The hardest current question in Venture Capital: What Moats are still effective in the era of the Great Commoditization of AI?

One of the questions I have been asked most frequently by fellow investors this year is, what businesses are defensible in the future? Or said more precisely, what moats are still effective in this era of AI?

The short answer: Things that cannot be built or created with AI.

*Anything that is built with AI cannot serve as a sustainable moat, because AI is globally commoditized.

We all know by now that the Great Commoditization of AI is here (if you somehow missed this or still don't understand why this is not the GenAI era but the Great Commoditization, please read one of my long-form essays on the topic). Many of us can already begin to see how this force is integrating AI into every element of our work, lives, and society. And yes, it is moving faster than any past technological revolution.

Those who read my stuff know that for several years now I have been writing about the demise of the SaaS business model (specifically the erosion of SaaS pricing power) and the Aigentic takeover of line-of-business applications. You have heard me talk about the coming Golden Era of Small and Medium Business and my strong conviction that soon vertical tech will be dominant over horizontal tech (in most cases). But the thing I have been thinking about most is business moats and the fact that it seems to me that these are changing before our eyes.

For those who are not native to VC land and the Startup world, moats are an important and frequent topic. Especially for those of us whale hunting to find a business that can get really big. We are trained to build moats, to protect them, and to invest in them. This is why this question is so critical for founders and investors alike. So, rather than waiting until my next long master essay is ready, here are some thoughts and observations now.

Old Moats May be Dying (or at least losing their protective power)

It seems to me that perhaps the best and most powerful early-stage technology Startup moat of the past 25 years was a differentiated technical founder coupled with an enviable technical team. This was determinative in many investment decisions over the past decades. It's why every investment deck tried to show a Google, Apple, Stanford, or similar logo near the pictures of their technical team. And why we all know investors who've said, "I'm investing because of what he/she built at ___(fill in the big 7 tech company here)___".

Further, I think it is indisputable that the best long-lasting moat of the past 20 years was the Aggregation of Demand. As Ben Thompson has taught us, over the past 20 years of the internet, it has been those aggregators that can amass demand that win the day. Facebook, Uber, Airbnb, Expedia, etc. were the big winners of the first quarter century.

I think both of these moats are up for change and with that a lot of what venture investors think we know about investing in early-stage technology businesses.

AI-generated code and AI Employees are quickly eliminating the 'enviable technical team' moat that has been so dominant. A company's ability to deploy high-quality code and custom software are no longer limited by the size or caliber of its engineering team. And with a lot of apprehension and humility (because of my deep respect for Ben Thompson and his work) think that Aggregation Theory may lose its power; at least in the e-commerce arena. Because at the end of the day, the core consumer value add of an Aggregator like Expedia or Airbnb is putting all the options in one place. But while you would rather not visit 134 hotel websites for your next London trip, the AI assistant on your phone is happy to visit them all, to break down your options, and is likely to do so with access to a lot more data about you than Expedia Group or Airbnb have. And thus, it is unclear what value the OTA and other aggregators in the realm of commerce will provide.

My point, a lot of our most valuable rules of thumb and common knowledge are changing, and we must pay close attention.

Emerging Moats

While the power and efficacy of some moats are quickly waning. It seems that there are some new moats growing in dominance, and perhaps some old moats making a comeback. We are still on chapter 1 (even if we are nearing the end of this chapter) of the Great Commoditization of AI, and thus, there are more things we still do not know than things we can be certain about.

But a few of the moats I feel at least a bit of confidence in include...

Differentiated Data Access - or what my brilliant friend Court Lorenzini refers to as a 'Organic Self Perpetuating Data Moat' or data creation strategy. In a world where nearly every startup can credibly claim to be utilizing the world's most cutting-edge AI (because we all use APIs from the top few AI model builders such as OpenAI or Gemini), one of the only sustainable differentiators is a company's access to proprietary differentiated data to train its models on and keep them from drifting or from becoming commoditized.

Founder Quality - one thing that seems certain is that due to the pace of evolution of both AI as a technology and its impact on our economy, moving forward Founders and Startups will need to pivot, adjust, and restrategize more frequently and more quickly than ever before. As many of us have been saying for a while, the pace of change will only continue to accelerate. While founder quality has always been one of the key considerations for investors, startup employees, and the market, in an era where everything is up in the air, founder quality is the only constant. Smart investors understand that it is literally impossible to guess where AI and our economy will head over the coming years and when the turns in the road will materialize, and thus a bet on an incredible founder with domain expertise and great agility may be as good a moat or competitive advantage as we have these days.

Speed - directly connected to Founder Quality; a Startup's speed, flexibility, and agility, or said simply, its ability to respond to technological and market changes more quickly than its competitors, is likely a meaningful moat in changing times like these.

Access to Rare or Limited Inventory - for much of the past 30 years most of us have focused on digital goods and digital inventory, which for the most part is infinitely extensible and effectively free to transmit, and due to AI, increasingly cheap to create. But in the real world, most inventory, especially most aspirational inventory, is limited or rare. Thus, access to limited non-digital inventory will likely serve as an effective moat for certain types of businesses. The easiest examples are found in the realm of travel and luxury. An upstart can build a better AI Travel Agency or a more digital native luxury retailer but these upstarts with amazing digital offerings cannot get you a suite on the new Four Seasons Yacht nor help you get the Berkin Bag, and thus, in this race it is likely to be the legacy player with differentiated access that will win the day over the AI native upstart.

