Chi-Hua Chien saw Facebook coming – now he says the real winners of AI won’t sell AI

Chi-Hua Chien has spent more than two decades as a capitalist, but he thinks like a cultural anthropologist. As the founder of Goodwater Capital, a firm focused on consumer and prosumer technology, he has built a portfolio spanning entertainment, healthcare, fintech, and live experiences – with investments in companies such as MIDI Health, Fever, and Monzo. And, as a 27-year-old partner at Accel, the person who originally founded a six-person company launched at Harvard called Facebook.

That ability to learn human behavior at scale informs everything in his opinion that Americans will never trust a single application that has both their social life and their finances, in his belief that the gap between the most advanced AI model and what you can use on your phone – and a range of two years – will shrink to three months within the next year.

These days, he is also willing to say out loud what many are only thinking: that the commercialization of the model layer is already underway and that the big winners of the AI ​​era will not be companies that sell AI at all.

We talked last week. This interview is edited for length and clarity.

More founders and investors have been publicly sharing their grievances about VCs lately. What has changed?

It’s part of the meme-ification of everything – you see what’s happening in politics bleed into the business side, and maybe it’s a sign of something rising in the market. The reason you see some of these very vocal investors speaking publicly is because the venture firms are very syndicated, so the really big ones have enough capital that they don’t really want syndicate partners. There has been a decoration in wanting to maintain a good relationship with other co-investors, because you have to work with them at different points in the line. As firms become larger and more vertically integrated, there is less demand.

What about “fast track” rounds – where firms invest a large amount in one valuation and a smaller amount weeks later in a higher position, making the headline number look more impressive than it really is? Is this really new? How full is it?

I think it has been going on for a long time. The leading companies are raising successive rounds very quickly – there can only be three to six months between rounds now, and the prices change really fast… The prices are sold very much as a way to show market leadership, to attract talent, possibly to prevent competition. There is probably something to deflate the bubbles, because what these fast currencies show is that there is more demand than supply. An investor can come in, set a price, complete the financing, and then a few weeks later there is excess demand – and the company can quickly sell a new round at a higher price.

He argued that infrastructure companies are for sale and that applications hold great value over time. Are we already seeing that play out this cycle?

If you look at the PC cycle, the web cycle, and the mobile cycle, they all follow pretty consistent patterns. The infrastructure markets peaked in the year 2000 – but fast forward 25, 26 years later, and in dollar terms, the market cap of those infrastructure companies did not exceed the 2000 peak. In the web era, new infrastructure entrants generated $400 billion of new market capitalization. Application companies made $3.1 trillion – 88% of the new value. In the mobile era, it’s very similar: Infrastructure generated about 700 billion, while application companies generated 3.7 trillion. Companies like Netflix, Spotify, Meta, Uber, Airbnb.

Again [last week] you saw something very interesting: Google announced that their subscription to the AI ​​product drops in price from $7.99 per month to $4.99 per month and doubles the storage. We are already in an era of price competition – and companies like Google, with structural advantages in direct integration and distribution, can begin to integrate and price compete for the average consumer.

You keep coming back to personalization as a byline. Is that what separates the next wave of winners?

Hyper-personalization is key with the line, because what does personalization give you? When done right, it gives you higher customer satisfaction, deeper engagement, and higher ARPUs over time.

We have entertainment companies in our portfolio – companies like Triumph and Ritten and Flow GPT – where the customer can say, “This is an AI program.” They say it’s an entertainment program. These companies are getting into 100 million, 400 million, 600 million ARR very quickly, with good margins, because AI makes the experience customizable and personal – but it’s not the core strength that they sell.

We also have a women’s health company called Midi Health. One of the basic barriers to women’s health is that there are not many providers well trained in hormone replacement therapy for menopausal women. By using AI, they are able to greatly expand the delivery of care and treat hundreds of thousands of patients who would otherwise be unreachable. And they can make it cost-effective, which expands access to a market that was previously tied to supply. You can play that forward in all the besieged sections to provide where human expertise is the bottleneck.

How far are we from an AI that feels human and has an environment?

I don’t think we are that far off. You can run your site now on your phone for AI models that are just as good as the best models were six months ago – and that latency is reduced. If you go back two years ago, the lag between what you could run on-premises and what was in the cloud with borderline models was probably 18 to 24 months. It is now six months. It will probably drop to three months by this time next year.

What we don’t have yet are very well defined use cases. You’ve seen this on mobile – when the iPhone was introduced in 2007, people thought it would be all web applications ported to mobile. It takes time for entrepreneurs to analyze what is happening now.

What LLMs do, if you deviate from the way they work in what they do, is basically two things: They make it possible for you to process large amounts of context and understand everything, and they allow you to customize your preferences down to the individual, cost-effectively, with a feedback loop that makes the product better and better over time.

You’ve watched Facebook try and fail for years to build a great app. Why is it so difficult to combine financial services and social entertainment for American consumers?

They’ve taken a lot of shots at the goal — Facebook Credits, launched in 2009 … Facebook Pay, Libra … They’ve never been able to find a real real app. I think people have the right idea about trust, and there is a trust gap between entertainment and social products, and commerce, banking, financial services – especially in the West.

There is a seriousness to financial trading that is very different from the seriousness of social media. And don’t get me wrong – that dullness created a trillion-plus company. But financial services are actually the complete opposite: While the audience has a very high time and low income, financial services activities are very high income and relatively low time. You don’t want to hang out on your banking app. You want to do and be done – but with the highest confidence in the security and reliability of what is being done. Those psychological expectations from customers are very difficult to match.

Are you betting people crave personal connection as the antidote to all of this?

We really believe in this. What do people long for in a world where there is an endless supply of digital content? They yearn for something much more delayed, which is real human connection, real world experience.

We have an investment in a company called Bump, based in Paris – from the original founders of Zenly, which was acquired by Snap … They have created an interface that allows people to interact in the virtual world, which is made up of digital information. We also have Fever, based in London and Madrid – basically the Live Nation of Europe. They started with small, extravagant events – candlelight concerts, the Bridgerton Experience – and have since gone viral.

I think we’re moving in the opposite direction from pure online consumption, and AI as enabling technology, knowing where you go, who you hang out with, where you tend to spend time, can generate a ton of relevant interests that make what you do in the real world even more useful and personal. That makes us very happy.