The Fund Is Dead, Long Live the Deal

Venture capital was built on a simple premise: give us your money, trust us for a decade, and we'll return multiples. For fifty years, that contract held. Limited partners committed capital to blind pools. General partners deployed it across portfolios. Everyone accepted the J-curve, the illiquidity, the opacity. The structure worked because LPs had no alternative. Finding, evaluating, and backing startups required infrastructure that only dedicated firms could provide.

That infrastructure advantage is evaporating. And the data from early 2026 suggests the consequences are arriving faster than most VCs want to admit.

On April 8th, Exec Sum posted a single line that captured the moment: "Family offices are ditching VCs for direct AI startup bets." It wasn't news to anyone paying attention. But the fact that it's now a headline -- rather than a whispered trend at LP conferences -- tells you where the Overton window has shifted.

The thesis is straightforward: more LPs will invest directly into companies, and more investors who would have raised traditional funds will instead operate as deal-by-deal SPV shops. The traditional venture fund -- ten-year lockup, 2/20 fee structure, portfolio of 25-35 companies -- is not disappearing. But it is losing its monopoly on how venture capital gets deployed.

The Numbers Behind the Shift

J.P. Morgan's 2026 Global Family Office Report surveyed 333 family offices across 30 countries, with an average net worth of $1.6 billion. The headline finding: 65% plan to prioritize artificial intelligence investments. But the more revealing statistic is that 57% currently have zero exposure to growth equity or venture capital. Only 3.3% of their portfolios sit in growth and VC allocations.

That gap between intention (65% want AI) and current allocation (3.3% in VC) is the story. These families want exposure to private technology companies. They're just not sure the traditional fund is the right way to get it.

The behavior data confirms this. By February 2026, family offices had already made 41 direct startup investments -- a pace that, if sustained, would dwarf prior years. These aren't small checks. Positron, an AI chip startup building energy-efficient inference hardware, closed a $230 million Series B led not by Sequoia or a16z, but by Arena Private Wealth, Jump Trading, and the Qatar Investment Authority. A billion-dollar valuation. No traditional VC lead.

The broader numbers are staggering. AI venture funding hit $211 billion in 2025. In Q1 2026 alone, it reached $239 billion. AI infrastructure specifically -- chips, networking, cooling, power -- drew $109.3 billion in VC funding in 2025. Capital of this magnitude doesn't flow exclusively through ten-year blind pools. It flows through whatever channel offers speed, precision, and control.

Why LPs Are Going Direct

Four structural forces are driving this.

Diligence costs collapsed. The historical moat of venture capital was information asymmetry. GPs had proprietary deal flow, deep networks, and the analytical infrastructure to evaluate hundreds of companies. Today, a family office CIO with AI-powered research tools, access to data rooms, and a few well-placed advisors can replicate much of that capability. The screening and evaluation work that used to justify 2% annual management fees is increasingly automatable.

Speed mismatch. AI company value curves now move on 6-to-18-month timescales, not the 7-to-10-year cycles traditional funds were designed for. Companies like Positron go from founding to billion-dollar valuation in under two years. LPs watching from inside a 2023-vintage fund see their committed capital sitting in reserves while the best deals close without them. SPVs and direct investments match the velocity of the market. Traditional funds, by design, cannot.

Precision over diversification. Instead of committing to a fund with 30 portfolio companies -- some of which are "deployment discipline" fillers rather than high-conviction bets -- LPs increasingly want to pick their spots. As Acquinox Capital put it in January: "Capital is becoming more selective, not scarcer. Investors are willing to deploy, but only when risk, structure, and pricing align." The deal-by-deal model lets LPs own fewer, better assets instead of broader exposure at any cost.

The fee revolt is real. A $200 million fund charging 2/20 collects $4 million annually in management fees before deploying a dollar. Over a ten-year life, that's $40 million in fees on a $200 million vehicle -- plus 20% of profits. For LPs watching median VC returns compress, that math has become harder to defend. SPVs typically charge only carry (no annual management fee), and direct investments eliminate the fee layer entirely.

The New VC Is Not a Fund

The shift isn't just happening on the LP side. A new generation of investors is building careers without ever raising a traditional fund.

