Andreessen Horowitz’s $90B Board Problem Is Coming for Venture Capital
Two board seats have put Andreessen Horowitz’s $90 billion model under DOJ scrutiny. Every founder who treats a VC board seat as free advice should pay attention.
The venture-capital board seat may be the most expensive “free help” in business.
Andreessen Horowitz has roughly $90 billion under management, and the US Department of Justice is examining whether two of its board seats crossed an antitrust line. That should make every founder with a crowded cap table sit up straight.
The core story: Ben Horowitz, Martin Casado and two increasingly similar companies
The DOJ has been investigating Andreessen Horowitz — a16z to anyone who has spent more than five minutes around Silicon Valley — for nearly a year, according to reporting first published on August 17.
The issue is not some smoking-gun email or a bloke walking out of a boardroom with a USB stick tucked into his sock. It is more mundane than that, which is exactly why it matters.
A16z co-founder Ben Horowitz sits on Databricks’ board. A16z general partner Martin Casado sits on Fivetran’s board. Both companies are backed by the firm. Both help enterprises collect, organise and analyse large quantities of data — an increasingly strategic position in an AI economy where data plumbing is hardly plumbing anymore. It is the mine, the railway and the toll booth.
Fivetran also acquired dbt Labs in June after a review that the DOJ ultimately cleared without conditions. Casado had held a dbt Labs board seat too. The merger appears to have made the competitive overlap harder to ignore, while the government’s separate inquiry continued.
No final decision has been made. The DOJ could close the matter without action. Databricks and the DOJ declined to comment to Bloomberg, while a16z and Fivetran did not respond to requests for comment.
Still, the question is deadly serious: can one venture firm effectively occupy boardrooms at companies that are now competitors, even if different partners hold the seats?
That is the sort of question founders usually wave away with a confident “we’ve got proper information barriers.” Maybe they do. Maybe they have a policy document, a lawyer and a Google Drive folder labelled Governance Final FINAL v7.
None of that changes the economic reality: board members receive more information than ordinary shareholders, and venture firms exist to turn information, access and judgement into returns.
The 112-year-old law is not the interesting part
The DOJ is looking at Section 8 of the Clayton Act, a 1914 law that restricts “interlocking directorates” — directors serving on the boards of competing corporations.
The lazy take is that this is regulators dusting off an antique law because they have nothing better to do. Rubbish.
The law is old because the problem is old. When competitors share board-level access, they can gain visibility into pricing, product road maps, customer concentration, hiring plans, acquisition targets and the exact points at which management is bluffing. You do not need an explicit conspiracy for that to distort competition. The incentive structure does plenty of work on its own.
What makes this case unusual is the venture-capital angle. A16z is not alleged to have placed the same human on both boards. The potential theory is broader: the firm itself may be sufficiently represented through Horowitz and Casado that the arrangement creates the kind of conflict Section 8 was designed to prevent.
That distinction is where the lawyers will earn their watches. It is also where the industry’s standard operating model gets uncomfortable.
Venture firms are built to back categories, not isolated companies. They invest early, take board seats, help recruit executives, shape strategy, make introductions and then continue funding winners as the market changes beneath them. The trouble is that modern software companies do not stay in neat little boxes.
A database company becomes an AI platform. A data-pipeline business becomes an analytics layer. An analytics layer becomes an application platform. Before long, yesterday’s adjacent investment is today’s direct competitor.
That is not a rare edge case. That is the game.
The second-order implication: portfolio maps are now risk maps
For founders, the practical question is not whether a16z is found to have done anything wrong. The practical question is whether your investor’s portfolio has become a competitive threat to you while everyone was busy congratulating themselves on “ecosystem value.”
A board seat is not merely a badge of confidence. It is access to the guts of the business.
The good investor uses that access to help you make better decisions. The wrong structure creates a problem even when every individual involved behaves honourably. I do not need to accuse anyone of bad faith to know that incentives matter. I have made enough investment decisions to know that smart people can talk themselves into believing a conflict is manageable right up until it is not.
Think about what happens if this kind of scrutiny expands across venture capital.
First, big firms may take fewer formal board seats in overlapping categories. Axios has pointed out that this would change venture investing’s value proposition. A firm can still write a cheque, but a founder buying “deep operational support” may receive less of it.
Second, more investors may use board observers, advisory roles or informal operating relationships. That may reduce one legal exposure while preserving much of the practical access. Founders should not mistake a different job title for a fundamentally different relationship.
Third, firms may need tougher internal conflict systems: clearer competitive maps, mandatory recusal processes, formal information barriers and a willingness to give up a seat when two portfolio companies begin colliding.
The last one is the real test. Most governance policies sound terrific before the money is at stake. The quality of a firm is revealed when it has to choose between preserving access and doing the clean thing.
The overlooked angle: this could be good for founders
Most founders reflexively want a marquee VC on the board. I get it. The logo helps recruiting. The firm knows buyers. The partner has a direct line to the next round of capital. In rough markets, that can be worth a lot.
But founders often underprice independence.
If a single investor sits across too many related companies, the founder can become one node in someone else’s portfolio strategy. Your company may still win. But your strategic options can get foggier: which potential acquisition gets shown to whom? Which executive candidate has already been discussed elsewhere? Which product direction is deemed “too competitive” with another holding?
Again, this is not a claim that any particular investor is misusing information. It is a governance reality. The fewer conflicting loyalties around your board table, the easier it is for directors to pursue one thing only: the best outcome for your company.
There is a contrarian point here too. A less board-heavy VC model might force founders to build stronger companies.
Too many startup boards are outsourced executive teams with better clothes. Founders wait for investor introductions, investor hiring recommendations, investor strategic wisdom and investor approval. Then they call it support.
Support is valuable. Dependency is not.
A board should sharpen management, challenge assumptions and help during genuine moments of leverage or crisis. It should not become a substitute for the founder having a clear market view and enough backbone to make a decision.
What this means for you
If you are raising money, running a company or sitting on a board, do these five things this week.
1. Map your investors’ portfolios properly. Do not just scan the companies listed on their website. Ask which investments, funds, board roles, observer roles and operating relationships touch your market now — and which could plausibly touch it in 12 months.
2. Treat “adjacent” as a temporary word. Your product category will move. So will theirs. Put a regular conflict review on the board calendar, especially after a major product launch, merger or strategic pivot.
3. Get the information rules in writing. Who can access board materials? What is shared with the wider investment firm? When must a director recuse themselves? Vague trust is lovely until it costs you a deal.
4. Build a board that can disagree. If every director comes from the same investor ecosystem, you do not have governance. You have a group chat with fiduciary duties.
5. Remember what the board seat is for. It is not a trophy for the investor or a signal for LinkedIn. It is a legal and strategic responsibility. Give it to people who are willing and able to act solely for the company when the pressure arrives.
The DOJ’s a16z inquiry may end quietly. But the underlying issue will not. AI is making formerly separate software categories collide at speed, and venture firms are embedded across the whole map.
That means governance is no longer paperwork for when the lawyers turn up. It is strategy. Founders who understand that early will keep more control, make cleaner decisions and avoid discovering too late that their most helpful investor also had a seat at the other table.