Investment tycoon Bill Gurley: The AI wave interrupts the market correction that should have occurred, and China's intense competitive environment can instead shape stronger companies

On June 19, the Invest Like the Best podcast invited renowned venture capital giant Bill Gurley to discuss the current issues in the U.S. primary market, as well as the valuations and opportunities for AI companies.

In the interview, Bill analyzed the current state of the venture capital industry, stating that super venture capital funds continue to rise, with many investment scales increasing tenfold, which has given rise to some "zombie unicorns" that are large in scale but whose real value is questionable. Currently, both GPs and LPs lack the motivation to correct valuations.

Bill pointed out that the zero-interest-rate environment has led to an abundance of capital, allowing companies to easily raise funds without the motivation to go public, resulting in stagnation in the U.S. IPO and M&A markets over the past few years. The liquidity issues faced by LPs are also worth noting; while doing some non-consensus but accurate things can yield significant profits, problems arise when everyone allocates 50% of their assets to illiquid assets.

Regarding AI development, Bill believes that the arrival of the AI wave has interrupted the market corrections that should have occurred, with AI being viewed as a historic platform transformation that has sparked a new wave of investment enthusiasm and valuation bubbles. However, he noted that most AI companies' revenues currently come from computing power resale, and economic benefits ultimately need to be addressed.

He also mentioned that China's intense competitive environment can actually foster stronger companies; if China's tech giants open-source their AI models, it would be incredibly powerful. With open models, they can train each other, and everyone can use them, which he believes will lead to a plethora of choices and experiments that won't happen in the U.S.

Here are the highlights from the interview:

1. Nowadays, many well-known funds have increased their investments from $500 million every three to four years to $5 billion, which is a tenfold increase. They are very actively participating in what we call the late-stage, which is a euphemism for large checks.

2. The emergence of LM is a watershed moment; everyone is excited about this new platform transformation. There are about a thousand private companies that have raised over $1 billion, but what their real value is remains questionable.

3. Managers of VC groups in large endowment funds lack the motivation to attempt to adjust valuations; founders have often multiplied their ownership stakes by the company's historical peak valuation and then used that number as their net worth.

4. The zero-interest-rate period has delayed any VC corrections while simultaneously creating a lot of capital and speculative behavior. When companies raise so much money, those that should have been eliminated earlier can survive, leading to a large number of "zombie unicorns."

5. I believe the market had a small correction around 2022-2023, but with the emergence of AI, everyone is excited, and investment enthusiasm is high, with ordinary companies achieving valuations and revenue multiples of 10x or 20x.

6. The IPO and M&A markets have stagnated over the past few years; successful companies now have no need to go public, or at least they don't have to rush to do so. If they can become trillion-dollar private companies, do they still need to go public?

7. Among companies with revenues exceeding $100 million, 87% are now private enterprises, which is quite striking. Taking Stripe or other companies as examples, that's just one company, but the real concern is the 1,500 companies that cannot all perform like Stripe

8. Many companies in the market are actually repackaging basic models and cloud services for resale. Many companies are actually operating at a negative gross margin. Buying products from these companies may be cheaper than directly purchasing models or cloud services, and these revenues are counted multiple times, resulting in a negative gross margin. Until we truly care about unit economics, this does not matter during the all-in phase of the capital war, where everyone is just fighting for market share.

9. In the liquidity issues of LPs, time is a significant problem; capital costs and internal rate of return (IRR) continuously erode profits. Everyone likes to say that the important metric is distributed profits multiple (DPI), not internal rate of return (IRR). However, if time doubles, the internal rate of return (IRR) becomes truly important. In addition to time and capital costs, there is also the issue of equity dilution, as each zombie unicorn has to issue 3-6% equity incentives to employees every year. When these two factors are combined, it becomes a real problem.

10. Theoretically, Google should be in the best position, dominating every AI use case. However, basically no one I know uses Gemini or Google for collaborative generation or as tools for daily use. Instead, they are using startups, Cursor, Anthropic, and OpenAI. Even if large companies act quickly, the tech companies themselves are still repeating the same phenomenon.

11. In the first two years of the internet, all startups were building on Sun and Oracle, but five or six years later, no one was doing that anymore, so it is very important to pay attention to this shift.

12. Alibaba has open-sourced Qwen, Xiaomi now has its own model, and Baidu's Robin Li originally had a closed-source model, but he said it would be open-sourced in June. If this level of fierce competition ultimately gives rise to open-source products from four well-funded companies, it will be incredibly powerful. These models can train each other, help each other, and improve. So if there are four open-source models that can train each other and everyone can use, I believe this will bring a lot of choices and experimentation, which is something that won't happen in the U.S.

13. If the market corrects, everyone will look for new opportunities. Everyone understands the effects of compound interest, network effects, and cyclical cycles, and has seen both prosperity and recession. The stock market decline at the beginning of COVID-19 lasted about three weeks, and then everyone started bottom-fishing. Therefore, confidence in the AI field is high enough that even if people feel AI is overvalued for six months, it will quickly rebound.

14. If AI can achieve more natural interactions and more personalized experiences in the consumer sector, then the consumer market may once again become a hotspot for venture capital. If four or five AI companies achieve breakthroughs in the consumer sector in the future, I would not be surprised. This may be the reverse-thinking opportunity we have been looking for, as most resources are currently concentrated on the enterprise side.

15. Yale University has announced that it is seeking to sell $6 billion in private equity on the market. Yale University is the institution doing this, which is very important and interesting. From a historical perspective, no institution has had a greater impact on endowment fund management strategies than Yale University, under the leadership of David Swensen (former CIO of Yale University) He is the pioneer of this model.

Below is the complete transcript of the interview:

Patrick: Our guest today is Bill Gurley. Bill was a general partner at Benchmark Capital. This is his sixth appearance on Invest Like the Best, and it is also his most comprehensive market analysis, discussing the realities reshaping the venture capital industry. Bill confronts the unsettling mathematical issues behind today's venture capital returns, especially the phenomenon of companies remaining private for longer periods. He also explains why there is insufficient motivation for anyone—from GPs to LPs to founders—to accurately mark assets, leading to a coordination problem for the entire system. We also delve into the investment implications of AI as a platform transformation, from assessing the quality of AI revenue to international competitive dynamics. Bill provides key perspectives on how to navigate the current and future landscape. Please watch my conversation with Bill Gurley.

So, Bill, this marks your reclaiming of the title of "most frequent guest," beating our good friend Michael Moritz. Welcome back.

Bill: I can't think of anyone more suited to ride alongside me than you.

Patrick: Interestingly, this is the first time since 2019 that you and I have done a show together alone, which is hard to believe. Time flies. Since it's just us, I want to keep the topic broad and start with your views on the current situation. I know this is something you often do at Benchmark, like delivering a "State of the Market" address. I hope you can do that for us and talk about the market in the summer of 2025.