16VC made this explicit in February 2026, publishing "Why 16VC Invests via SPVs Instead of a Traditional Fund." Their argument: the traditional fund structure optimizes for capital deployment, not for conviction. By operating deal-by-deal, they give their LPs transparency on every company, every valuation, every set of terms -- and the ability to opt in or out. No blind pool. No deployment pressure. No filler deals to meet allocation targets.

Elad Gil represents the logical endpoint of this evolution. A single-person, multi-billion-dollar venture operation. His third fund raised over $1 billion. He's made 241 investments. Companies like Anduril, Instacart, and Stripe sit in his portfolio. No partner meetings. No investment committee. No platform team. One person with conviction and capital, deploying at a pace and scale that rivals institutional firms.

Patrick Ryan, co-founder of Odin (a platform that has onboarded over 10,000 VCs, angels, and founders globally), published a piece in February 2025 with a title that left little ambiguity: "No, You Shouldn't Raise a Fund." His argument: for emerging managers, the SPV path lets you build a track record, prove your sourcing ability, and return capital to LPs -- all without the two-year fundraising slog and the regulatory overhead of a pooled vehicle. The fund comes later, if it comes at all.

The infrastructure makes this possible. Platforms like AngelList, Carta, Allocations, Odin, and Sydecar have turned SPV creation from a six-week legal project into a same-day workflow. You can spin up a compliant vehicle, onboard investors, collect capital, and close a round in days. Administration is automated. As the New York Venture Hub observed in January 2026: "SPVs went from an edge case to a core strategy."

And the performance data supports the model. VC Beast's analysis of solo GP returns found that single-partner funds are generating top-quartile returns at higher rates than multi-partner firms. The intuition is simple: concentrated conviction, lower overhead, faster decisions, and no committee dynamics diluting the portfolio.

The Contrarian Case: Why the Fund Model Survives

Now for the part that most SPV evangelists skip. There are real, structural reasons the traditional fund model persists -- and they're not just inertia.

Venture returns follow a power law, and power laws demand portfolios. The defining feature of venture capital is that a small number of investments generate the vast majority of returns. In any given vintage, 5-10% of companies produce essentially all the value. This is not a theory; it's the most empirically robust finding in venture data. An LP making deal-by-deal decisions is, by definition, making a concentrated bet that they can identify the outliers in advance. History suggests most investors -- even very good ones -- cannot do this consistently. The traditional fund structure, with its 25-35 company portfolio, is specifically designed to capture power-law dynamics. A deal-by-deal investor who passes on one "unexciting" Series A because the metrics don't look clean might be passing on the next Stripe.

Decision fatigue is real, and it compounds. When an LP commits to a fund, they make one decision. When they invest deal-by-deal, they make dozens -- each requiring independent evaluation of the company, the terms, the co-investors, the market timing. For a family office running a small team, this is an enormous analytical burden. The whole point of allocating to a GP was to outsource that work. Many LPs who experiment with direct investing discover they lack the bandwidth, the deal flow, or the pattern recognition to do it well. The Whalesbook report flagged this directly: there's a growing concern about "tourist capital" -- investors who show up for hot deals but lack the operational depth to add strategic value or weather downturns.

Management fees fund things that matter. The anti-fee narrative treats the 2% management fee as pure rent extraction. In practice, it funds recruiting teams that help portfolio companies hire, go-to-market support, regulatory navigation, and follow-on investment capacity. The best firms -- Andreessen Horowitz's platform model is the canonical example -- use management fees to build genuine operational infrastructure that makes their portfolio companies more valuable. SPVs, by design, cannot fund this. A deal-by-deal investor writes a check and wishes the founder well. That's fine for a $10 million seed round. It's less fine when a company needs to navigate an enterprise sales cycle, a regulatory filing, or a complex international expansion.

Adverse selection is the quiet killer. When a GP raises an SPV for a specific deal, which deals get SPV'd? Often, it's the hottest deal -- the one where the GP couldn't get enough allocation from their main fund, or the one that's already been de-risked by market validation. The less glamorous but potentially higher-returning early bets stay inside the fund. LPs who only participate via SPVs may systematically get offered the deals with the most hype and the least remaining upside.

