Every venture firm offers startups money, and money on the same terms looks much the same no matter who provides it. Michael Ovitz, the Hollywood agent who advised Marc Andreessen and Ben Horowitz when they started their firm, said it plainly: without something more, "their money is the same as anyone else's." Andreessen Horowitz, known as a16z, was built on two ideas about what that something more should be.
Software is eating the world
The first idea is the firm's investment thesis. In 2011, Andreessen wrote in The Wall Street Journal that "software is eating the world," and predicted that "many more industries [will] be disrupted by software, with new world-beating Silicon Valley companies doing the disruption." a16z has spent the years since backing those companies.
More than money
The second idea is how the firm helps those companies win. The New Yorker described most venture firms as operating like a guild, where each partner looks after their own companies with a small shared staff. a16z built a company instead. Its partners took far less pay than the industry norm, and the savings paid for specialists in recruiting, business development, and marketing. a16z calls this its platform. The model borrowed from Ovitz's talent agency, CAA, which replaced the single agent handling every part of a client's career with a team of specialists.
Scott Kupor, a16z's first employee, explained the reasoning: "If you're a young CEO, you probably don't know a bunch of customers who could buy your product." You probably don't know the right executives to hire either. At most firms, those introductions depend on the partner who leads the investment. a16z built teams to own those relationships for the whole firm, so partners could focus on strategy and the board while the platform handled the network. When the firm launched in 2009, Andreessen wrote that the goal was to help technical founders "develop the skills necessary to reach his or her full CEO potential."
The firm reports the results in volume. By 2017, a16z said its Executive Briefing Center, where large companies come to meet its startups, had hosted more than 12,000 briefings and events and produced more than $3.5 billion in potential deals. Founders describe what that looks like up close. Cursor's COO told writer Packy McCormick that a16z's team introduced the company "to nearly 200 CTOs at key target customers during our first year." Applied Intuition's CEO said, "Our first defense customer came through" one of those briefings. And when Databricks CEO Ali Ghodsi had tried for years to reach Microsoft's CEO, Satya Nadella, Horowitz sent Nadella a note about the company. Forbes reports that the partnership that followed put Databricks inside Microsoft's Azure cloud, and Microsoft later invested.
How it spread
Rivals were skeptical at first. In 2012, one venture capitalist told Forbes, "When you go to the doctor's office, you want to see the doctor, not the physician's assistant." By 2015, Okta's CEO told The New Yorker that every firm Okta talked to was promising recruiting help and customer introductions: "It's become the table stakes."
A platform made of people has three limits. It grows by hiring: a16z had 45 people helping its startups in 2012, and Kupor described a 100-person operating team by 2019. Much of what those people know lives in their experience and relationships, so it grows only as fast as the team does. And the most valuable help depends on a few people with rare access, like a partner who can reach Microsoft's CEO.
Software eats the platform
a16z's own thesis applies to its platform. Andreessen argued that software would disrupt industry after industry, and a platform team runs on the kind of work software keeps taking over: research, preparation, meetings, and handoffs. Take one task: preparing a startup to meet a large company's technology leaders. Someone has to read the buyer's public filings and job postings, learn what it is trying to fix and how it chooses vendors, check what happened when similar companies met this buyer before, and write a briefing for the founder. An AI agent, software that uses a model to carry out a task, can do that preparation by following a defined workflow and drawing on shared knowledge from every earlier meeting. After the meeting, what the team learned goes back into that knowledge, so the next startup starts from it. A person reviews the briefing and makes the introduction.
Software addresses the first two limits. It grows without hiring for each new company, and what it learns stays with the firm. The third limit is different. Horowitz's note to Nadella worked because of who sent it, and that kind of access still comes from people. But the people with the relationships and the system that does the work do not have to sit inside the same firm. An expert network built from agents, knowledge, and workflows can work alongside the investors who already have those relationships, with each doing the part it does best.