Bill: I'm excited to do this. Yes, I used to often start our LP meetings with an overview of the VC landscape. So this is a process and presentation I'm accustomed to. Recently, I've noticed many different things happening in the world. The venture capital space may be undergoing permanent changes. Many of my talks have been based on what seems to be inherent cyclicality in the venture capital industry, but this cyclicality has recently been somewhat disrupted or overturned, becoming a bit chaotic, and we'll discuss that later.

Before we dive deeper into this issue, I want to make two preliminary points. First, Michael would agree that I am a staunch fan of systems-level thinking. There is a book about how to think systemically, and it's really cool. My time with Michael at the Santa Fe Institute was largely related to this theory that the behavior of a system is different from that of its individual components. It's not easy to see the connections between systems; it's a challenging task. But as we dig deeper, I believe many components of this industry are colliding with each other, and the combined effects of all these factors are the most interesting. So you have to step back a bit and look at the big picture from a distance.

The second point I want to clarify in advance is that I will not judge the participants in these activities. Just like some people and companies have taken actions that changed the state of this field, I believe their actions are all rational and in their own best interests. The negative effects may not have a positive impact on the world, but I don't see it that way. I'm not accusing anyone of having malicious intent. I want to make this clear first If you allow me, I will now begin to delve into some market realities that I have observed. In the first part, I do not want to overanalyze; I just want to emphasize a few things. If you are in the VC market, by the way, I believe what we are discussing is important for VCs, founders, LPs, and anyone involved in this ecosystem. This is very high-level stuff. So let me talk about the realities, and then we can exchange interpretations on some of them.

Current State of the U.S. Primary Market

1. The Rise of Super Venture Capital Funds

So the first thing I want to bring up, which everyone is discussing, is the continued rise of super venture capital funds.

In the beginning, you know, everything was customized. Most well-known funds focused on early-stage investments. They did not participate in late-stage rounds, and the fund sizes were not large compared to today. Now, many well-known funds have increased their investments from $500 million every three to four years to $5 billion, which is a tenfold increase. They are very actively participating in what we call the late stage. Although I have always felt that late stage is a euphemism for large sums of money, you know, because if someone is willing to invest $300 million in an AI company that has only been established for 12 months, right? So that’s not late stage. That’s just a big check.

Many companies have shifted to the upstream market, and they have also created different sector-specific funds, etc., all of which have led to a significant increase in capital managed by many brands. Then there are many new players entering the late-stage market in different ways. Some established institutions occasionally participate, such as Fidelity and Capital Group. But I think Atreides, Coatue, Altimeter, and Thrive are very active, and I feel they are doing something distinctive in the market. Also, Masa (Son Masayoshi) is back. We haven’t heard much from him in recent years, but he is now active in the market again.

Patrick: He is a metric in himself.

Bill: Yes, I agree with that. So there is more money in the market now.

1. Zombie Unicorns

The second reality that people talk about, which is somewhat surprising, is the "zombie unicorns." I don’t particularly like this term, but it is the most commonly used. If you look at the number of those companies, I like to use the emergence of LLMs as a dividing line because it really is a watershed moment, and everyone is excited about this new platform transformation. There are approximately a thousand such private companies that have raised over $1 billion. ChatGPT tells me it’s 1,250, while NVCA (National Venture Capital Association) says it’s 900. Let’s just say it’s about a thousand.

Bill: It is said that each of them raised between $200 million and $300 million, you know, adding those up gives you $300 billion. NVCA estimates that there are $30 trillion in assets on LPs' books. I have had one-on-one communications with LPs, and their participation and investment scale have slowly increased from 5%-7% to now 10% to 15% Some even reach half of their private equity allocation. So we see that some private equity firms have much larger PE scales, but venture capital is increasingly significant on the balance sheet. So this is important. I think there are many questions about a group of companies. First, what is their true value? You know, their last round of pricing was still in 2020.

Patrick: Around 2021, right?

Bill: Yes, that was when the market really peaked, which was the second year of the COVID-19 pandemic. If you remember, all the tech stocks surged at that moment, and Zoom also skyrocketed at that time. During that period, everyone performed exceptionally well. So the question is, what is their value geometrically? The investment community seems uninterested in this type of company, and their overall growth rate is not high. I want to talk about why I think this way.

Most people may not believe it, but I assure you it is true; no one has the incentive to clarify valuations. So if you don't understand this world, you know, private equity investment, whether PE or VC, is a strange model where GPs (the people responsible for investments) quote LPs and set their own prices.

Now there are auditors messing around behind the scenes, and you will hear LP complaints. Some companies are more conservative and price lower, while others price higher. So they receive mixed signals.

1. Distortion of Corporate Valuation

Patrick: The same asset has different prices from different GPs, right?

Bill: Yes, but people may not realize that the managers of VC groups in large endowment funds have no incentive to try to adjust this number. In fact, many of them are based on book valuations. So if there is any difference, they even have a reverse incentive to correct it.

Patrick: But don't founders have the incentive to get these things right? Isn't it better for the company in the long run not to create a farce?

Bill: That's a great question. I think there are two things that can counter this. First, every founder I know has multiplied their ownership stake by the company's historical peak valuation and then taken that number as their net worth. I think that's incorrect.

Patrick: But who cares? I mean, it doesn't mean anything.

Bill: Let me say again, I don't mean to judge; I think it's natural for you to do so. But reducing that number by 70% is still a difficult task.

Then, another issue, frankly, is liquidation preference. This is another technical detail, and let me explain it to the audience. The total amount of financing directly affects the importance of liquidation preference. During mergers and acquisitions, investors can choose liquidation preference to get back their principal instead of converting to common stock. So if a company raises $300 million with a valuation of $2 billion, liquidation preference has little impact. If the valuation drops to $40 million, liquidation preference could account for 75% of the company's valuation. You know, this is a real issue for people Patrick: If we go back to the list of those thousand zombie unicorn companies and delve into this area, you will find out how many companies are profitable. Therefore, this question can go on forever until they are willing to reprice, rather than being companies that will eventually go bankrupt at some point and have to finance and reset their prices.

Bill: Well, I have to admit I haven't done a statistically significant survey, but it might be interesting if someone did this survey. Maybe someone in a fund of funds, or Pitchbook, or other relevant institutions can provide this data. I will tell you, this is to lead into what I think is happening. We are in a very long period of zero interest rates, now called ZIRP, which is unprecedented in 100 years, with zero interest rates lasting for five, six, or seven years.

Patrick: A long time.

Bill: On one hand, it has delayed any VC corrections, and on the other hand, it has created a lot of capital and speculative behavior. The funniest thing is that I received an invitation. I have only seen Mr. Buffett once in my life, at a small fundraising event with 20 people, where each person could only ask one question. I told Warren, you know, if interest rates are zero, your DCF doesn't work. He replied, "You're right." And that was it, a brief encounter with a great man.

In short, speculation is rampant. The amount I mentioned, 20-30 billion dollars, is unprecedented before that.

When companies raise so much money, something happens. I think too many participants enter a single field, and companies that should have been eliminated earlier can survive. This makes market expansion more difficult because the number of surviving companies goes from 1-2 to 3-5. When you overfinance, you can do anything; there are many articles and studies that show that constraints bring creativity, and it's better to focus on one or two main products. But when there's too much money, you end up doing seven projects.