Brand & Trust - I had written about this a lot in 2021 & 2022 when some were predicting that creators were going to become so dominant in our economy that Brands would no longer matter. As I wrote back then, in a world overrun by AI, DeepFakes, 3D printing, and more, it is a Brand whose incentives are aligned with the long-term interest of the consumer to have a product that maintains both its utility and its style well into the future. With another 4 years of data in the books, I believe this more strongly today than I did back then. We have all seen the general level of public trust diminish across society, and to address that, we have seen the rise of Creators, Web3, and the Social Commerce movements. All with the same simple goal to establish trust with the next generation of consumers. And while all 3 of these forces will continue to grow, I think Brands will remain a critical way to establish and maintain trust.

As we see the massive proliferation of AI build platforms, websites, apps, and other digital experiences, Brands and the trust they convey will play a determinative role in who wins in a given market.

Direct Access to Clients & Distribution - A company's ability to sign and maintain high-value clients and defensible or preferential access to distribution channels has always been an important moat. But this one seems to be growing in dominance. In an era when technology and product moats have lost their protective power (due to AI Generated Code and such), traditional business moats such as a distribution advantage will become ever more valuable and differentiating for investors.

Related Market Shifts

When engaging in these discussions, there are at least 2 common refrains, which to some feel like material pushback on the above assertions but I think lack a broader market perspective.

Market Size Shift, or more accurately, a decrease in the number of viable market participants - one common, and correct, pushback is that most operating businesses, like a hotel chain or tour operator, will not end up building many of their own tools and software, and will still license and utilize many outside technology providers. But even in this scenario, it seems likely that the market for Startups and companies providing tech to these players will shrink due to the twin forces of at least some tools and technologies being build in house (and I think with time we will see a lot of this; just think about my building v. buying a house analogy) while at the same time individual outside technology vendors will also benefit from AI Generated Code and AI Employees and will continue to expand and extend their offering. Deploying the age-old land and expand strategy, which has always been a good idea but is even more valuable now because of data interoperability.

Thus, it seems likely that a medium-sized hotel chain that today has let's say, 150 technology vendors will have 120 in two years and maybe only 75 a year or two after that. Effectively shrinking the market by 50% and with that, reducing the number of successful upstarts by the same amount.

Pricing Power Shifts - The second element of this market shift is pricing power. I have written for a long time that SaaS is dying, and what I mean by that is that SaaS's pricing power, and with that its dominance as techs' preferred business model, will wane. Even in those cases where an operating business (the tech startups client) will not be able to or will not choose to build their own tech, the rest of the market participants will still have access to all the new AI capabilities and a similar outcome will emerge. If today a tech provider to a medium sized hotel chain charges that chain $100k per year for their technology, a new upstart can spot the opportunity, utilize no-code tools to build a meaninful competitor and do so with a dramatically lower cost basis, and thus undercut the market and offer the same, or a very similar solution, for $70k. And six months later, when the AI-generated-code tools are even better, the next entrepreneur will recreate the same thing, maybe even better due to better technology, and will do so at an even lower cost basis and thus will undercut the market again and offer the solution for $25k per year. And the cycle continues.

These two scenarios together (fewer market participants with lower pricing power) point to a future with fewer viable, viable from a VC perspective that is, startups ready for VC investment. But of course, as I have written, this may create a Golden Era for small and medium businesses.

Conclusion

My point is simple: AI is changing many things including the way we build technology and the cost basis for standing up technology companies. Due to this massive technological shift some, or many, of the fundamentals of our tech economy are changing. And those of us who build, invest in, or work at Startups must pay attention!

PS - a bit more Insight on Moats from Smart Friends

A couple weeks back, I had the privilege of hosting 15 of Seattle's leading VC investors at my house for a fun summer afternoon of shmoozing along with a focused conversation about investing during this era of the Great Commoditization of AI. Being in Seattle, one of the global hubs of AI, many of these investors have been thinking about AI nearly as long as I have and have been investing in it much longer. As it often does these days, at one point the conversation turned to business moats.

A few additional Moats they focused on include...

Workflow Expertise - one area in which the LLMs still perform quite poorly is in the realm of complex business processes. There are many reasons for this, including the fact that data about these processes is almost always kept within corporations and held as trade secrets. Thus, Startups and Founders with a deep and differentiated understanding of a specific submarket or complex workflow may have a sustainable moat in this era of AI.

Organic Self-Perpetuating Data Creation Capabilities - I already highlighted this as one of my top moats but this was an area of deep discussion and one where there is a lot of innovation on the horizon. Whereas product innovation is likely not defensible, perhaps these data flywheels can be.

Balance Sheet - this one is probably only applicable in markets and segments where new models must be trained but in those markets, a large warchest is certainly a meaningful moat.

Other Moats discussed at this and other recent VC conversations include (even though for the record, I do not agree with most of these as effective moats) - Product and User Experience, SEO as a moat, High Switching Costs, and more.