travel, its largest deal yet, signaling that the model extends well beyond HOA management.
Titan MSP (IT services) demonstrated it could automate 38% of typical MSP tasks through pilot programmes. GC led a $74M investment in August 2025, funding Titan's acquisition of RFA (now more than 700 financial-services clients).
Eudia (legal) went straight for enterprise clients. Enterprise customers include Cargill, Del Monte, and Stripe. Fixed-fee legal services powered by AI, rather than hourly billing. Recently acquired Johnson Hana, an alternative legal services provider.
Dwelly (UK property management) offers the clearest before/after comparison. General Catalyst says that wherever its AI is fully deployed, EBITDA margins have doubled and repair wait times have fallen by 40%. Dwelly has kept buying: it says it made 13 acquisitions in 2026 by late October, some of them lettings portfolios, taking it past 16,000 properties. The question is whether the margins scale with it.
Accrual (accounting) raised $75M in February 2026 to sell AI software to accounting firms. It buys technology, not firms: its one deal so far is for Puzzle's accounting-firm product.
What Makes GC Different
Three things stand out.
First, they build before they buy. The AI capabilities exist and are proven before acquisition capital gets deployed. That's the opposite of PE firms, which acquire first and figure out the technology later.
Second, permanent capital. No fund lifecycle forcing exits. GC can wait a decade for the transformation to compound. A PE firm with five-year funds doesn't have that luxury.
Third, founder retention as a strategy, not courtesy. Owners roll equity, stay involved, keep their brands. The portfolio companies benefit from the founders' knowledge, while GC provides technology and capital.
Free self-test · One page
Sell, invest or build?
Knowing the buyers is half of it. The other half is which seat you're in. The one-page self-test runs three readiness checks, then shows which of four moves fits you now: sell to a platform, invest in one, build your own, or wait. Free with the Complete Edition of this guide.
Get the self-test →Thrive Capital: The OpenAI Partnership
Thrive Capital is known for concentrated bets held for years—early investments in Stripe, Instagram, Spotify. More recently, they've become one of OpenAI's largest investors, first backing the company at a $27B valuation in 2023, then leading a $6.6B round at a $157B valuation.
That relationship just got much deeper.
In December 2025, OpenAI announced it was taking an ownership stake in Thrive Holdings—not as a passive investor, but as an embedded partner.
What This Actually Means
The deal goes beyond typical vendor relationships. OpenAI will embed research, product, and engineering teams within Thrive Holdings portfolio companies. These teams build custom AI tools for specific industries while gaining access to real-world data that helps train OpenAI's models.
OpenAI receives equity, not cash. If Thrive Holdings' companies succeed, OpenAI's stake appreciates. The structure aligns incentives: OpenAI is motivated to make the AI transformation work because it profits directly from the results.
Some call it a "circular deal"—Thrive invests in OpenAI, OpenAI takes equity in Thrive Holdings, everyone profits from each other's success.
The optimistic read: aligned incentives for long-term value creation.
The sceptical read: a closed loop where it's hard to tell whether success comes from market traction or from advantages that only work with direct OpenAI support.
I lean towards cautious optimism. But founders considering selling to Thrive should understand they're betting on this relationship staying productive. And they should ask: what happens to my AI capabilities if Thrive and OpenAI ever part ways?
Portfolio
Current, a Thrive-backed accounting network rebranded from Crete Professionals Alliance, is Thrive's flagship. The growth metrics are impressive:
- Founded: 2023
- Revenue: $500M+ (as of June 2026)
- Employees: 2,000+ (as of June 2026)
- Partnerships: more than 30 partner firms (company, September 2026)
- New acquisition budget: $500M over 24 months
That $500M budget means roughly 20-30 more accounting firms at typical multiples. Enough to reshape the competitive landscape for any regional CPA wondering whether they're an acquirer or a target.
The AI results are concrete. Bennie Lewis, President of Assurance Dimensions (a Crete-owned firm in Tampa), reported AI tools saved his team "hundreds of hours every month" in audit testing alone.
Shield Technology Partners (MSPs) launched in June 2025 with $100M+ funding from Thrive Holdings and ZBS Partners. A second $100M followed in February 2026, and Shield now runs its own Forge automation platform.
Current state:
- Acquisitions: 20 MSPs (including ClearFuze, IronOrbit, Delval, OneNet Global, NetAscendant, BCS365, SK Tech Group, Nerds Support)
- Targeted doubling by Q1 2026, since exceeded: 20 MSPs by September 2026
- CEO: Jim Siders, former Palantir CIO (started as helpdesk engineer, ended as CIO)
- Internal products: Sentinel and Spectre (auto-resolve repetitive tickets)
Siders' hire signals serious intent. Palantir's approach to enterprise AI deployment is widely respected. Bringing that operational playbook to MSPs suggests Shield isn't just rolling up businesses—they're building a fundamentally different kind of IT services company.
Everything here is free. Subscribing just tells me the content is useful — and helps me decide what to write next.
Other Players Worth Watching
Bessemer Venture Partners co-invested in Crete and in CRI (Carr, Riggs & Ingram, a top-25 US accounting firm), alongside Centerbridge Partners. Brian Feinstein, a partner focused on enterprise software, offers the most honest framing I've seen: "There's not some magic AI accounting product today that automates the audit function or the tax function."
His positioning is instructive: "In a base case, we expect this to be a really good private-equity deal. The tech automation and the AI tailwind is a source of upside that helps us get to a home-run case."
That's the right mental model. PE economics as the floor, AI transformation as the upside.
