a16z Partners on Competing Boards; DOJ Asks Uncomfortable Questions

Andreessen Horowitz, the $30 billion venture capital firm that has spent the last decade positioning itself as the thinking man's tech investor, now finds itself defending an arrangement that would make a seasoned antitrust lawyer spit out their coffee. Ben Horowitz sits on the board of Databricks, while his partner Martin Casado occupies a seat at Fivetran—two companies that compete directly in the data pipeline and transformation space. The Department of Justice, apparently having nothing better to do than read board rosters, has been investigating the arrangement for nearly a year, according to reports. When federal prosecutors start asking questions about your governance structure, you know the definition of "operational synergy" may need some revision.

Let us be clear about what these companies actually do: Databricks provides a lakehouse platform for data engineering and analytics, while Fivetran specializes in data integration and replication. They are not adjacent players—they are direct competitors courting the same enterprises, the same engineers, and the same budget lines. The theoretical conflict is not subtle. A venture partner with voting power at one company has every incentive to steer strategic decisions, partnership opportunities, and perhaps even technical roadmaps in a way that advantages his investment over the other. Whether this happens in practice, or whether both partners pinky-swear to recuse themselves at the right moments, is a question the DOJ apparently deemed worth 365 days of investigation time.

This is not a16z's first rodeo with board complexity—the firm runs one of venture capital's most ambitious portfolio strategies, where partners sit across dozens of companies simultaneously. That aggressive networking is partly why a16z has outperformed peers; it provides intelligence, leverage, and optionality. But there is a meaningful difference between sitting on boards of non-competing companies and volunteering to referee a cage match between two of your own investments. The SEC has historically been casual about dual board seats in venture, treating them as a feature of the ecosystem rather than a bug. The DOJ apparently disagrees, or at minimum, suspects the feature might constitute something closer to a pattern.

a16z's public defense, should it arrive, will almost certainly wrap itself in the language of "strategic value creation" and "portfolio diversification." Translation: our partners are smart enough to manage conflicts, and besides, we've already made money on both sides, so what's the harm? This is the venture capital equivalent of a bank saying it can safely advise both sides of a merger because its compliance department is "really good at preventing leaks." The argument is not technically false—it is just legally insufficient when regulators decide to start asking whether it should be allowed at all.

The practical risk here is straightforward: if the DOJ concludes that dual board seats on competing platforms constitute an antitrust violation or unfair competitive practice, a16z faces potential restrictions on future board appointments. More dangerously, any existing agreements or strategic decisions made under these conditions could theoretically be challenged by competitors or shareholders claiming market manipulation. Databricks and Fivetran investors have every reason to worry that their cap tables now carry regulatory uncertainty that was never properly disclosed. Even if the investigation concludes with a handshake and a memorandum of understanding, the precedent has been set: the antitrust era has finally noticed venture capital.

This moment crystallizes a broader reckoning in venture: the era of aggressive board consolidation is meeting the era of aggressive antitrust enforcement. For years, large firms have accumulated positions across entire market verticals, claiming that venture partnerships naturally require deep portfolio presence. The DOJ's investigation suggests that regulators are no longer content to assume good behavior, especially when the conflict is so direct and the competitive stakes so obvious. When a federal agency spends a year investigating your governance choices, you can no longer claim those choices are merely clever—they are now suspect.

The real scandal is not that a16z tried to optimize its position across two competing data companies; it is that anyone was surprised the DOJ noticed.