a16z's Opening Viewpoint: When Supply-Side Leaps Forward, We Need a New Thinking Framework

Source: Sense AI

"If there is a fundamental breakthrough on the supply side, market size analysis becomes invalid."

In mid-January 2026, shortly after a16z announced raising over $15 billion for its new fund, Marc Andreessen and Ben Horowitz engaged in a deep conversation in Silicon Valley about AI, media, and the future. This was not a conventional interview but a systematic discussion centered around the concept of a "supply-side revolution."

Their core thesis is: true innovation does not come from optimizing existing demand but from creating unprecedented supply, forcing new demand to emerge. From Elon Musk's restructuring of Twitter to AI reshaping all market boundaries, this logic runs throughout. The result is not 10% growth but 10x, 100x, or even 1000x opportunities. This article is based on and expands upon content from the a16z official podcast, *The Ben & Marc Show*.

Speaker Bios

1. Marc Andreessen: Co-founder of a16z. At age 22, he created the popular graphical browser Mosaic, then co-founded Netscape, igniting the first wave of the internet revolution. His investment portfolio includes Facebook, Twitter, GitHub, among others.

2. Ben Horowitz: Co-founder of a16z. Formerly a product lead at Netscape, he later founded Opsware and served as CEO, taking the company public and successfully selling it. Author of *The Hard Thing About Hard Things* and *What You Do Is Who You Are*, he is regarded as one of Silicon Valley's most respected "startup mentors."

3. Packy McCormick: Author of the tech business analysis blog Not Boring, known for his deep, optimistic, and insightful analytical articles.

4. Erik Torenberg: General Partner at a16z, founder of the tech media network Turpentine, and host of *The Ben & Marc Show*.

Media Liberation: From 'Controlled' to 'Free'

Ben Horowitz: Packy, in your article, you mentioned that 2015 *New Yorker* piece, calling it the last golden age when media was still willing to be honest with you. I think no one is better suited than Marc to summarize what has happened in the media world over the past decade.

Packy McCormick: Marc, to what extent do you think this shift in the wind is due to your and a few others' public pushback, pulling the entire media environment back to its current somewhat "uncontrolled" or "liberated" state?

Marc Andreessen: I would describe today's information environment with three words: the neutral "uncontrolled," the pejorative "anarchy," and the complimentary "liberated." We are indeed moving towards a more open world, but I don't want to package myself as some moral hero.

This story starts in 1993. Back then, I refused to add censorship features to the Mosaic browser. If I had, the world today might be a different dystopia. Later, starting in 2007, I spent over a decade on Facebook's board, personally experiencing that company's rollercoaster-like crazy ten years. But the real turning points were two things. First, Elon's acquisition of Twitter changed everything. Second, I must give a huge shout-out to Substack. As one of their earliest and largest external investors, we are incredibly proud of them. Under immense pressure, they held the line on free speech, standing firm, purely on principle.

Ben Horowitz: That's right. They were truly besieged by various "anti-speech" forces at the time.

Marc Andreessen: Yes, but they held firm. Elon fought the censors in a very public, visible way. Substack, as a small company, had to fight more difficult battles behind the scenes to preserve the integrity of its platform.

Supply Precedes Demand: The Investment Logic Behind Substack

Packy McCormick: That's a perfect segue to talk about your investment in Substack. I use Substack too, but I still think it was a crazy investment. Were you purely considering commercial returns, or did you believe it was beneficial for the future?

Marc Andreessen: We don't invest for sentiment; the goal is always returns. We believed then, and believe even more strongly now, that Substack has the opportunity to become an industry cornerstone. We fell in love with those guys (the founders); they are the kind of people who are easy to like. Also, I lived through the early internet, especially the golden age of blogging, which was a very special period. Blogs created a massive amount of intellectual content that wouldn't have existed otherwise, but there were always a series of problems, one of which was that bloggers struggled to make money.

We invested in Substack because we deeply believed it was a "supply-driven market." In an internet era flooded with free content, the question seemed to be, "Will people really pay for content?" But we saw the other side: If you can provide creators with monetization capabilities, you will create authors and content that don't exist today, and this will generate new demand that is currently invisible.