Lightspeed raised $9B in December 2025, the largest raise in their 25-year history. Partners have made roll-up plays in engineering services and healthcare, though with less public detail than GC or Thrive.
8VC takes a sector-focused approach. Sequence Holdings, whose founders came from Scale AI, Cognition and Lone Pine according to Newcomer, describes itself as a holding company that buys established service businesses and rebuilds them around AI. Arcos reinvents transactional law as a technology-first platform. Joe Lonsdale's Palantir background shows—the emphasis is on understanding deep workflow structure before automating.
Slow Ventures backs roll-up startups with lower upfront capital. Partner Yoni Rechtman remains sceptical of accounting specifically, citing operational complexity that exceeds what most investors can handle. Worth noting when everyone else is optimistic.
How the Strategies Compare
Rather than abstract comparison, here's what actually differs:
| Dimension | General Catalyst | Thrive Holdings | Traditional PE |
|---|---|---|---|
| AI approach | Build in-house first | OpenAI partnership | Buy tools off-shelf |
| Capital timeline | Permanent (decades) | Evergreen (decades) | 5-7 year funds |
| Founder retention | Core strategy | Core strategy | Variable |
| Brand preservation | Yes | Yes | Often consolidate |
| EBITDA target | 30-40% | 30-40% | 15-25% |
The difference in capital timelines matters more than it might seem. GC and Thrive can wait for AI transformation to compound. A PE firm racing a fund clock may sell half-transformed because capital needs to return. I've seen this happen—"platform" businesses sold before the platform thesis played out.
The difference in AI approaches matters for the depth of transformation. GC develops capabilities internally, controls the roadmap. Thrive gets direct OpenAI access but depends on that relationship. PE firms buying off-the-shelf tools compete with everyone using the same tools.
The margin targets reflect fundamentally different ambitions. PE aims for incremental improvement within existing business models. AI roll-ups aim for step-function changes to the model itself.
Complete guide · PDF
Founder’s Guide to AI-Enabled Roll-Ups
The full 13-chapter playbook in one designed file — every chapter refreshed to July 2026, plus three you won’t find on the site: building your own roll-up, when roll-ups break, and the UK and Europe map. 103 pages, free to download.
Download the guide →What This Means If You're Evaluating This Space
For investors:
Competition for quality acquisitions is increasing. As more platforms chase similar targets—profitable services businesses with fragmented ownership and automation potential—multiples will rise. The arbitrage that existed when this was novel is narrowing.
The differentiation is now execution. Which platform can actually deliver 30-40% automation? Which can retain founders and maintain service quality through integration? Ask for evidence, not projections.
Early movers have real advantages. GC has been deploying this playbook for three-plus years. Integration processes refined, AI tools trained on real data, portfolio companies sharing learnings. Late entrants face a steeper curve.
For founders considering selling:
You have options. Multiple well-capitalised buyers are competing, which shifts negotiating leverage towards sellers. Terms are better than traditional PE typically offers—brand preservation, equity rollover, operational continuity.
The questions to ask any potential acquirer:
What AI capabilities do you have deployed today? Crescendo's 80% automation rate and Crete's "hundreds of hours saved monthly" are concrete. Vague promises about future AI integration are not. Ask to talk with acquired founders about what actually changed.
What happens to my brand, my team, my role? Get specifics. GC and Thrive emphasise preservation. Some platforms consolidate. The answer matters if you care about legacy.
What's your capital structure? Evergreen means no forced exit timeline. Traditional PE funds have clocks ticking. Ask when their current fund needs to return capital.
Who else have you acquired? Talk to those founders. Their experience tells you more than any pitch deck.
The competition for quality targets will intensify. Multiples will rise. Some platforms will overpay for mediocre businesses and struggle to turn them around. Others will build real capabilities and compound for decades.
For founders watching this unfold, too much capital and too many smart people are committed to AI roll-ups for this to be pure speculation. What separates the winners from the overpromisers is evidence—deployed AI, transformed margins, and founders who'd sell to them again. When evaluating offers, understanding complete investment strategies helps you assess whether a buyer's thesis is sound or speculative.
Ask for evidence. Talk to founders who've sold. And remember that permanent capital only matters if the people deploying it actually know what they're doing.
Common questions
Why is venture capital, not private equity, behind AI roll-ups?
The plan needs software built and improved over years, which is what venture firms fund. Private equity is set up to improve existing businesses within a fund's life, often 3 to 5 years per deal. The result is a hybrid: a private equity acquisition strategy with venture-style technology development on top.
How do General Catalyst and Thrive Holdings differ?
General Catalyst builds the AI in-house or incubates a software company first, then buys firms for it to run. Thrive Holdings works closely with OpenAI, which took an ownership stake in December 2025 and puts research, product and engineering teams inside Thrive's companies. Both pitch sellers on long-term ownership, with founders rolling equity and staying involved.
What should founders ask before selling to an AI roll-up?
What AI is running today in firms the buyer already owns, with measured results; what happens to your brand, team and role; how the capital is structured and whether a fund deadline sits behind it; and what founders who have already sold to them say. Evidence from existing firms counts for more than projections.
Continue to Chapter 3: Industries in the Crosshairs
Or return to: The Founder's Guide to AI-Enabled Roll-Ups (Hub)
Capital Founders OS is an educational platform for founders with $5M–$100M in assets. Frameworks for thinking about wealth — so you can make better decisions.
Explore more: Playbooks · Capital Signals · Wealth Architecture · Investment Strategy · Business Building · Life Design
Found this useful? Forward it to a founder who's thinking about this stuff. Got a question or disagree with something? Get in touch.
New here? Subscribe for one email a week.
