Continuation Vehicles Become Exit Infrastructure, AI Valuations Split by Stack Layer and Proof Point

By DripPublished Updated

The short version

Private equity and growth teams are being forced to operate with new exit plumbing and a more granular AI pricing playbook, changing how they source, underwrite, and defend value.

This week’s developments

  • Continuation vehicles are now core exit infrastructure, not rescue deals — sponsors need secondaries fluency, LP-liquidity structuring, and longer-hold portfolio management.
  • AI valuation is splitting by stack layer and proof point — investors must benchmark each company with the right metric set, not rely on a single AI multiple.

Continuation Vehicles Move from Tactic to Exit Infrastructure

Single-asset “trophy” continuation vehicles have become a scaled PE workflow, not a one-off fix: GP-led secondaries reached about $28B in H1 2024, continuation vehicles were roughly 90% of that volume, and they represented an estimated 13–20% of PE exits in 2024–25. That matters because sponsors are now using them to hold top assets longer while still delivering LP liquidity, and the model is spreading into venture and growth where weak IPO/M&A markets make traditional exits unattractive. For deal teams, the work shifts toward secondary pricing, LP option design, conflict management, and packaging products for both institutional and private-wealth channels.

How should your team adapt to continuation vehicles becoming core exits?

If you're an individual contributor

Continuation vehicles are turning into a core PE exit path, so your edge is shifting from pure deal execution to being the person who can price secondaries, model LP outcomes, and spot conflicts before they become problems.

Build fluency in continuation-vehicle mechanics, LP option structures, and secondary valuation now, because the juniors who can package a hold-and-liquidity story will be more useful than those who only know how to run a standard sale process.

If you manage a team

Your team’s work is moving beyond traditional exit prep into a more complex workflow where pricing, structuring, and stakeholder management matter as much as the asset itself, and the people who can handle that complexity will become the go-to bench.

Coach your team to think in terms of LP choice architecture, conflict management, and cross-channel packaging, because the next layer of capability is not just running processes faster but managing more nuanced liquidity solutions.

If you lead the organization

Continuation vehicles are no longer a fallback tactic; they are becoming exit infrastructure, which means your platform now needs a repeatable secondary capability, not an occasional bespoke solution.

Reassess whether your org has the right pricing, legal, and distribution talent to run GP-led processes at scale, because firms that can industrialize continuation vehicles will preserve hold optionality and liquidity while slower platforms get trapped in weak exit markets.

AI Valuation Splits by Stack Layer and Proof Point

Benchmarking moved down a layer this week: investors are no longer paying a generic AI premium, but valuing foundation labs, infrastructure/cloud, and vertical applications against different anchors, including segment-specific ARR multiples, secondary pricing, revenue per employee, valuation per employee, and compute utilization. Late-stage ranges now diverge sharply: roughly 15–50x ARR for foundation labs, 10–20x for AI infra/cloud, and 20–45x for vertical AI apps. OpenAI and Anthropic are increasingly treated as ceiling cases, not universal comps, while secondary markets are reinforcing that discipline with reported private price discovery around OpenAI at about $500B, Anthropic at $183B+, and Databricks above $100B.

The market is still rewarding picks-and-shovels, but with harder proof tests: contracted GPU/cloud economics for infra, revenue scale and monetization for model providers, and measurable productivity or workflow gains for the application layer. That is a clear break from 2023 through early 2025, when broad EV/revenue medians of roughly 24–30x and narrative premiums of about 33% often carried deals, including many at zero revenue. For PE and growth teams, the implication is direct: diligence now has to separate real AI economics from branded software exposure, and the people who matter most are those who can benchmark at the stack level and translate operating data into underwriting conviction.

How should our team adjust AI valuation comps by stack layer?

If you're an individual contributor

Your edge is no longer spotting “AI” broadly — it’s proving which layer of the stack actually deserves a premium, so the analysts and associates who can benchmark ARR, secondary pricing, and unit economics by segment will be the ones trusted on live deals.

Build fluency in stack-level underwriting now: know how to compare foundation labs, infra/cloud, and vertical apps against the right anchors, and get comfortable translating operating metrics like revenue per employee and compute utilization into a defensible valuation view.

If you manage a team

Your team can’t rely on narrative-driven AI comps anymore — juniors who still pitch generic premium stories will look behind the market, while the ones who can separate real economics from branded software exposure will become the go-to support on diligence.

Coach the team to stop using one-size-fits-all multiples and start segmenting every AI asset by proof point, with recurring workstreams around contracted economics, monetization quality, and measurable productivity gains so your coverage is sharper than the market’s headline comps.

If you lead the organization

This is a valuation discipline reset: if your org still underwrites AI with broad software heuristics, you are already mispricing risk and missing where the market is actually paying up, especially as secondary marks harden the ceiling cases.

Rebuild the investment process around stack-layer benchmarks and operating evidence — your team needs dedicated capability to test model economics, infra contracts, and workflow ROI, because the next winners will be the firms that can distinguish true AI economics from expensive label-chasing.

Stay ahead in Private Equity & Growth Investing

Get the weekly Private Equity & Growth Investing brief in your inbox — the developments, what they mean by seniority, and what to do next.

Want this for the accounts and people you track? Explore Drip.