Why the venture concentration headlines say nothing about your firm

Two things called concentration

Search for concentration in venture and you will find a great deal of well-reported writing, almost all of it about the industry. Capital concentrating into fewer funds, dollars concentrating into fewer deals, returns concentrating into fewer companies. It is a real story and the numbers are striking.

None of it is about the thing this engine measures, which is how one firm's own capital is distributed across the funds it runs. The two share a word and answer different questions, and a General Partner who reads the macro coverage and feels informed about their own position has not learned what they think they have.

What the macro numbers actually say

The industry story is worth knowing on its own terms, and it runs at two levels.

At the fund level, PitchBook found that in 2024, 30 firms raised 75 percent of all capital raised by US venture funds, and just nine of them took in $35 billion, half the total. Andreessen Horowitz alone brought in more than 11 percent of everything raised. The same reporting notes how quickly this arrived: ten US funds over $1 billion in 2019, thirty-five by 2022. Emerging firms outside the top 30 closed $9.1 billion between them, 14 percent of all US commitments.

At the deal level, SaaStr reported that 41 percent of US venture dollars deployed in 2025 went to ten companies, $81.3 billion of $197.2 billion. OpenAI, xAI, and Anthropic between them raised $65 billion, close to a third of everything, with OpenAI alone at $40 billion.

TFX Capital draws the parallel that makes this more than a set of statistics. Writing in March 2026, from the perspective of someone who was a credit trader at Bear Stearns in 2008, they note that the five largest US banks then held more than half of total banking assets, and that what turned concentration into contagion was that so many institutions held near-identical exposures. Their reading of venture today names two dynamics: too much capital controlled by too few decision-makers, and market participants chasing overlapping opportunities with limited valuation sensitivity.

Why none of it tells you about your firm

Every one of those figures is computed across the industry. The unit of analysis is the market, and the finding is about how the market's capital is distributed among its participants.

This engine computes across one firm. The unit is your funds, and the finding is about how your capital is distributed among your positions.

They are also independent. A firm can be a small participant in a highly concentrated market and hold a beautifully distributed book. A firm can be a small participant in that same market and have most of its capital in two companies in one sector. The macro numbers are identical in both cases, because the macro numbers were never about either firm.

The mechanism TFX names is the firm-level one

There is a genuine connection between the two, and it is worth being precise about where it lies.

What made 2008 dangerous, in TFX's reading, was not that banks were large. It was that they held the same things, so a single correction hit all of them at once. Applied to venture, the risk is not fund size but overlapping exposure: many firms holding correlated positions in the same small set of companies and the same handful of categories.

That is a description of concentration inside portfolios, which is exactly the level this engine reads. The macro data cannot see it, because it counts dollars raised and dollars deployed rather than what any individual book holds. The industry-level statistic is the symptom; the firm-level distribution is where the exposure actually sits. Reading the first and skipping the second means watching the weather instead of checking the roof.

Your LPs have been asking the firm-level question for years

There is a further reason the firm-level view is the one that matters, and it long predates the current cycle: it is the version institutional LPs already ask about.

ILPA's Principles 3.0 addresses this directly. Its guidance is that General Partners should seek to limit the number of overlapping investments between their own funds, should avoid transfers of assets between funds, and should disclose overlapping investments, along with the fees earned on them, to their limited partners. The language is qualitative rather than a numeric cap, but the expectation is plain: overlap between a firm's funds is something an LP expects to be managed and told about.

Mark Suster set out the practitioner version well before that, and his framing is the most useful one for understanding why LPs care. The problem with investing in the same company from two funds is that the funds have different LPs. Fund II's investors reasonably ask whether their capital is making an independent decision or protecting an earlier fund's position. Fund I's investors who did not come into Fund II end up owning less of the company than they otherwise would. And in a down round the conflict is at its sharpest, because the second fund's money can be used in a way that damages the first fund's. Firms manage this with advisory committees and, for the genuinely conflicted cases, dedicated conflict committees.

Which means a firm's cross-fund single-company exposure is not only a risk measurement. It is a governance fact that a sophisticated LP will ask about in diligence, and a firm that has not measured it will be answering from memory.

What to do with each

Read the industry numbers as conditions. They tell you what kind of market you are raising and deploying into: that LP capital is harder to reach outside the top tier, that the largest rounds are absorbing a share of dollars that has no recent precedent, and that a correction in one category would be widely felt. Those are things to plan around, not things to fix.

Read your own distribution as a position. How much of your deployed capital sits in your largest sector, whether that is consistent with the scope you declared, and how much sits in any one company across every fund you run. Those are things you chose, mostly without deciding to, and every one of them is still adjustable with the capital you have not deployed yet.

The engine's hub explains what it reads, and the single-company exposure page covers the reading LPs are most likely to ask about.

Sources

• Venture Has Never Been More Concentrated: 41% of VC Going to Just 10 DealsSaaStr, 2025Deal-level concentration: $81.3 billion of $197.2 billion deployed in the US going to ten companies.
• 9 VC firms collected half of all money raised by US funds in 2024Rosie Bradbury, PitchBook, December 2024Fund-level concentration, and the primary source for the mega-fund figures. Carries a published correction that raised the top-nine share.
• Too Big To Fail, Part Deux? Capital Concentration in Venture CapitalTFX Capital, March 2026The 2008 parallel, written by a former Bear Stearns credit trader, and the clearest statement of why overlapping exposure is the actual mechanism.
• ILPA Principles 3.0: Fostering Transparency, Governance and Alignment of InterestsInstitutional Limited Partners Association, 2019Guidance that GPs should limit overlapping investments between their own funds and disclose them, with the fees earned on them, to LPs.
• Can VC's Invest Across Two Funds?Mark Suster, Both Sides of the Table, April 2010A practicing GP on why cross-fund investment is a conflict question, and what the differing LP bases across vintages actually disagree about.