Ben & Marc: The 10x Future of Tech, Media, and Venture
Marc Andreessen and Ben Horowitz reveal how they built a16z's reputation machine: their controversial public presence isn't ego—it's deliberate infrastructure that lets portfolio companies "borrow" their force as a slingshot, and why AI's reinvention of computing makes traditional market sizing obsolete.
Read OriginalMy Notes (2)
Supply-Driven Markets
The core Substack thesis: provide the monetization mechanism and you bring into existence writers/content that don't exist today. That new supply creates demand that's currently invisible.
Everyone assumes demand is the constraint. It's not. There's enormous untapped appetite for smart, high-quality stuff. The bottleneck is supply. The infrastructure to monetize quality content didn't exist, so the content didn't exist.
People don't know what they want until you give it to them. No one asked for a Macintosh. No one asked for an iPhone. Supply-side innovation creates demand that surveys and market research can't detect.
From the transcript:
On the Substack investment thesis:
- "There was this chicken and egg thing right up front, which was like, well, can you really see, the internet's awash with content. People aren't paying for almost any of it. Like are they really gonna, you have to kind of squint and kind of say okay are people actually going to pay for any of the stuff that they're getting today."
- "I think the thing that in that case that we had faith in was basically this could be what we call a supply driven market. Which is, if you provide the monetization capability, then you're going to bring into existence writers and content that don't exist today. And that is going to create new demand that's not visible today. And then that demand is going to come back around and it's going to incent... so it was basically a bet that there's like an entire generation of high quality content that doesn't exist because the monetization mechanism doesn't exist."
- "That's certainly what Chris and his partners believe. And I think that's absolutely what's proven out. And so it's like a great example of founders that really see a future that doesn't exist yet. And it's obvious to them."
On the broader principle:
- "There's also, I think, just like enormous latent demand for actual smart high quality stuff. And particularly in media and in every kind of media. And I think the issue is not lack of demand. And I think the issue is lack of supply."
- "This kind of goes back to consumer marketing 101, which is people don't know what they want until it's given to them. Nobody ever asked for a Macintosh. Nobody ever asked for an iPhone. Like these things had to be designed and built and provided on the supply side before the demand materialized. And then the demand of course turned out to be far higher than anybody expected. And I think that same thing is exactly true of media."
On long-form podcasting as proof:
- "I think a great early existence proof of this was the success of long form podcasting. Which is, I remember my early conversations with some of the early long form podcasters and they're like it's the strangest thing we've ever seen because everybody tells us that the consumers have short attention spans but people are literally watching three-hour podcasts and we get the analytics and people are watching all the way to the end of the three-hour podcast."
- "So I view this very much as like this is one of these classic markets that's a barbell. Which is yeah you have a certain amount of whatever just mainstream filler on the one side, but you have this massive sort of untapped market for high quality content in basically every domain. And there just had, technological transformation, the existing structure of the media company was a structure that was designed for a world of centralized media. You need a new structure today and that's why we're so high on Substack."
On the potential scale:
- "When we look at Substack it's the exact same thing that you were just talking about with Databricks which is like wow this thing could be orders of magnitude, orders of magnitude multiples larger than anything we've seen so far. And frankly, I think we're starting to see that."
- "I think Substack could be like a thousand times the size of the existing content industry, whatever you want to call it, media news industry."
Market Sizing Discussion
Market sizing presumes you can predict future markets from current dynamics. This breaks completely when there's a supply-side breakthrough.
On the venture capital approach to market sizing:
"The venture lens I would put on this is, in terms of the mechanics of venture capital, the sort of classic venture capital triangle is team, product, and market. And you're always trying to kind of evaluate all three of those and people have always had different theories over the years of which is more important than the others and how they interact."
"The thing that basically every investor, public market investors, private investors, the thing that we're all trained to do is basically market sizing. We're trained to do technology analysis. We're trained to do background checks on people. And then we're trained to do market sizing. Like, okay, how big is this market? Because the classic adage is if you put a huge amount of effort into going after a small market you still get a small outcome."