I think we had a small correction around 2022-2023. But this was still before AI took off, and most people were working towards what you mentioned as the breakeven point. So once you shift to breakeven, you cut those seven projects down to two. But those seven projects and the over-expanded sales teams brought in revenue, but the revenue is not sustainable. So when you cut expenses and move towards breakeven, your growth rate will naturally be affected, and I think that's the reason for the low growth.

I agree with your point that many companies have enough funding to achieve breakeven or come close to breakeven. Based on all my previous discussions about the beautiful nature of building traditional companies, you might think this is a good thing. Of course, I support that. But there is a potential reality that they might really be able to exist indefinitely, which is where the "zombie" label comes from.

Patrick: Yeah, what does it all mean? Since no one has the motivation to correct valuations, will it just stay this way?

Bill: Let's revisit this question. Let me continue, I want to clarify these market realities first, and then we can delve into the possible situations 4. Exit Window Closure

Patrick: The next question is about exits, specifically how these companies will be priced in the real market.

Bill: Exactly, we have super funds, zombie unicorns, and the capital markets. For some indescribable reasons, the IPO and M&A markets have stagnated in recent years. In 2021, both were actually quite good, but now things have come to a standstill. I think if you look back at last year, which is 2024, you'll find this very important. That year, the Nasdaq index rose by 30%, but the window remained closed, which seems to be the general perception outside. In my history of watching the capital markets or engaging in venture capital, I've never seen a situation where the Nasdaq market performed well but the exit window was closed.

Patrick: No IPOs, right.

Bill: Yes, it doesn't make sense. In the past, these two were correlated, so something else must be happening now. I've been paying close attention to IPO discounts, especially the discounts imposed by well-known large firms on the market. But some believe that the cost of going public is too high, and others think that the cost of being a public company is too high. Of course, money is everywhere. We'll come back to this topic later—successful companies today have no need to go public, or at least they don't have to rush to go public.

M&A is even harder to explain. Everyone blames Lina Khan (Commissioner of the Federal Trade Commission), but she has left, and there have been no record-breaking M&A deals in the first five months of this year. I think this may be related to the "Seven Giants." These seven companies have an absurd amount of cash, which should lead to large-scale M&A, and I believe they would be happy to use that money. But Washington doesn't like it, and the EU is even less keen on them being active, so the situation is stuck. No one wants to take the risk of an M&A deal that might not go through. Even large deals like Wiz, once announced, say it will take over a year to complete. It's hard for boards and management teams to wait a year; it's too difficult.

Patrick: Do you think we will soon see a private company valued at $1 trillion?

Bill: How far is SpaceX from that goal?

Patrick: About a third, I guess. OpenAI is also a third. Stripe is a tenth. There are several others that could potentially achieve this if they can maintain their success. What I'm saying is, if you can become a $1 trillion private company, do you still need to go public?

Bill: It's a bit crazy. We'll talk about this. The last factor that might affect M&A is overvaluation, just like what we did in 2021; we are still pushing the most exciting companies to historical highs today, which will also affect M&A.

Patrick: In your view, why does this situation continue to occur? Is it the feedback loop we just discussed?

Bill: I think ZIRP (Zero Interest Rate Policy) was the main reason before LLM. After LLM, everyone believes that AI is the biggest technological platform shift in our lifetime. So if you believe that... One more thing, I recall thirty years ago when I was with Mauboussin at First Boston, the network effect and compounding effect were not fully understood or recognized Now everyone is completely convinced, so those who have seen the valuations of Google or Meta rise from $12 billion to $3 trillion will feel that if a certain company could reach such heights, they cannot miss such an opportunity—it's reasonable for independent investors to think this way. If everyone thinks this way, the market will factor expectations into prices, but we shall wait and see.

5. LPs face liquidity issues

Next, we need to talk about how many LPs are facing liquidity issues. This is a new phenomenon related to the closing of IPOs and M&A windows. There is also a unique piece of data: in the first quarter of 2025, U.S. universities issued $12 billion in bonds, the third highest quarterly amount in history. If you are using debt to fund capital commitments, it is because your endowment funds do not have enough liquidity to cover the 3% or 5% expenditures annually as they used to.

Recently, you may have seen Harvard announce the sale of $1 billion in private equity assets in the secondary market. More interestingly, Yale University announced that they are seeking to sell $6 billion in private equity on the market. It is very important and interesting that Yale University is the institution doing this. Historically, no institution has had a greater impact on endowment fund management strategies than Yale University. David Swensen (former CIO of Yale University) was the pioneer of this model.

Patrick: He is the godfather of this model.

Bill: Yes, without a doubt. Under David Swensen's management at Yale, the annual average return over 35 years reached 13%. He is known for the Yale model, which invests much more in illiquid assets than in liquid assets. The reason no one did this initially was due to a lack of transparency, a lack of liquidity, and the difficulty of managing these assets. But he did it and succeeded. What I want to say is that what we are seeing now may be the result of everyone imitating the Yale model. You know, Howard Marks has a famous saying: when you do something that is non-consensus but accurate, you can make a lot of money. But what if everyone imitates David Swensen? If everyone puts 50% into illiquid assets, can they still succeed? I think this is a very challenging question, but that may be the reality. And the Yale University that led us to adopt this strategy is now trying to exit, which I find very interesting.

Patrick: If you consider the liquidity issues of LPs, could this be the key to breaking the deadlock you just described?

Bill: It’s possible. If I can overcome these real issues.

6. The AI wave interrupted the market correction

The AI wave came at a very opportune time, which is the fifth point of the six realities I mentioned. We were heading towards a small correction. You remember, Patrick, at that time everyone was tightening their belts, laying off employees, pursuing breakeven, and worrying about whether they could refinance.

In my thirty years of venture capital experience, every time the industry overheats, there is a correction, and then everything calms down. I have seen Morgan and Goldman Sachs open and close offices on Sand Hill Road, and I have seen Fortune and Forbes focus on Silicon Valley and then withdraw I have seen it several times.

But this time there is no complete correction because AI has emerged, and everyone is very excited. I'm not saying we shouldn't be excited; if this is indeed the biggest technological platform transformation of our lifetime, then we must be excited, as it will impact the zombie unicorns and everything else.

But suddenly, investment enthusiasm is soaring. What are the multiples for AI company valuations and revenues? My goodness, a regular company is 10 times, 20 times, right?

Patrick: About that, some are even higher.

Bill: Yes. Although traditional LPs are facing funding constraints, they are still able to find money elsewhere. The Middle East is a major source of funding. How many friends have you had go to the Middle East in the past 12 months? Quite a few, right? They are all talking to fundraisers, so the money is there, and everyone is chasing this opportunity; no one wants to miss out. This is a very important component of the whole situation.

1. Companies are more willing to stay private

Bill: The last reality you mentioned is that there is a new trend in the late-stage market. I think Thrive's Josh and his team have taken the lead in this, though they are not the only ones; they have been at the forefront.

