top of page

Moats for AI Companies. Do They Exist?


MAY 2023

An investor recently asked us about the moats being created by companies focused on AI. The topic is interesting, so we thought we’d share some thoughts here.

The AI landscape is changing extremely fast and even the concept of who has a moat is up for grabs. A few months ago, orthodoxy was that training a Large Language Model (LLM) cost several millions of dollars, and as such only incumbents or the best-funded startups could even compete. That changed when a memo from Google leaked, showing how quickly the open-source models were catching up. For a few thousand dollars, a company could train a model that was meeting, or in some circumstances for specific tasks, exceeding the models generated by the big players.

Given the shifting landscape, it can be difficult to predict the right moat. The ideas we have been discussing include:

  • Grounded Business Fundamentals. Companies solving a specific business use case that they have internal expertise around can help defend against platform incumbents.

  • Proprietary data. A unique dataset will likely produce models that others can’t match. By this, we mean model results could be stronger, and competitors simply will not have the same data access.

  • Strong Teams. Good experience and adaptability are helpful in a rapidly evolving market. Venture has always been a people business, and even with AI calling human intelligence into question, we tend to think that focusing on the human element is the best margin of safety

If you are going to invest, ‘Be careful out there!’

Below are a few of the articles that caught our attention this month. Moreover, we’ve inserted one or two sentences in italics, summarizing each article’s importance. We hope you enjoy and appreciate the material.


Here's a curated list of things we found interesting.

Google "We Have No Moat, And Neither Does OpenAI"

Leaked Internal Google Document Claims Open Source AI Will Outcompete Google and OpenAI

As discussed above, the question of moats for AI and who will capture the value being created is a hot topic; here is a leaked memo from Google that lays out the case for open source models.

Here’s What Happens When Your Lawyer Uses ChatGPT

A lawyer used ChatGPT to prepare a court filing. It went horribly awry.

Large Language Models (LLMs) tend to hallucinate when they can’t find the answer they want. Funny enough, that can lead to huge problems.

AI for Security is Here. Now We Need Security for AI

Despite all the attention AI received in the industry, the vast majority of the discussions have been focused on how advances in AI are going to impact defensive and offensive security capabilities. What is not being discussed as much is how we secure the AI workloads themselves.

Written by a friend of the firm this is an excellent look into the landscape of companies looking to secure AI models.


Deals that caught our eye.

Absolute Software to be Acquired by Crosspoint Capital Partners for Enterprise Value of US$870 Million

The purchase price represents a premium of approximately 34% over Absolute’s closing share price of US$8.58 on May 10, 2023


What we're listening to.

This month we are throwing you a bit of a curveball, instead of a podcast we are sharing a Keynote by NVIDIA CEO Jensen Huang. With all of focus on AI this month we thought this really had great information about where the industry is heading and is worth watching.

GTC 2023 Keynote with NVIDIA CEO Jensen Huang

Watch NVIDIA CEO Jensen Huang unveil a host of accelerated computing breakthroughs -- from AI training to deployment, semiconductors to software libraries, systems to cloud services -- that will help transform every industry.


Lytical Ventures is a New York City-based venture firm investing in Enterprise Intelligence, comprising cybersecurity, data analytics, and artificial intelligence. Lytical’s professionals have decades of experience in direct investing generally and in Corporate Intelligence specifically.


bottom of page