A16z co-founder: The physical laws of the old world are dead; encryption has become the key infrastructure for AI.

Source: a16z

Compilation: Felix, PANews

At the a16z Fintech Connect Conference, Alex Rampell and a16z co-founder and general partner Ben Horowitz discussed how AI is rewriting the fundamental rules of software competition, why crypto infrastructure is crucial in an AI-dominated world, and the future trends of venture capital.

PANews has summarized the highlights of the conversation.

Host: You have been deeply involved in this industry for many years. I want to start with your previous book about the tough situations CEOs face, during which you experienced market crashes and company transformations at LoudCloud and Opsware. Today’s startups are mostly “AI-first,” but for CEOs of traditional companies with 5 to 10 years of history, in an era before AI took off, how should they respond when even the financial markets are no longer optimistic about them?

Ben Horowitz: I believe that in the face of huge industry upheavals, you must recognize that some of the most basic “physical laws” have changed. Compared to how we built tech companies in the past, AI introduces two fundamental differences: First, everyone used to know that “throwing money at the problem doesn’t solve software issues” (for example, being two years behind couldn’t be caught up by hiring a thousand engineers in a month), but that’s no longer true. As long as you have enough money and good data, you can buy enough GPUs to solve almost any software problem. Second, the software industry once believed that “market dominance grants 90% of the power” (such as data lock-in, migration costs, UI lock-in), but these moats are now largely gone. Code is easy to copy, data is easy to transfer, and in the future, interactions won’t even be between humans and software, but between AI and users, with AI’s use of user interfaces being extremely flexible.

As a CEO, you must first realize that these advantages are disappearing. So, where does your value lie now? What do you offer? It turns out many things still hold value. But if you try to maintain good pricing based on old advantages, you’ll face enormous pressure. Your pricing must be based on other, more unique value you provide.

Host: We often discuss internally that in the past, having a good product might give you a decade or at least five years of growth; but now, it might only be five weeks. We also talked about IPO issues. Companies are staying private much longer now. If you’re facing a life-or-death crisis, being a private company is definitely easier to handle than being public. But the “end of days” in SaaS (Software as a Service) is also because people are doubting the ultimate value of companies. Everyone starts a business to create economic value and aims for financial gains. But if you wait too long, your company might become worthless. That’s terrifying. If you’re at risk of being disrupted, what different approaches should CEOs take now?

Ben Horowitz: Yes, I think you must be honest about what you truly own. Some companies are naturally淘汰, others are not. If you push ideas to their logical extreme, you might feel everything is worthless, because if there are no employees left in the company, who will buy your lousy software? The evolution of these things is often more subtle than we imagine and takes longer. So the question is, during this transition, are you getting stronger or weaker? If no one is buying your product, and customer funds are shifting elsewhere, you face a huge problem—you might need to lay off a lot of staff and pivot.

On the other hand, some companies are ruthlessly slaughtered in valuation but remain strong in their core business. For example, Navan, where I serve on the board, is in the travel business. Clearly, under the “SaaS end of days” narrative, people think they’re doomed, that there’s no future in travel. But a closer look shows the situation is much more complex. In travel, you need a clear network of relationships. If your company is sizable and needs global travel, you must establish relationships with every airline, hotel, and train operator worldwide. You have to manage all these relationships and connect your systems back to your clients’ budgeting systems, etc.

Furthermore, giants like OpenAI or Anthropic don’t want to sell products directly to travel managers. No one has channels to reach travel managers directly, and it’s hard to imagine that being a good idea. So, you want to keep moving forward, doing what companies like Intuit are doing—transforming into more AI-oriented companies and retaining customers. By the way, AI-driven travel experiences are actually much more complex than people think. I don’t know if this situation will last, but that’s the current reality. So, I believe it depends on the specific company. Not all companies are the same, but I do think this is a “brave new world”; if you look at it with old-world eyes, thinking it’s still governed by the same physical laws, you’re doomed.

Host: The boundaries are now blurred. Creating a feature is easy, but a feature isn’t the same as a product or a company. Since you founded your venture capital firm during the 2009 global financial crisis, the world has changed dramatically. How does today’s VC landscape differ from back then?

Ben Horowitz: It’s very different. Our first fund was $300 million, raised from all traditional LPs, endowments, foundations, fund of funds, etc. Now, we’ve just raised $15 billion across four of our seven funds. And we’ve raised from different types of investors. Initially, we had almost no international investors; now, about 35% of our capital comes from around the world. Technology has become increasingly important. I think we have to think about this world in ways we never did before.

The reason we raised so much capital is that the U.S. must rebuild its entire infrastructure immediately: we lack enough rare earth minerals, sufficient power resources, and manufacturing capacity. The existing chips are extremely power-hungry and were originally designed for gaming. Investing in the future world requires enormous funds—this is entirely new and very critical because America’s power resources are nearly exhausted, while China’s capacity graph is growing vertically. We even invested in a physical transformer company because we need more efficient, easier-to-manufacture power transformers. Since we invented electricity, transformers haven’t really changed.