"But there's a presumption in there which is that you can actually predict market sizes on these things. And the problem with that again is this sort of presumption that you can predict market sizes based on the dynamics that exist in the market today."
"But if there's a fundamental change on the supply side, if there's a fundamental breakthrough, a fundamental capability that doesn't exist yet, you're not going to be able to accurately model the market size because you can't see it yet. Like you've changed one of the major variables and you can't do that."
"You could call that the leap of faith or whatever, but it's like, okay, if you make the change in the supply side, then all of a sudden the market gets 10 or 100 or thousand times larger. Like you can almost never validate that with math at the time of the investment, but that's the thing that makes the outperformers really go."
"As Ben and I kind of go through our careers, we just see more and more examples where the mistake that we or others make is, oh, this must be, the market for Uber and Lyft must be the market for taxi cabs. Or the market for cloud software must be the same as the market for on-prem software. Or the market for GPUs must be the market for people who like to play games"
"We just keep seeing example after example after example where the significant enough technology change, product change on the supply side unlocks much larger markets. And I think that's going to be the single dominant trend in investing for the next 30 years. I think that's just going to telescope way out."
Summary used for search
TLDR
• a16z's loud, political brand was designed as transferable power—portfolio companies borrow their reputation at critical moments (first fund: 6 months of meetings for $300M; latest: two AMAs for $15B)
• Supply-driven markets create demand that doesn't exist yet: Substack monetization brought writers into existence, cloud software was 10x bigger than on-prem because technology unlocked latent demand
• AI breaks traditional market sizing—when you reinvent the computer itself, markets become 10-1000x larger than historical data suggests (mythical man-month is dead, Elon caught OpenAI by throwing money at it)
• The firm's culture document requires signing: "dream builders not dream killers," never publicly criticize entrepreneurs, reputation compounds but one mistake destroys more than ten good deeds build
• Genius inventors know technology but not the messy real world—VC is increasingly art over science because change accelerates faster than neural nets can train on history
In Detail
Andreessen and Horowitz explain a16z's counterintuitive strategy: they deliberately built a dominant public brand—appearing promotional and political—specifically so portfolio companies could borrow that reputation as a "slingshot" at critical growth moments. This wasn't ego; it was infrastructure. The proof: their first $300M fund took six months and countless meetings to raise; their latest $15B fund was raised with essentially two AMAs. Reputation is their compounding competitive advantage, and it transfers directly to founders who need to recruit top engineers, close customers, or navigate regulators.
Their investment thesis centers on supply-driven markets where breakthrough technology creates demand that doesn't exist yet. Substack worked because providing monetization capability brought high-quality writers and content into existence—the "non-fungible writer" who could finally leave publications and build independent brands. Similarly, cloud software wasn't just migrating the on-prem market; it was 10x larger because the technology unlocked latent demand. This pattern repeats: when you change a fundamental variable on the supply side, traditional market sizing becomes useless. AI represents the ultimate version of this—a complete reinvention of the computer that makes every problem solvable, breaking the mythical man-month (Elon caught OpenAI by throwing resources at it, something previously impossible in software).
The firm stays small while getting powerful through organizational design borrowed from original HP: autonomous groups (crypto, infra, apps) that feel like small companies but can tap into centralized brand and fundraising. Their culture document—which everyone must sign—mandates being "dream builders not dream killers," never publicly criticizing entrepreneurs. One mistake destroys more reputation than ten good deeds build. They're betting heavily on "Zoomers" (Gen Z founders) as unapologetically ambitious, AI-native, and freed from millennial guilt culture. The meta-insight: venture capital might be the "last job" because it requires the intangible art of betting on people with dreams in asymmetric domains—and this is becoming MORE art, LESS science as change accelerates beyond what historical data can predict.