LPs are not VCs. The deepest contrarian argument is also the simplest: evaluating startups is a skill. It requires pattern recognition developed over hundreds or thousands of company evaluations, deep domain expertise, reference networks, and -- crucially -- the ability to say no to 99% of what you see. Family offices that made their wealth in real estate, manufacturing, or public equities may be excellent capital allocators in those domains. That expertise does not automatically transfer to evaluating whether a pre-revenue AI infrastructure company has defensible technology. The 57% of family offices with zero VC exposure in J.P. Morgan's survey aren't there by accident. Many of them understand that they lack the apparatus to do this well.

What Actually Happens: The Messy Middle

The reality will be neither the death of funds nor the triumph of SPVs. It will be a structural bifurcation.

At the top, mega-funds consolidate. Firms like Sequoia, a16z, and Founders Fund have brand gravity, proprietary deal flow, and operational infrastructure that no SPV shop can replicate. LPs will continue to commit to these funds -- not because the structure is optimal, but because access to the best GPs is the scarcest resource in venture. These firms may increasingly offer co-investment rights and deal-by-deal opt-ins alongside their main funds, creating hybrid structures that give LPs more control without abandoning the portfolio model.

In the middle, funds get squeezed. The $100-500 million generalist fund with a mediocre track record and no clear differentiation is the most vulnerable. These firms can't offer the brand premium of a top-tier fund or the precision of a deal-by-deal model. LPs will increasingly ask: "Why am I paying 2/20 for a portfolio I could assemble myself through SPVs?" The answer, for many mid-tier funds, will be unsatisfying.

At the emerging end, SPVs become the default on-ramp. New managers will build track records deal-by-deal before raising pooled vehicles. The traditional path -- work at a VC firm for a decade, then raise a Fund I -- is being replaced by a path that starts with syndicates and SPVs. Platforms like Odin, AngelList, and Allocations have made this not just possible but preferable. You prove yourself with real returns, then institutionalize.

Family offices build internal capabilities. The largest family offices will hire dedicated venture teams. Some already have. This mirrors what happened in private equity twenty years ago: the biggest LPs (endowments, sovereign wealth funds, large family offices) built direct investing capabilities and reduced their dependence on external managers. The same arc is now playing out in venture, accelerated by AI tools that compress the analytical work.

The fee structure evolves. Expect more 1/20 or 1.5/15 structures. Expect more hurdle rates. Expect more co-investment rights built into fund LPAs. The 2/20 standard persists at the top but erodes everywhere else. The SPV model -- carry only, no management fee -- becomes the benchmark that every emerging manager is compared against, even if they ultimately raise a fund.

The Deeper Shift

What's really happening isn't a structural debate about SPVs versus funds. It's a power transfer.

For decades, GPs controlled the relationship. They chose their LPs. They set the terms. They decided which companies got funded and at what price. LPs were, structurally, passive. The information asymmetry, the operational complexity, and the illiquidity all favored the GP.

That asymmetry is compressing. LPs now have better data, better tools, more options, and less patience. The 57% of family offices with no venture exposure aren't staying out forever -- they're waiting for the right entry point, on their terms. When they arrive, they'll demand transparency, deal-level discretion, and aligned economics. Some will invest through funds. Many will invest alongside funds. An increasing number will invest instead of funds.

The venture capital industry isn't dying. But the venture capital fund, as the sole vehicle for deploying risk capital into startups, is losing its monopoly. What replaces it is messier, more fragmented, and more competitive. For founders, that's almost certainly good news: more sources of capital, more leverage in negotiations, faster closes. For LPs, it's a mixed bag: more control, but more work. For GPs who built their careers on the structural advantages of the fund model, it's a wake-up call.

The deal, not the fund, is becoming the atomic unit of venture capital. The investors who understand that -- and build their operations around it -- will define the next era.

Sources: J.P. Morgan 2026 Global Family Office Report; Whalesbook, "Family Offices Cut Out VCs in AI Funding Surge" (April 2026); New York Venture Hub, "SPVs and the Reshaping of Venture Capital" (January 2026); Acquinox Capital, "Why Some Investors Are Dumping Funds for Deal-by-Deal Opportunities" (January 2026); 16VC, "Why 16VC Invests via SPVs Instead of a Traditional Fund" (February 2026); PitchBook, "Elad Gil Raises Over $1B for Third Fund"; VC Beast, "Why Solo GPs Are Outperforming" (March 2026); Patrick Ryan / Odin, "No, You Shouldn't Raise a Fund" (February 2025); Exec Sum (@exec_sum), April 8, 2026.