They will look for companies that were originally planning to go public, and the media has reported they are going public, and make an offer that is hard to refuse. Founders cashing out will encourage employees to cash out, and may also encourage angel investors to cash out. And basically, they will persuade your company to stay private. A recent example is Databricks. Patrick and John from Stripe have also mentioned this in different podcasts and talks; at first, they said, "Maybe we will go public, but not in a hurry," and later it became more like what you said, "Maybe we will never go public." I have talked to some LPs I have interacted with, and this is quite unusual. They have traded Stripe shares, and the company is relatively accepting of this. This is very novel and unique in our world.

Patrick: These companies can get the funding they need, whether it's employees cashing out or early investors selling shares, basically like a "reservation-based public market"?

Bill: Yes, it's like the old-fashioned pink sheet market trading, done by reservation.

Patrick: Stripe is undoubtedly a great company led by outstanding founders. If you can have your own private market, why take on the extra work, regulation, data disclosure, and let competitors know your situation? This makes sense for everyone, so I wonder if this model will continue. If LPs can gain liquidity by transferring Stripe shares, then the liquidity issue is also resolved?

Bill: We are about to delve into all of this.

I want to add one point—these investors who encourage companies to stay private have another motivation. In traditional IPOs, banks are very cautious about allocating shares. If a large public or private fund applies for allocation, it is usually oversubscribed by 100 times, hoping to get 1-2% of the allocation rights; they cannot get 30%. But when these investors go for large private financing, they can get 30% of the shares, which is much more than in an IPO. This way, they can obtain a higher equity stake than through the traditional IPO process This is somewhat like an oligopoly, taking the IPO growth dividends away from the public market. When Amazon went public, its market cap was less than $1 billion, and now it exceeds $1 trillion, with the public market enjoying this compound growth. If you delay going public and obtain a high percentage of shares in advance, these investors are better off than those who buy in after the company goes public.

Another important point is that if they turn around and tell LPs: the company will no longer go public as it did in the past, if you want to benefit from these high-growth tech companies, you must invest in me. This is very persuasive.

Where does the U.S. capital market need improvement?

1. High IPO costs, companies have no need to go public, tokenization may be a solution.

Patrick: You have already understood the reality of the market, and now I want to delve deeper into all these situations. My framework is that the interesting premise for me is that I have always hoped for a healthy operation of the capital market. The U.S. capital market is an extremely important innovation engine in world history, driving countless innovations.

So, in my view, a well-functioning capital market can price risks effectively. Therefore, I support anything that can achieve this. I am curious, from this perspective, considering these market realities, where do you think the system needs the most improvement, and how would you like it to change?

Bill: Yes, I agree with your wish, and I think our situation would be much better if more companies participated. One thing I didn't mention in reality, and most people know, is that the total number of publicly listed companies in the U.S. has significantly decreased from its peak, and fewer companies are going public.

I believe a large part of the reason is the IPO process and those well-known investment banks. I asked my friend Jay Ritter to reanalyze the data, and now the IPO discount is about 25% to 26%, plus a 7% fee, which brings your capital cost to 33%.

I know a CEO who is preparing to go public, and when discussing with the investment bank, the bank said you should issue at price X, and the founder said I can raise $1 billion in the private market at a price 20% higher tomorrow.

As you said, if the private market is so liquid, flexible, and optimal, why go public? I don't know what changes are needed. I feel that as long as it involves financing, everyone will avoid this part.

I know of a CEO with records showing he talked to bankers, and the bankers said, we think you should price it at X. The founder said, I can raise $1 billion at a price 20% higher tomorrow. Speaking to your point, since the private market is so flexible, liquid, and well-optimized, why go public? I don't know what changes are needed. I feel that as long as it involves financing, everyone will avoid this part.

Hester Peirce wrote an article titled "A Balancing Act of Creativity and Collaboration," which is about eight pages long and well worth reading. She was the longest-serving commissioner at the U.S. Securities and Exchange Commission (SEC). Currently, she is one of the four commissioners most supportive of cryptocurrency. The article suggests that blockchain technology may be a way to fix the IPO market, a controversial view, but I personally support it Patrick: What do you mean by "free trading of private assets"?

Bill: I mean tokenizing securities. In the future, people will no longer allocate cryptocurrencies like they do in IPOs, but will turn to distributed ledger (DL) technology. In fact, initial coin offerings (ICOs) are already operating in this way.

Indeed, this is a very interesting topic. I will continue to pay attention to it. Currently, there is a lot of regulatory pressure. We have seen some strange "workaround acquisitions" in the artificial intelligence (AI) field, such as signing licensing agreements first and then hiring people, but we haven't seen any real big cases for a long time; this seems to be a roundabout strategy.

Moreover, when asset pricing is too high, it is difficult to reach a deal. For example, some AI companies have financing rounds valued at up to $15 billion. I understand that Apple might want to acquire a company like Perplexity, but such a high valuation makes the deal difficult. Your point about the capital markets is interesting . Many people claim that we have the most sophisticated capital market functions in the world, which is envied globally, but I do not fully agree with that.

Patrick: You mentioned that the Middle East has been very proactive in this wave of technology, and they are working hard to engage with the most interesting companies, technologies, and infrastructure. Are there any other capital market innovations you find interesting?

Bill: I don't know if you would call it innovation, but a recent announcement about Coatue caught my attention. I haven't discussed it with Philippe yet, but I think they have adopted a minimum standard. In the past, their minimum subscription amount was $5 million, and now it has dropped to about $25,000, and they will work with an investment bank to handle related matters. This is similar to what I mentioned earlier about marketing to LPs, but this approach utilizes a funding pool, sometimes referred to as the "dentist and doctor funding pool," which previously couldn't invest in funds like Coatue, but now can.

I've heard that similar things are happening in private equity, where a large private equity firm is lobbying in Washington to allow 401(k) plans to invest in private equity to open up different funding sources.

Interestingly, when I tested this series, someone countered that U.S. institutional funds could be found elsewhere, but that just allows the higher-ups to earn more. It can be likened to a pipeline with imports and exports, where the exports are blocked. I can't think of a better analogy; the human digestive system might be the best comparison—eating more food doesn't solve the constipation problem.

2. Companies are burning cash, but only a few stand out

Patrick: When you talk to limited partners (LPs), you must mention specific names. What have they said to you privately? Do you think the things they haven't said are important?

Bill: I think people have become more aware of the realities of the market. They have to make decisions. For LPs, this is a long-term decision. If you work at a donation fund, you don't have much time to make decisions because your feedback cycle might be 10 or 15 years. It's difficult, but you have to start thinking about whether the changes we are discussing are temporary or permanent If they are permanent, you need to change the way you do things. For example, one LP I spoke with has gone in and out of Stripe, knows the person responsible for capital markets in the company, and has started to consider that this might be permanent, also thinking about how they need to prepare for such a world.

Patrick: Apollo released an interesting report showing that among companies with revenues exceeding $100 million, 87% are now private enterprises. Of course, if calculated by market capitalization, the proportion in the public market would be higher due to the presence of large tech companies. This phenomenon is quite exaggerated; even at the minimum revenue of $100 million, many companies have reached this level. So I want to say that we live in a world where the private equity market is very active, and this is an undeniable fact.

