Sabertooth VC’s SPV model shows how AI capital is being rewired around deal access, not funds

Sabertooth VC is making a pointed argument about how venture capital can work in the AI era: you may not need a traditional fund to move serious capital into the hottest late-stage companies.

According to TechCrunch, Justin Ernest’s firm has used Special Purpose Vehicles, or SPVs, to pool money from about 30 family offices and smaller institutions and invest nearly $400 million over the past year into 10 companies, including Anthropic, Anduril, Databricks, PsiQuantum, and SpaceX. The pitch is simple and structurally unusual for a firm operating at this scale. Instead of raising a conventional multi-year fund, Sabertooth is assembling single-deal vehicles, one transaction at a time, and using them to give its investors access to cap tables that are usually closed to all but the largest and best-connected buyers.

That distinction matters more than it sounds. In traditional venture, the fund is the core operating unit. It aggregates LP commitments, invests over a fixed period, and centralizes governance in a manager who makes portfolio-level decisions under a defined mandate. Sabertooth’s model flips that sequence. The sponsor sources or secures allocations in individual late-stage rounds, then wraps those allocations into SPVs so a small group of LPs can participate without waiting for a fund close or building a long-term blind pool. Ernest told TechCrunch that forming a new VC fund can take 12 to 18 months. For AI markets, where the tempo of financing can matter as much as product iteration speed, that lag is not trivial.

A new funding chassis for AI scaleups

The practical appeal of the SPV model is access. Family offices and smaller institutions often want exposure to late-stage AI companies, but they do not always have a direct path into those cap tables. Sabertooth’s setup gives them one. The firm acts as the operator, coordinating allocations and packaging them into SPVs that function like single-deal funds. The LPs are not underwriting a diversified venture portfolio in the classic sense; they are buying into specific names, with the upside and concentration risk that come with that choice.

For founders, the benefit is speed and reach. A sponsor that can mobilize capital without waiting for a traditional fundraise can, at least in theory, help close financing faster and with fewer structural delays. That can be useful in late-stage AI rounds where demand can outstrip available allocations and where the difference between moving now and three months from now can affect hiring plans, inference capacity, cloud commitments, and product launch timing.

How the SPV model actually works

SPVs are not new, but Sabertooth is using them as the core operating model rather than as an occasional sidecar.

Mechanically, an SPV is formed for a single investment. The vehicle pools money from a set of investors, buys shares in one company, and then passes the economics of that deal through to the SPV’s participants. In Sabertooth’s case, the LP base is reportedly made up of about 30 family offices and smaller institutions. That structure gives those investors a way to participate in late-stage rounds without appearing directly on the startup’s cap table in the way each underlying investor might if they invested individually.

That simplicity has tradeoffs. A traditional fund has a clearer governance stack: one manager, one mandate, one fund-level set of economics. With SPVs, the sponsor’s influence is more transactional and more fragmented. The sponsor must source deals, negotiate access, handle vehicle formation, and communicate terms to LPs on a deal-by-deal basis. Fund-level oversight can become thinner, because the capital is assembled around individual opportunities rather than through a single pooled portfolio model.

In other words, SPVs can make access easier, but they do not make governance disappear. They just move it.

Technical implications for AI product rollout

For AI companies, especially those already operating at scale, the capital structure can shape operational decisions in ways that are easy to miss if you look only at the headline valuation.

A financing process that is faster and more modular can reduce uncertainty around follow-on capital and expand the set of investors able to participate in growth rounds. That matters for product teams that need to lock in compute, hire applied research and infrastructure engineers, and commit to deployment schedules that depend on predictable cash flow. If an AI startup can close capital with less friction, it can move faster on roadmap execution, cloud spend, and customer-facing rollouts.

But the influence is indirect rather than centralized. With SPVs, there is no single permanent fund manager sitting above the company for years shaping a full portfolio strategy. Instead, the structure can create a distributed cohort of LPs and deal sponsors whose incentives are aligned at the transaction level, not necessarily at the operating level. That can make the funding process more nimble, but it can also complicate the feedback loop between investors and founders when a product is moving from model training into production deployment.

For teams shipping AI tools into enterprise environments, that distinction is important. Late-stage capital can accelerate deployment, but it does not automatically improve governance around safety review, compliance, model updates, or customer rollout discipline. Those remain operational responsibilities of the startup, not the SPV.

Risks, governance, and market dynamics

The same structure that broadens access also concentrates certain decisions.

Because Sabertooth is the operator assembling and managing the deals, the sponsor becomes the key gatekeeper for sourcing, allocations, and vehicle setup. That can create a dependency on a small number of dealmakers and raise questions about how transparently each SPV is managed. LPs may have access to high-profile late-stage exposure, but they also inherit the limits of a single-deal structure: less diversification, less liquidity, and less visibility than they would have inside a conventional fund with longer-term portfolio oversight.

There is also the matter of market signaling. When a sponsor can rapidly aggregate money for a small set of companies, the structure can reinforce winner-take-most dynamics in AI by channeling additional capital toward already dominant players. That is not unique to SPVs, but the mechanism can make it easier to extend that pattern.

Regulatory scrutiny is a more cautious topic, and there is no evidence here of a specific enforcement issue. Still, as more capital moves through SPVs, the complexity of the underlying investor mix, disclosure practices, and secondary liquidity expectations becomes harder to ignore. Those are not abstract concerns. They are the practical frictions that show up when a financing structure scales faster than the administrative systems around it.

What this means for the AI funding playbook and developers

Sabertooth’s model suggests that the next chapter in AI financing may be less about who can raise the biggest traditional fund and more about who can secure the best access and distribute it efficiently.

If SPV-based financing proves durable, it could widen the set of institutions that can participate in late-stage AI without waiting for a standard venture fund cycle. That would matter not just for investors but for the companies building infrastructure, models, and tools on top of those rounds. Faster access to capital can shorten fundraising timelines, support quicker deployment, and shape hiring and infrastructure decisions earlier in the product lifecycle.

For developers and operators, the lesson is operational as much as financial. The investor base behind a company may become more fragmented, the cadence of deals may become more transactional, and the expectations around diligence may shift toward deal-by-deal execution instead of long-horizon fund relationships. For VCs, the challenge is obvious: if sponsors like Sabertooth can combine network access, speed, and institutional demand without a traditional fund, the old fund-led model will have to justify itself on more than branding or scale.

That is what makes Sabertooth interesting. The firm is not just buying into AI’s biggest names. It is testing whether venture itself can be unbundled into a faster, thinner, more targeted instrument — one that trades the cohesion of a fund for the velocity of an SPV and the promise of direct access to the companies shaping the market.