What Silicon Valley’s Venture History Can Teach Global Investors

Q&A with Sebastian Mallaby, based on his recent book The Power Law


Originally published on my Forbes column here.

The Power Law by Sebastian Mallaby is the astonishingly frank and intimate story of Silicon Valley’s dominant venture-capital firms—and how their strategies and fates have shaped the path of innovation and the global economy.

The Power Law' Is a Funder-Friendly Look at the World of Venture Capital -  The New York Times

The book details the fascinating history of tech incubation by way of storytelling and analysis and helps anyone in the venture capital world think more effectively about their future in it.

Insight into the venture ecosystem as a whole can provide a new perspective on how specific emerging startup industries, notably fintech, can emerge from the capital standpoint and endure over the long-term.

I sat down with The Power Law’s author, Sebastian Mallaby, to get some of his thoughts.

Key takeaways for me:

  • Building venture capital moat requires balancing embeddedness and differentiation

  • New differentiation will emerge from new geographies, specialization and unique bridges

  • The importance of governance, and how it failed in some later stage deals

  • The pros and cons of blitzscaling

  • The role (and lack thereof) of AI in predicting market behavior

  • The coming diversification rather than institutionalization of VC

  • Advice for young VCs – get a few years of experience first

Alex Lazarow: Your book does a great job of highlighting the differentiation between more prominent, established firms – and smaller startups. Given the explosion of new emerging managers, what advice would you give a new venture capitalist?

Sebastian Mallaby: A moat in venture capital derives from the tricky mixture of differentiation and embeddedness. In short, you need to be differentiated and deeply embedded in the mainstream.

Let's start with embeddedness. Who are your connections? Who trusts you? Who could you bring on to work for a portfolio company? You need strong, mainstream answers to these questions.

Embeddedness can also include non-obvious networks of value. Being deeply embedded in the Carnegie Mellon ecosystem, for example. It is not in Silicon Valley or off the beaten path. Embeddedness thus trends towards the mainstream.

At the same time, you need to be different from other funds out there. Successful launches, like Benchmark, link deliberately small funds, high fees, and a promise that the high fees represent greater performance.

Accel pioneered the Prepared Mind and a specialized VC fund. Founders Fund differentiated itself by being super founder friendly. A16Z differentiated with investment services.

They are all differentiated, even if perhaps their underlying value derives from being equally mainstream.

Lazarow: For new fund managers today, what are the vectors of differentiation?

Mallaby: By definition, today's differentiation will not be what was done differently before.

New differentiation will come from things like:

  • Atypical geography: e.g., North East US, or an emerging market

  • Network: for example, a unique network among CTOs for a SAAS fund or financial institutions for a fintech fund. In defense technology, there is an enormous value derived from having connections inside marine navy airfare. Getting the first contract is a big advantage.

  • Bridging between geos. In many ways, this was Masayoshi Son's original value proposition. Whether that was Cisco or Yahoo – or other deals early in the US – part of the attraction was launching the Japanese version of the product and getting into the market. There will be new ways to scale this in the future.

Lazarow: What's the one thing you would change about the VC model if you could that wasn't mentioned in your book?

Mallaby: The thing I’m most troubled by in the VC model is growth investors. Described in chapter 14, before the conclusion, are the WeWork and Uber stories. What started out really well for Uber almost went off the rails – as it did with WeWork.

I blame that on how growth investing has become ‘governance light.’ I would venture to say growth investors are the worst offenders in American capitalism in terms of owning major stakes in companies but not taking care of governance.

We have different types of governance in the US for different types of companies. In public companies, if you have bad governance, the CEO can get fired, or the business can get shorted. So there is discipline.

In the PE model, an investor buys the full company, which creates strong incentives. And in early-stage venture, where a VC might own 25% of the company, they have strong incentives and aligned interests.

But at growth, problems emerge. Founder deference is good, to a point. But some oversight is missing.

Lazarow: What is your view on ‘blitzscaling’ today in Silicon Valley?

Mallaby: For network-based businesses, the value you create is the square of the people on it. The winner takes all the advantages. Blitzscaling is a path to capture those advantages.

People raised criticisms of blitzscaling. One is it pushes founders too aggressively. They then make mistakes, which is bad for mental health as it creates undue pressure.

When there is commercial logic (in a network-based business), it is up to the founder to decide if they are up for this.

But for non-network-based businesses, the logic is trickier. The case of WeWork is a perfect example. The network effects there are weak. When Masayoshi Son said ‘bigger faster’, that was a mistake. The downside of being bigger was worse than the upside.

[For my own thoughts on the right approach to sustainable, resilient long-term growth, see my work on Camels here.]

Lazarow: What do you think about different fund structures like Evergreen funds or other deal structures like revenue-based finance?

Mallaby: One clear lesson from early VC history is that debt is a big mistake for startups. The whole point of a startup is that it grows fast. Therefore, startups should not pay interest on debt. They should retain capital. On that note, Softbank structured itself with debt. That was a red flag for me.

An expansion of how long you want to hold companies can make sense regarding the Evergreen story. For Sequoia, it can make sense to do seed, series A, growth, and public as well.

If you understand technology in one stage, you have a foot in the door for other stages. If you deeply understand the disruptive technologies you're backing, it might tell you what to go short or go long in public markets. In that context, Evergreens can make sense.

The risk, though, is that over time a company that plays in public markets will find that overwhelmingly the profits (and risks) will be dominated by the big checks in a later stage. And the early stage team will lose influence at the partners' table. It’s still too early to know how that pans out – but it is a trap.

Lazarow: A lot of talk has been given to the rise of AI and computerized decision-making (e.g., Two Sigma, Convergence, etc). This can potentially remove bias. But it also removes the deals' humanity and the founders' selection. What is your thinking on this?

Mallaby: There are a number of players here. Hedge funds have a much longer history – think DE Shaw, Renaissance, and others. But today, qualitative hedge funds are claiming a bigger chunk of AUM in hedge funds.

I would argue that the discretionary human element will not go away. It is about changes in the game – data on past episodes which take you back to Spanish flu are irrelevant to dealing with Covid in the future. The pandemic was huge. The only way to understand that is through discretionary measures.

So, AI can be helpful, but it will not be the whole story to get you into deals.

Lazarow: Will we see an institutionalization of venture capital? Will there be a Goldman or Bridgewater of VC?

Mallaby: There is probably space for all sizes of players. A16Z is aiming to be the JP Morgan of the technology industry. Just like JP Morgan was the dominant financier of the industry. A16Z would like to do that with tech. And they might succeed. Sequoia is less explicit but is also heading the same way.

At the same time, Benchmark continues to stay focused on a certain stage. And while some solo capitalists might burn out, some may do very well. Heterogeneity will always be part of it. In the future, venture capital won't be dominated by one model.

Lazarow: VC investors are often poor managers of people, and VC funds are not always the best place to get trained early in your career. What advice would you have for leaders of VCs in developing their firms and for up-and-comers trying to enter the industry?

Mallaby: Most successful people share a certain set of qualifications:

  1. Technical qualification - masters in software engineering or some other

  2. Business school or equivalent familiarity with marketing and sales

  3. Having been a founder or early employee of a startup

There are exceptions. But ultimately, venture capital is something you can do in your late 20s or early 30s. It is worth getting experience first.

Be prepared to do it for 5 years and only then start becoming good at it. You need to develop experience. All successful VCs share a steep learning curve for the first 5 years.