This was essentially a bet: a bet that an entire generation of high-quality content had not yet been born due to a lack of monetization mechanisms. Just as no one asked for the Macintosh or the iPhone, demand doesn't manifest until the supply-side product is provided. When the brand shifts from *The New York Times* or *The Wall Street Journal* to the author themselves, Substack is a huge enabler. They created what you might call the "non-fungible writer."

Ben Horowitz: At the time, Hish, one of Substack's founders, told us that many writers were trapped in traditional media institutions. I asked, "Are they really trapped, or did they build their own prison?" He assured us that given an independent path, many would be thrilled to break free and write from different angles.

When Supply Leaps, Market Sizing Becomes Invalid

Packy McCormick: Ben, I really liked that email you sent to Ali, the founder of Databricks, saying he was underestimating his business and that it would be 10 times larger than Oracle. What's the mechanism for this "10x" judgment?

Ben Horowitz: It's actually quite simple. If you compare on-premise software to cloud software, like PeopleSoft vs. Workday, the cloud version is 10x larger. Oracle can be seen as the on-premise version of Databricks. AI wasn't as hot then as it is now, and the emergence of AI has only accelerated my prediction. What I was doing at the time was essentially a psychological game: Ali is very cautious, and I had to tap into his specific psychological frequency to steer the company's trajectory in the right direction.

Marc Andreessen: I think Substack could be worth 1000x the existing content industry. The reason is that the latent demand for high-quality, in-depth content is enormous. The problem isn't a lack of demand but a lack of supply.

Marc Andreessen: I grew up hearing people complain about TV, calling it a wasteland of nonsense. Today, people are having a "moral panic" about short videos and TikTok, saying everyone just wants to watch cats playing the piano. By the way, I love cat videos; AI-generated cat videos are my favorite genre now. But the reality is, this is a classic "barbell market." On one end is a massive amount of time-filling filler content; on the other end, there is a huge, unmet craving for high-quality content in every field. The success of long-form podcasts is the best proof. The existing media structure was designed for a centralized world. Today we need new structures, which is why we are bullish on Substack.

Marc Andreessen: This leads to the core of our investment philosophy: When there is a fundamental breakthrough on the supply side—for example, when AI introduces a previously non-existent capability—you simply cannot frame it using existing market sizes. Traditional "market sizing" becomes invalid at that moment. From a VC perspective, the classic investment triangle is team, product, and market. If you put enormous effort into a small market, the result is still small. But this prediction is based on the market's existing dynamics. If the supply side undergoes a fundamental breakthrough, introducing a capability that didn't exist before, you cannot accurately model the market size.

Marc Andreessen: The classic mistake we've made is thinking Uber's market was just the taxi market, or that the GPU market was just for gamers. It turns out, when supply undergoes a qualitative change, it creates a new market 10x, 100x, or even 1000x larger than the original. This will be the dominant investment trend for the next 30 years.

Ben Horowitz: We believe we are reinventing the computer, and this new computer is far more powerful than the one from the past 50 years. Internally, we say every day: there is no problem AI cannot solve. The way humans do everything will be reshaped. We recently raised over $15 billion, but this is just the beginning because there is so much to build.

Marc Andreessen: I often have these "out-of-body experiences" about AI now. When I'm pondering a complex problem, I suddenly realize, why don't I just ask the AI? It will not only give me 18 steps but might even interrogate me back, digging into my thoughts. The old-era computer would just stare at you blankly.

From Inventor to CEO: The 'Slingshot' You Need

Packy McCormick: In the AI era, products seem to achieve success faster. In this context, does a16z's platform functions like market access and policy become even more meaningful?

Ben Horowitz: We are constantly thinking about how to help founders not only succeed but build their companies the way they want. One key is helping them transition from "inventor" to a competent "CEO."

Ben Horowitz: This is essentially a "confidence game." When an inventor doesn't know how to manage an organization, they receive all sorts of advice, and their confidence can enter a vicious cycle. Our entire firm is built to place founders into a virtuous confidence cycle. When you can access important CEOs, hire top engineers, or connect with key government figures, your confidence builds. With confidence, you make decisions faster and build the company more effectively. The whole firm exists to enable the inventor to become the CEO and run their own company, connected to anyone through the network.