Bill: I want to adjust the order; I have about five analyses. I think this is a more chaotic world. The best world you mentioned is one with an efficiently operating capital market, where capital markets are efficient, going public is easy, liquidity is strong, and transaction costs are low. I do believe that such a world is better.

If we enter a new world where ordinary consumers enter the high-growth tech sector, putting their 401K and Individual Retirement Account (IRA) funds into venture capital funds that charge a 2% management fee and a 20% performance fee, I just feel that the information will be less transparent, and the transparency will be lower. Fraud will increase, and transaction costs will also rise; this is an inevitable result.

Taking Stripe or other companies as an example, this is just one company; in this example, we might mention five companies, but what we are really concerned about is 1,500 companies that cannot all perform like Stripe.

Patrick: You taught me something years ago: you have to play well on the field while also thinking about future rule changes to prepare for the future. But if we adopt this more chaotic, more private market-dependent, liquidity-reality field game approach, I’m curious how you think different groups should act, starting from the founders all the way to the entrepreneurs who truly create value, funded by these capital markets. In the AI world, if they can raise funds at a $15 billion valuation, how would you advise them to make the optimal choice under the current game rules?

Bill: They are forced to act under the rules of the field. This is also what I think is the worst part of this world. I recently came across a term called "gavagetube"; do you know what it is? The French use a gavagetube to force-feed geese to produce foie gras. In this world, the reality is that in 2021, as long as there is a bit of a trend, someone will come knocking, trying to offer them $100 million, $200 million, or $300 million.

I think for those founders who have struggled their whole lives to raise funds, this must sound like the most absurd comment, but this is the reality. I experienced this during the Uber-Lyft competition, and now there is a capital war unfolding in various fields You mentioned the concept of building traditional companies, not spending $100, but rather burning $100 million or $150 million each year, which all large AI companies are doing, and they might be consuming even more.

This is not the startup or venture capital of your grandfather's era. This is a completely different world. If you are a founder, ignore all of this and build your company the way you want. But if your competitor raises $300 million and their sales scale expands by 10 times or 50 times, you might find yourself dead before you even realize it.

You have to play a game on the field, and the good news is that because these investors are eager to give you money, you might be able to gain founder liquidity. I think this is detrimental to the long-term success potential of the company, but because it aligns with their strategy, they are all encouraging it. So I think if someone is willing to pay 30 times revenue and forces you to play a game you don't like, burning hundreds of millions each year, then you probably have no reason not to cash out a little.

I believe that if we remove all the small to medium achievements and only think about hitting home runs all day, it is harmful to the ecosystem. But if that’s how I feel, it seems we haven't learned anything from the zero interest rate policy era. All the issues we talk about that cause zombie unicorns are just being repeated in AI companies. It’s as if this might be part of the reason for not experiencing market adjustments, but the way we are funding these AI companies is exactly the same as how we funded those companies back then.

Most AI company revenues come from computing power, and economic benefits are the ultimate deciding factor

Patrick: I want to start by discussing a very important point, which is to understand your view on artificial intelligence as a new general-purpose enabling technology, which is the key difference between now and 2021. We have never seen a technological wave like this, nor have we seen companies with such revenue growth.

I know you, like me, love technology. I use this thing every day. It is the most amazing technology I have ever encountered. So I hope you can improvise a bit; I call it a bull market situation, where people are not actually being irrational because we can really achieve 5% GDP growth or other crazy numbers. Because this is indeed a different level of technology, comparable even to the internet.

Bill: First of all, I agree with your point. I would never take an opposing stance and say this is not a reasonable platform shift. If this is a platform shift, like mobile internet, the internet, or PCs, then it is big enough; it doesn’t have to be better than those.

Patrick: Even if it’s just another platform transformation?

Bill: Definitely one of them, and it might be bigger, which leads to everything we are discussing. Before we start, I won’t judge any participant. I think that’s just how it is. I have an idea in my mind, although I haven’t fully thought through all its implications, that some revenue growth is actually counted in the resale of computing power.

Many companies in the market are actually repackaging and reselling foundational models and cloud services. Many companies are actually operating at negative gross margins. The products of the companies buying these might be cheaper than directly buying the models or cloud services, and this revenue is counted three or four times, resulting in negative gross margins. **

Until we truly care about unit economic efficiency, it doesn't matter during the all-in phase of the capital war; everyone can only grab market share. Until things are resolved, the window of opportunity in front of us is wide open. I have no doubt about this. Even without discussing the foundational models, like what Bret Taylor did at Sierra, I have no doubt that AI will fundamentally change every business. I believe many things are rational responses to what is happening.

Patrick: How do you view this technology? You have experienced many such technological paradigm shifts and invested in them. What excites you most about this project, especially compared to other projects you have been involved in?

Bill: My answer is personal and relates to what you just said. I might conduct 40 or 50 searches on AI platforms every day, which is more than my searches on Google. It’s almost a very rapid way of learning, like learning details at super speed. I forget some things I didn’t know, and this happens every day. I think about how astonishing it is for those who are natural self-learners to accomplish things and get promoted at such a pace.

Then I think beyond the large models, from Tesla's FSD to other problems solved with traditional AI. These are also super interesting to me. Perhaps even more profound.

I do worry about the limitations of large language models. It’s possible to address this, but their language is as difficult to understand as a bumpy road. They are not very good with numbers. When people say that general AI will replace all computation, I don’t think they have to fix certain things or merge it. Now, when you ask an AI a math problem, it runs and writes a lot of Python code. You have to do more of that kind of work to reach that level. If you support it. But isn’t that not the real argument? I can’t refute that.

Issues Faced by GPs and LPs

Patrick: Let’s take a step further and talk about GPs (General Partners). The same question: what is the rational approach under the current rules of the game? There are two versions here, one is the "Spock-like" rational answer, and the other is the "Kirk-like" emotionally driven answer.

The Spock-like answer is: since the market is like this, I will create my own company to rationally maximize investment returns.

The Kirk-like answer is: if you were to restart a venture capital firm today, what would you do? Would you do small funds like you did at Benchmark? Or would you create a fund that invests anywhere with a different fee structure? I’d like to hear your answers from both perspectives.

Bill: I want to emphasize one of the last two points, which is that time is a big issue. We have extended the time for these companies to become profitable from 5 to 7 years to 10 to 15 years. I don’t know the exact numbers, but every LP is aware of this issue.

I think I have forwarded you this chart from NVCA, which shows the percentage of committed capital returned by venture capital funds within 5 to 10 years. In the past, it averaged about 20%. It peaked at 30%, and last year it dropped to 5%. Now it’s roughly in the 5-7% range, which is precisely the liquidity issue for LPs The reason time is a big issue is that the cost of capital and the internal rate of return (IRR) continuously erode returns. Everyone likes to say that what matters is the distributed profit multiple (DPI), not the internal rate of return (IRR). But if time doubles, the internal rate of return (IRR) becomes truly important. In addition to time and capital costs, there is also the issue of equity dilution, as each zombie unicorn has to issue 3-6% equity incentives to employees every year. When you combine these two factors, it becomes a real problem.

