Andreessen Horowitz’s 2 Board Seats Face a Clayton Act Probe
Two board seats. A $190 billion valuation. And a DOJ probe that could turn a16z’s “value-add” into venture capital’s biggest conflict problem.
Two board seats. A $190 billion valuation. And a DOJ probe that could turn a16z’s “value-add” into venture capital’s biggest conflict problem.
That is a much bigger problem for venture capital than one awkward investigation. If the US Department of Justice decides a16z cannot keep partners on the boards of companies that grow into competitors, plenty of firms will discover their prized board seats are not an advantage. They are a liability with a nice leather chair.
The core story: a16z, Databricks, Fivetran and a 112-year-old law
The Justice Department has reportedly spent nearly a year examining whether Andreessen Horowitz has run afoul of Section 8 of the Clayton Act, a 1914 antitrust provision aimed at “interlocking directorates.” The reported focus is straightforward: Ben Horowitz, a16z’s co-founder, sits on the board of Databricks; a16z partner Martin Casado sits on the board of Fivetran.
The complication is that Databricks and Fivetran now overlap in the market for data tooling. Databricks has pushed further into AI data pipelines and application connectors through Lakeflow, while Fivetran’s core business is moving data between systems. They may not have been direct competitors when a16z made the investments. That is exactly the point. In modern software, the edges of a market move faster than a board calendar.
This is not an allegation of proven wrongdoing. The DOJ has not publicly announced charges, the department declined to comment on the reported inquiry, and the investigation could end without action. But Bloomberg reported the probe, and the ensuing coverage has given the entire venture industry something it hates: a hard look at the machinery behind its claims of founder-friendly guidance. ([news.bloomberglaw.com](https://news.bloomberglaw.com/antitrust/andreessen-horowitz-focus-of-doj-probe-over-board-directors?utm_source=openai))
The scale matters. Databricks raised $5 billion at a $190 billion valuation in August 2026. That makes the board seat less like a ceremonial quarterly meeting and more like access to one of the most strategically important private software companies on the planet. If you sit in those rooms, you hear product plans, pricing pressure, customer losses, acquisition targets, hiring decisions and the bits of bad news that do not make the investor update. Pretending a firm can casually hold that information while advising an overlapping company is not sophisticated. It is convenient.
Why venture capital is suddenly nervous
Most people outside Silicon Valley think VCs primarily write cheques. The good ones do more than that. They recruit executives, help close customers, pressure-test strategy, arrange financing and, when necessary, tell a founder they are being an idiot before the market does it publicly.
A board seat is the formal version of that promise. It is also the mechanism by which a VC gets the deepest possible view into a company. That information advantage is valuable precisely because it is not available to everyone else.
Now take a large multi-stage firm with hundreds of investments. Its portfolio companies do not stay neatly inside the category printed on their original pitch deck. A database business becomes an AI platform. A data-pipeline company adds governance, analytics or agent tooling. A payments company becomes a bank-lite. A cyber company becomes an identity company. By the time lawyers agree on a market definition, founders have already built into each other’s patch.
That is why the a16z inquiry cuts deeper than a dispute over two directors. It tests whether the venture model has been relying on an unspoken exemption: that a VC firm can hold influence across a market as long as it calls the overlap accidental and asks its partners not to compare notes.
VCs call the internal separation a “Chinese wall.” Fine. Call it whatever you want. The real test is whether the wall works when the economic incentive is to make the whole portfolio more valuable. The firm owns shares in both companies. The partners share a platform, a brand and a carry pool. Founders are entitled to ask whether the wall is made of concrete or PowerPoint.
The overlooked angle: this could make boards worse, not cleaner
Here is the contrarian bit. Plenty of people will cheer if regulators force big VC firms to give up overlapping board seats. The instinct is fair: less conflicted advice sounds better for founders.
But there is a trade-off. If board seats become a compliance headache whenever a portfolio company changes direction, firms may simply take fewer of them. Axios noted that a broad Section 8 application to venture capital could push the industry in exactly that direction. ([axios.com](https://www.axios.com/2026/08/18/doj-andreessen-horowitz?utm_source=openai))
That sounds harmless until you remember what replaces a proper board relationship: a distant investor with economics but no formal accountability. The investor still owns stock. It still has opinions. It still has influence through follow-on financings, introductions and the ability to make a founder’s next round easier or harder. But it no longer has the same fiduciary obligation or formal governance role.
That is not automatically better for founders. It may mean less useful help when things are going badly, precisely when an experienced director earns their keep.
The right outcome is not “ban VCs from boards.” That would be daft. The right outcome is to stop treating a board seat as a collectible trophy and start treating it as what it is: a serious duty that becomes untenable when two businesses genuinely compete.
There is also a structural winner here: specialist firms. A focused investor that knows one industry deeply, makes fewer overlapping bets and manages conflicts with discipline can look more attractive than a giant generalist spraying capital across every fashionable part of software. TechCrunch reported that the broader venture fundraising market has already become barbell-shaped, with capital flowing to giant brand-name managers at one end and tightly focused specialists at the other. This investigation could accelerate that split. ([techcrunch.com](https://techcrunch.com/2026/08/18/reach-capital-raises-265m-fund-v-to-back-ai-founders-building-to-expand-human-potential/?utm_source=openai))
The second-order implications for founders and investors
For founders, the lesson is not to avoid elite investors. That would be like refusing to hire a brilliant operator because they have seen your competitors’ businesses before. Experience is useful. Pattern recognition is useful. Capital is useful.
The lesson is to be much more explicit about conflicts before you take the money.
Ask the questions that feel mildly impolite in the partner meeting, because they become bloody important after the wire lands:
- Which companies in your portfolio could become competitors if we expand our product line? - Does anyone at your firm sit on their board, advise them or hold observer rights? - What exactly happens if we overlap in 12 months? - Who gets access to our board materials, data room and strategic plan? - Will you agree in writing to information barriers and a resignation process if a conflict emerges?
If the answer is vague, you have your answer. “We manage that internally” is not a governance policy. It is a sentence designed to end the conversation.
For investors, this is a warning to audit the portfolio before the regulator does it for you. Map every board seat against current product overlap, not the market category that existed when the investment committee approved the deal. Review observer rights too. Check who receives board packs. Decide in advance who steps back when a collision happens.
Most importantly, do not confuse financial exposure with entitlement to information. You can own shares in two companies without believing you deserve to sit in both war rooms.
What this means for you
If you are a founder, build a conflict register this week. It can be a simple document listing every investor, their relevant portfolio companies, board roles, observer rights and any areas where your roadmap could collide. Update it every quarter, not after a problem blows up.
If you are raising capital, make governance part of diligence on the investor. Founders obsess over valuation, liquidation preferences and pro-rata rights, then accept a board relationship on vibes. That is amateur hour. A slightly lower valuation from an investor who will be fully in your corner can be far more valuable than a flashy logo with competing loyalties.
If you are an operator, do not wait for the board to ask the hard questions. Put market overlap and confidential-information controls on the agenda yourself. The best operators do not treat governance as legal housekeeping. They treat it as protection for the business they are building.
And if you are an investor, remember the blunt version: access is not the same as value. The reason founders tolerate a VC in the room is that the VC is supposed to make the company better. The moment the room becomes a conduit for portfolio-wide advantage, the trust is gone.
The DOJ may ultimately do nothing to Andreessen Horowitz. But the sensible conclusion is already clear. In a world where every software company expands into everybody else’s market, board-seat conflicts are no longer edge cases. They are part of the business model. Smart founders and smart funds will act like it before the lawyers make them.