Marc Andreessen: Let me add to that. These super-genius founders have typically spent 10 to 20 years in a lab staring at a screen. They have the capacity to understand the world but haven't done it yet. There's a misconception that if the product is good enough, the world will naturally adopt it. But the real world is vast and messy and may not always be friendly to new ideas.

Marc Andreessen: There are 8 billion people in the world, and their opinions don't necessarily align with yours. Many people have a real "vote" on your product and company. The real world is not necessarily friendly to new ideas; it might even want to reject them. Building a company around the product and the founder is an art and a science.

Marc Andreessen: Our job is to help founders navigate this. As an inventor, you need a kind of "power-up." We want startups to be able to leverage our brand, network, and expertise to become incredibly powerful very quickly. The purpose of building this dominant VC brand is so that our portfolio companies, at their most critical moments, can use our influence in the world like a "slingshot."

Compounding Reputation: The Intangible Moat

Packy McCormick: You never publicly attack other technologies, founders, or companies. How do you manage that?

Ben Horowitz: This rule is written in black and white in our culture manual, and everyone who joins the firm must sign it. Our core is: we are dream builders, not dream killers. Anyone trying to move the world forward, whether we agree with their methods or not, we support. We always bet on the future.

Packy McCormick: If you had to choose one thing to compound, what would it be?

Ben Horowitz: Reputation. We've discussed this since the firm's founding day. Sometimes the investment is huge and the payoff is slow, but this is our core competency. We want everyone in the global tech world to think we are the best partners.

Marc Andreessen: And then that reputation transfers to the companies we invest in. When a company takes our investment, they can leverage our reputation to acquire customers, recruit talent, attract follow-on investors, and deal with regulators.

Ben Horowitz: You have to be very vigilant about maintaining reputation. The damage caused by one employee's rude behavior takes 5 to 10 correct actions to repair. But once reputation is built, it is the most powerful force. It took us 6 months and countless meetings to raise our first fund ($300 million). This time, raising over $15 billion, Marc and I each did one AMA and it was done. That's the power of reputation.

Packy McCormick: When facing attacks, do you find any enjoyment in it?

Ben Horowitz: Sometimes it's emotional because you know the intentions of these builders. Many attacks are often character assassinations. Now everyone thinks Marc is Jewish so they can attack him more.

Marc Andreessen: My new name in certain political circles is "Andy Horowitz" because people think Andreessen sounds Jewish.

Zoomers Will Save the World

Packy McCormick: Marc, you once said venture capital would be the last job on Earth. If that vision is true, what will a16z look like a century from now?

Marc Andreessen: I think I was taken out of context. I was talking about a recurring historical pattern: a person with a dream, operating in an area with asymmetric returns (high risk, high reward). Traditional banks or large corporations won't fund such projects.

But if you can build a portfolio of these dreams, the expected value is very high. Imagine Columbus's fundraising pitch: I have a 60% chance of not coming back, and my initial idea will ultimately be proven completely wrong. This intuition for betting on uncertain outcomes has always existed throughout history and will only become more important.

Ben Horowitz: Exactly. There used to be a saying in software development called the "mythical man-month," meaning adding more people actually slows things down. But now it's different; you can throw money at problems. Elon invested heavily in foundational models, and the catch-up speed is incredibly fast, something that wouldn't have happened before.

Packy McCormick: If you were to train an AI to find the next "Columbus," what traits would you have it look for?

Ben Horowitz: First, independent thinking. They don't read the room; they have original ideas. Second, a certain degree of personal charisma that makes people want to follow them. Beyond that, great entrepreneurs come in all shapes and sizes.

Packy McCormick: Finally, what are you most excited about for the future?

Ben Horowitz: It's like the invention of the steam engine or electricity; we are entering a better world. All those mundane tasks that consume our lives will no longer be necessary, and the quality of life will be much better.

Marc Andreessen: I am fanatical about Gen Z (Zoomers). They are the "AI-native generation," the recipients of this strange period from 2015 to 2024, and they are no longer tolerating it. They don't carry moral guilt or feel they need to apologize for wanting to succeed. They've watched thousands of hours of tech videos online and are more knowledgeable than previous generations of founders. They are driven and uncompromising. I would spend 100% of my time with Gen Z.