For example, you originally expected to receive $100 in returns from an investment in the 10th year, but now you want to push it to the 15th year. If you only consider a 10% compound interest rate, then its value after 15 years should be $160. You know, if you think about these people. If investing in a project is to achieve substantial returns, then your cost of capital is not 5%. This is the risk-free rate. 15%, then 20% each year, 15% plus 5% for equity dilution. Now, if you wait another five years, guess how much you need to replace that $100? Delaying for five years to reach the expected return for that asset class requires a return of $250.

So I think this is a big problem. I believe there are a certain number of companies that have truly achieved acquisition or IPO at some stage, and then entropy increases, and all companies will encounter difficulties in long-term growth.

People like to discuss what the fund's return rate would be if you exclude the biggest winners. But I haven't asked anyone this question: what if you keep the big winners and remove the others? Because it feels like we are heading in that direction. So I don't know, after saying so much, I really don't know the answer to your question. I have spent my entire career in the early stages, and I still love that period. I think it is a time window where you can place the biggest bets and achieve the greatest returns.

I really don't want the next generation of GPs to see every company go through a situation like Uber-Lyft. When you walk into the boardroom, you know another company has raised $1 billion, and the negotiation decision becomes whether we should maintain a negative gross margin for another two years to capture market share? You can't find this in your Harvard case studies.

It's like playing a unique deck of poker cards, and the stakes are extremely high. This is a strategic poker game, and you can't learn these strategies from the book "Good to Great." This is not the traditional way of managing a company, nor is it what Warren Buffett writes in his letters to shareholders; it doesn't apply in this world of capital warfare.

Patrick: I now want to talk about LPs and the trend of capital seeking the highest returns after risk adjustment. Generally speaking, from a rational perspective, over time, capital pools will flow around seeking the highest returns under risk prompts. That's the point. So I'm curious, what do you think are the factors that hinder this from happening? In other words, what should LPs do now? It's as if they are capital owners, representing the capital owners. On the surface, their job is to maximize risk-adjusted returns. What should they do, and what is hindering them from doing so? Bill: This might be the last point I want to emphasize. You raised a very thought-provoking question at the beginning of the podcast: Could the LP liquidity issue become a catalyst for change in this world?

There are many factors driving this change. Time is a factor; LPs are still leveraging, and there is widespread discussion in Washington about the fund donation tax, which will bring more liquidity pressure. This is something they have never experienced before, along with cuts in research funding—not just the radical cuts at Harvard, but even the normal reductions in research funding from the National Institutes of Health (NIH) and the National Science Foundation (NSF).

These cuts in indirect costs or other areas have also led universities to demand higher spending from endowments, such as 5% or 6% annually instead of 3%. These changes could put LPs in a more difficult position. Yale University may be the first to enter the secondary market, which seems exciting. If you are a small endowment that has never invested in Sequoia, now you can get a piece of the action at Yale. However, as more big players join in, the ripple effects of secondary pricing could have a chain reaction throughout the entire market.

Additionally, I think another significant issue to watch is whether the Middle East will change its mind. Qatar's Chief Investment Officer, Sheikh Salim Al-Athabi, has stated that the time for private equity is running out. He has joined the ranks of investors increasingly concerned about the valuation methods in the industry. If this viewpoint spreads and becomes a common perspective affecting other players, it could have a huge impact.

So I think this is an area worth paying attention to. If I were an LP, what would I do? I think I would definitely prepare to trade in the later private markets, observe market dynamics, and feel the market situation. I don't want to cause a bank run, but you might really need to reassess whether the Yale model is still viable. I think it was definitely workable when only Yale was using it, but I don't know if it is still viable now. I think it would be interesting to find a private equity firm that actively seeks opportunities among those thousands of zombie unicorn companies, trying to uncover value. I feel optimistic about this aspect rather than pessimistic. So I might also be interested in that.

Patrick: If you only consider the return aspect, as your former partner Andy Ratcliffe often says, to make the most money, you have to go against the crowd; you must look for those opportunities that the market has overlooked. Part of the answer might be in the private markets outside of AI, where pricing and supply-demand conditions are completely different. It's like if you go to an ordinary company, the capital markets are not particularly eager to invest in them, and to some extent, they will use very strict criteria to evaluate these companies. This is completely different from their approach in the AI field; shouldn't we pay more attention to these areas?

Bill: I even feel that some of those previously considered late-stage investors are thinking this way as well. They are wondering if they can find a traditional company that may not yet realize that AI could enhance it, but we can do it ourselves; perhaps this is a disruptive way of looking at things Howard Marks was the first to propose the idea of "non-consensus yet correct." I have read a lot of his work. However, this idea conflicts with platform transformation. Because platform transformation has now become a consensus, if you want to go against the trend, you cannot invest in AI, which sounds absurd. Therefore, it is difficult to do both things simultaneously.

Regarding AI, there is something very interesting: large companies seem to be acting very quickly. If you visit the Service Now website, you will find AI's shadow everywhere. Microsoft's earnings call mentioned AI 67 times, and Satya spent two hours discussing AI. This is strange. In "Crossing the Chasm" or "The Innovator's Dilemma," we read that large companies should be slow to react in the mobile and internet fields, which would give startups an opportunity. But this time, I believe many large companies have started paying attention early on.

Patrick: Do you think this is just happening in a different form now? Perhaps as an example, theoretically, Google should be in the best position to dominate every AI use case. However, basically, none of the people I know use Gemini or Google for collaborative generation, or as tools for daily use, nor do they use them for everyday LLM tasks. Frankly, they are not used for many other things either. Instead, they are using startups like Cursor, Anthropic, and OpenAI. So, even if large companies act quickly, the tech companies themselves are still replaying the same phenomenon.

Bill: I think there is data on both sides, and I find your argument very interesting. Apple is an interesting point. You know, Microsoft missed an opportunity but still survived, which makes it more capable of being alert for the next opportunity, right? I saw an interesting interview where Friedberg interviewed Sundar and asked him if he had read "The Innovator's Dilemma," and he admitted he hadn't. So when your company is thriving, these theories seem to be for others, but maybe it's time to read them now.

Patrick: When evaluating an exciting new AI company, its revenue nature may differ from traditional models like Software as a Service (SaaS). As an investor, how would you assess the revenue quality of a new AI startup?

Bill: I think this is very difficult, for reasons I mentioned earlier. You might receive a $1 million order, but its gross margin could be negative. On the other hand, you might find that AI models from two generations ago are now priced at only one percent of what they were initially. You may have confidence in price optimization in the future.