Host: There’s an old saying that the cure for high prices is high prices themselves, but the problem is there’s a significant delay. So now, some computers arrive without RAM. When you buy servers from Dell, they say no RAM available because it’s sold out. You can build new factories, but that takes five years. In 1999, we were building fiber optics, but most of it was unused at the time. Now, all GPUs are running at full capacity. How do we solve these bottlenecks?

Ben Horowitz: Yes, there were bottlenecks when laying fiber optics too, just in different places. Back then, server transmission speeds weren’t even enough for video streaming. There were no load balancers, no application servers, nothing. So, you had fiber and bandwidth but couldn’t build applications, and most end users weren’t connected to the network. The system couldn’t operate, and that’s when the dot-com bubble burst. But now, almost every part of the chain is a bottleneck.

I think what might happen in the future is that we’ll have enough chips before we have enough power resources. Nvidia will produce enough chips, but then we’ll face shortages of memory and power. So, you need to carefully analyze where we stand in each part of the supply chain and find ways to alleviate bottlenecks. By the way, God bless Elon Musk’s “Terrafab” plan. That’s his idea—he wants to personally solve all bottleneck issues. That’s just how he does things, and that’s why we need him.

Host: Absolutely. You are an expert in AI and crypto. The most frightening thing now is that anyone can generate highly personalized scam emails or calls using Claude or ChatGPT, making all communication channels seem unusable. Do you agree?

Ben Horowitz: 100% agree.

Host: Because I often receive emails addressed to “Dear Allen from Index Ventures,” with the wrong name, I can just delete them and feel relieved that they got the name wrong. But what if anyone can perfectly personalize and forge content? The best way to view email inboxes now is that they are open permission-based to the public, like a to-do list. What should we do? This goes back to cryptography, which is why I mentioned hashes—they were originally designed to prevent spam. Do you see overlaps between AI and cryptocurrencies?

Ben Horowitz: Yes, I think everything starts with the problems AI creates. One night, I woke up thinking about how someone could use AI on Zoom to impersonate me, causing my finance team to wire $500 million to Nigeria—that would be a huge problem. We face several core issues: First, on social media, dating apps, or Zoom meetings, how do you prove you’re a real human and not a bot? Second, I often receive videos from family members that are convincingly AI-forged, and in the future, even AI might not be able to distinguish what’s AI-generated. So, we need cryptographically strong signatures to verify authenticity (e.g., this is really someone’s speech, not a forgery). For such provenance, you shouldn’t trust Google, Meta, or the U.S. government, but rather the mathematical game theory of blockchain. Third, if universal basic income (UBI) is to be implemented, government disbursements are highly inefficient (about $450 billion was stolen during pandemic stimulus), so everyone needs a secure address to receive funds—this is a crypto issue. Lastly, how can AI become an economic agent (like a business making money, receiving payments)? They need internet currency (cryptocurrency) as infrastructure. Therefore, AI has spawned many opportunities in the crypto space.

Host: So, what is the future of venture capital? Some say white-collar jobs will disappear, leaving only venture capital as a profession, because investing is about betting on the non-deterministic nature of entrepreneurs, and human relationships might survive in the AI era. Considering potential changes in white-collar work over the next 5 to 10 years, what do you think the VC world will look like?

Ben Horowitz: That’s hard to predict. Looking back at the industrial revolution, the risk investors in railroads and automobiles eventually became banks like JPMorgan and Goldman Sachs. There are many scenarios: one is that a few giant companies monopolize everything, and VC firms move upstream; another is that after AI labs reach their limits, they are nationalized and become public utilities like electricity, upon which everyone builds—this would be a very different VC world. Power shortages could also lead to big companies controlling all GPUs, or it could push computing toward edge devices like smartphones running small models. Future VC might become even larger and more exciting as everyone becomes an entrepreneur, or, like late-stage industrial revolution, it might become harder to start new companies.

Host: With 8 billion people now able to turn their ideas into reality—whether coding, composing music, or making movies—this is truly exciting. How can we make this technological shift less frightening and counter the dystopian narratives?

Ben Horowitz: From a macro perspective, the history of technology is a story of things getting better. But transitions are always scary. For example, in 1750, about 93-94% of Americans were farmers. If you told a farmer back then what a “product marketing manager” was, they’d think it’s the stupidest, most unbelievable job—because they weren’t producing food or building houses. Most people find it hard to see the other side of transformation. During the Great Depression, economist Keynes predicted that once material wealth and basic needs (housing, food) were met, people would only need to work 15 hours a week. But he didn’t realize how incredible human capacity for creating new needs is—we now want cars, TVs, computers, and even gourmet meals prepared by chefs for 10 hours. These needs didn’t exist then. In the next 15 years, I believe that in the U.S. and worldwide, everyone will enjoy a quality of life, luxury, and access to information better than the best lives in 1980. So, although the transition is unsettling, we should not be angry about it.