Partners at Benchmark have been studying and assessing when companies enter optimization mode and how they make different decisions compared to being in experimental and sandbox modes. With your amount of funding, you can run sandbox modes for a longer time. Before entering optimization mode, I want to emphasize that in the first two years of the internet, all startups were building on Sun and Oracle, but five or six years later, no one was doing that anymore, so it is very important to pay attention to this shift. **

If China's AI Models from Tech Giants Are Open Source, They Will Be Super Powerful

Patrick: How do you view the international competitive landscape in the AI field? In the transition of some other technology platforms, this competitive landscape is relatively weak, mostly with American or Western technologies at the forefront. China is clearly a region worth paying attention to, especially with projects like DeepSeek. Now, more and more Chinese startups are launching impressive products. How do you see the international factors in this race, especially the AI competition between China and the United States?

Bill: There is a very interesting development in China that is worth noting. When DeepSeek was launched and rapidly developed, we were all watching the American response, the American model, and Washington's policies.

But in China, Alibaba has open-sourced Qwen, and Xiaomi now has its own model, I can't remember the name, it might be called MiMo, which is also open source. Baidu's Robin Li originally had a closed-source model, but he said it would be open-sourced in June. If this level of competition ultimately gives rise to open-source products from four well-funded companies, it will be super powerful. We have already learned that these models can train each other, help each other, and get better. So if there are four open models that can train each other and everyone can use them, I think it will bring a lot of choices and experiments, which won't happen in the U.S. This is the most compelling part of the international AI narrative I have seen.

Patrick: Do you find yourself more loyal to certain groups, hoping they win over others? Who do you support the most?

Bill: The point you mentioned is interesting; I noticed that some of the most radical Chinese investors are actually betting on military companies supported by a new generation of venture capital. I hate that you might become a war profiteer, but I know this could happen because when I invested in Uber, I would defend it at all costs. It's natural, like your child, you would protect it, so your stance changes with your investment targets. I still have that feeling for any company related to Benchmark, and I'm not sure if one day I will lose that feeling. That's the reality, and that's how the world works.

In terms of the technology itself, I find some non-LLM directions very exciting. I'm looking forward to seeing what robotic intelligence can achieve. I hope we can make progress in the healthcare field. I don't think all diseases will disappear in ten years, as some AI founders say. I think that's an exaggeration, but the process will be interesting. As you said, I use these things every day. The pace of change is the fastest I've seen in my career. If you don't check the news for a week, it feels like you've entered another world.

Patrick: You just mentioned the defense startup ecosystem. I want to expand this to the physical world, the hard tech ecosystem, many of which are actually unrelated to war, such as mining companies, etc. What do you think of these types of companies? They are undoubtedly tech companies, usually operating in very large markets, but with high capital density and requiring a long time. For example, in fields like nuclear fusion and fission. What do you think of this type of private equity market and tech investment? I know you haven't invested in this type of project much before, maybe you don't like it?

Bill: Generally speaking, if I were a professor, I would say you can study it using mathematical methods; the returns in these areas are usually not high. You can look at it—15, 20 years ago, there was a lot of venture capital flowing into solar energy, and the results were not good. Of course, there is an exception to this rule, which is that anything Musk gets involved with tends to have returns. So SpaceX and Tesla are data points, but they are actually exceptions. And they are both related to Musk. So I think we need to see four or five non-Musk individuals do this to know if it's feasible.

From what I understand, have heard, and researched about his execution capabilities, as well as the speed he has demonstrated in these companies, I'm not sure if others can do this or if they have the capability to do so. If they succeed, that would be a great thing for the world. By the way, we have already seen that the abundance of capital has led to greater interest in companies with lower capital efficiency, as if there is some correlation between the two. So another thing to watch is whether market demand will still exist if capital becomes tighter. Many businesses involve regulation.

Patrick: It seems that this situation is about to happen; those companies that have received venture capital, gained early support from the private market, and are involved in heavily regulated large industries will be dealt with. I'm curious about companies like Anduril, which may have a valuation of $30 billion. Although it doesn't reach the level of SpaceX, it's an important data point that makes you feel we really have the capability to execute in companies that require a lot of funding.

Bill: Undoubtedly, from a regulatory perspective, this is true. So I think historically, it has been difficult for companies to make breakthroughs in these industries, mainly due to regulation. Before Tesla, there were probably seven electric vehicle manufacturers, right? There were other attempts to manufacture cars, but none succeeded. I think many people encountered obstacles in this process from a regulatory perspective.

Anduril obtaining certification from the Department of Defense and actively selling products to the military is undoubtedly a new milestone for a startup and quite impressive. I don't know if this means every VC should dive into it. You know, it's difficult. If you can start a software company, or as people often say, start a social networking company, let it thrive and achieve high-margin revenue growth, that is much easier than the way we are discussing making money now.

Patrick: What other areas in the current ecosystem that we haven't discussed are you particularly interested in, such as types of companies, investment strategies, or dynamics?

Bill: If I were still an active GP, I think I would consider verticals in AI and think about where AI performs exceptionally well. AI is very strong in language, and coding is actually a more refined form of language, so AI is stronger in programming, and these areas are important. There has already been a lot of exploration in fields like law and customer service. But I feel there are still some areas that have not been fully tapped. This intersection is very interesting to me

What Happens If the System Corrects?

Patrick: Returning to the LP perspective and the issues at the systemic level of the capital markets that we discussed at the beginning, you have articulated the reality and various incentive mechanisms, or the lack of motivation for change. What do you think will happen in the next five years?

Bill: My intuition is that we are in trouble. Although I have had some success in venture capital, I have always been more of an analyst than an optimist. I started out doing security analysis, and I am naturally inclined towards critical thinking, so my bias is in that direction. Someone could certainly counter me by saying that Gurley always predicts the next recession or something like that.

The current system has led to worse liquidity, less construction of traditional high-quality companies, and a faster burn rate. To me, this is not a good combination. Everything is self-reinforcing. All the components I listed, unless something happens at the LP level, I don't see a correction mechanism. I feel like we are increasingly trapped in that cycle.

You may have seen a great video where Josh Kopelman briefly explains some GP math from his perspective. But I find it hard not to agree with what he did there. From the prices we pay, the amounts we spend, and what needs to happen for VC returns to be comparable to historical levels, it all seems like a tricky situation that is hard to untangle.

Patrick: So, if a reset occurs, what will happen on the other side? Let's imagine a simulated scenario where we can apply public equity pricing reviews or mechanisms to every available asset, resulting in a large-scale pricing reset. If we go through a bad period, what are the pros and cons after the reset?

Bill: I think most people would find that terrifying, after all, having gone through a few resets like that. But I actually find that as an active GP, I am calmer and happier in a reset environment, and I find my work more fulfilling, efficient, and productive. Some others might prefer the bubble periods, like those with sales talent who enjoy being in the thick of it. But I find that the conversations about building traditional companies happen more efficiently and authentically during these windows.

When the internet bubble burst, those impostors left Silicon Valley, both B2C and B2B. There’s a joke that when making money is no longer easy, people leave Silicon Valley and go back to consulting and banking. I don't like those opportunists; I think their motives are impure. They often overhype, overraise, and overengage in secondary financing, and then pull out when things might take a sharp turn for the worse. When handled at that speed, it becomes part of the world.

There was a joke back then that when making money is no longer easy, B2C and B2B turned into "Back to Consulting" and "Back to Banking," because money is hard to come by, and opportunists leave. I don't like those opportunists; they often overhype, overraise, and overengage in secondary financing, and then pull out when things might take a sharp turn for the worse I don't like it this way, but in this fast-paced world, it's also a part of it.

If the market corrects, everyone will look for new opportunities. One of the reasons for this situation is that everyone has studied history, understands the effects of compound interest, network effects, and cyclical cycles, and has witnessed both prosperity and recession. Do you remember how long the stock market decline lasted at the beginning of the COVID-19 pandemic? About three weeks, right? Then everyone started to buy the dip. So I suspect that confidence in the AI field is high enough that even if people feel AI is overvalued for six months, it will quickly rebound. Of course, that's what I think.

Patrick: If you were to start a brand new investment company, what do you think is the most important component of brand building for that company? We are in an era where some emerging private market firms, like Green Oaks, Andreessen Horowitz, and Rivet, were established around 2010 and became very large and respected brands a few years later, each with their own way of doing business. We are in a new era. What advice would you give to emerging investors who have just established companies this year and hope to become industry leaders in 12 years?

Bill: What you just said reminds me of a point that is unrelated to your question. Another negative impact of systemic issues is that some companies squeeze into the shareholder list by writing big checks of three hundred million dollars and showcase their differentiation by becoming the best friends of the founders. It's easy for me to say this because I no longer write checks, and no one will stop me from being a director because of what I say, so it doesn't matter.

But they won't take on the responsibility of "helping you make better decisions." They will never say "no." An extreme example is the SBF's FTX incident, where no one joined the board, and everyone believed he didn't misappropriate funds, and in the end, it collapsed. In fact, having someone who can "hit the brakes" at a critical moment and drive the unit economic model is very useful. I'm worried that this situation is becoming increasingly rare now.

The best CEOs, I often categorize people like Barton (former CEO of Zillow) and Benioff (CEO of Salesforce) into this category. Even Meta's Mark has said that they believe going public will make their operations more efficient. Another downside of these companies remaining private forever is that they don't receive this kind of feedback.

Now, let me try to answer your question. I don't know, because with everything I've just said, it's hard for me to imagine what it would be like to embark on that journey. So I really don't know. I can only give up on this.

Patrick: Well, I'm glad my question sparked other thoughts. Perhaps at the end, you could leave some special thoughts specifically for founders; I always like to return to the founders. It makes sense. Because without them doing the work, all of this is meaningless.

Bill: Yes, if you are fortunate enough to enter a hot company, you will find yourself in the world we just talked about, and I would like to offer a few pieces of advice.

First, unit economics will one day be very important, but that doesn't mean you have to sharpen pencils right now. As I said, the model price from two generations ago was only a fraction of what it is now. You can plan for the future to shift towards that kind of model, and that's fine I think it's okay to burn money now, but the unit economic efficiency will ultimately become key. In the end, you must scale the company and operate in an efficient and productive manner. I believe that when you go through this comprehensive battle, you may lose your direction. I find that many founders are thinking about this issue. They believe that this is what big companies do. It's too "bureaucratic," and it's not the original intention of entrepreneurship. But when your revenue exceeds 100 million and reaches 1 billion, you can no longer operate this way. This is actually advice applicable to any cycle, but it is particularly important in times of abundant capital.

My favorite Reid Hoffman wrote an article about Uber, using the metaphor of "pirates-navy," saying that all startups start as pirates and eventually have to become a navy. This is true. For some, this transition is uncomfortable, but you have to find a way that works for you.

Another related idea is that Ben Horowitz also wrote a great blog saying they only want to support founders who can stick it out. This is a genius way of writing because founders love to hear it. In fact, every venture capitalist in the world is like this because the success rate of replacing a CEO is only 50%. Why would you want to take that risk with your portfolio? But there are two or three paragraphs in that article that say, "Of course, this is premised on the founder's willingness to learn how to lead. I feel that our industry often overlooks this—no founder is born with the ability to lead an organization of thousands, and some spend their whole lives studying how to excel at it. There are only a handful of founders, probably around 30, who worked with Bill Campbell to help guide them on how to do this. But this is not innate; it is not free. And you have to be willing to do it. There is a type of person who finds it hard to do this. I had a great conversation with Michael Dell. He did it, even though he felt he didn't want to, but ultimately found a way to do it while being happy. This part is difficult.

Finally, there are two more points. First, network effects are real, and if you pay attention to them, you can make them stronger. If you are in a hot market with a lot of business and gross profit, it's easy to overlook network effects. But you need to think: does your business have some kind of "data byproduct" or other mechanisms? For example, if you have a thousand customers, when it turns into two thousand customers, the experience of the two thousandth customer should be better than that of the first thousand. Can you design this mechanism into the system? If you can, it will have a profound impact on your company's long-term success.

The AI Consumer Field is Where the Opportunity Lies

Patrick: What do you think should be done in the AI era? Is it mainly a data issue? Do you just hope your product naturally generates more data and then improves the product?

Bill: Assuming you serve a functional vertical field, if a customer's learning experience can benefit the entire team, that would be very powerful, and I think this is entirely achievable. For example, there are some AI companies in the legal field. I am not involved in them, but they are researching all the information you would use in litigation while also studying all the precedents and historical legal cases. AI will do certain things, and if there is human involvement, you will find failures and continuously improve the model. These factors may allow early leaders to further expand their advantages in long-term competition Patrick: Do you think AI will make consumer internet investment-worthy again? Since the era of mobile internet, a large number of amazing consumer-facing businesses have been established, but American VCs have paid little attention to this area, and capital has flowed very little there. Do you think AI will make this field attractive again?

Bill: I've noticed that certain applications of AI in China have made significant progress, which may be a signal. While most AI investments in the U.S. are concentrated on the enterprise side, there may indeed be some untapped opportunities on the consumer side. We actually had early attempts with Character.ai. However, the first generation of LLMs has two issues that make them unsuitable for the consumer space: the first is voice capability, which is improving; the second is memory capability, which is also improving, although it's being achieved outside the main model, but that's okay, it will eventually be integrated into the context window. As technology advances, these shortcomings are gradually being addressed. If AI can achieve more natural interactions and more personalized experiences in the consumer space, then the consumer market may once again become a hotspot for venture capital. If four or five AI companies make breakthroughs in the consumer space in the future, I wouldn't be surprised. This could be the contrarian opportunity we've been looking for, as most resources are currently focused on the enterprise side.

Patrick: Bill, it's so interesting to talk to you. Maybe we should do a market status update every few years. Since you're no longer doing this directly for LPs, we can do it for the whole industry. Thank you for sharing your experiences and insights with us.

Bill: Of course, I know the LP community is very interested in what you're doing, Patrick. If anyone has feedback on what I said, wants to correct me, or has any suggestions, please feel free to reach out. I really enjoy this industry and hope my sharing is useful. So I encourage anyone who has feedback for me or wants to correct what I said to contact me. I'm happy to engage and learn more.

Patrick: We'll send out the signal. Bill, thank you very much for your time