I’ve been building and investing in AI since 2011 - first as VP of Product at Skype and Microsoft, and later as a Partner at Social Capital and Khosla Ventures. At Khosla, I was fortunate to be part of a firm that backed OpenAI early, one of the most important technology bets of our time.
So why leave to build yet another venture firm?
Because the next few years will be the most exciting and fertile time to build new category-defining AI companies. And I want to focus 100% on helping those founders at the seed stage.
Right now there is a lot of hype in AI. But that doesn't mean you can't have a differentiated view and thesis within that and anticipate what's next vs just focusing on what's obvious today.
In fact, this is what I have done throughout my investing career.
In 2018, I invested in Groq anticipating the need for new AI inference hardware 4 years before the launch of ChatGPT.
In 2019, I backed Fireflies.ai before most people even knew what the "AI application layer" was.
These 2 AI Unicorns have scaled far beyond what most imagined was possible.
And so once again, I am investing "under the radar" today, in what I believe will become the new categories of the future. To do this, I look for founders building net new capabilities—unlocking latent demand or creating demand where none existed before. These founders ask:
“How do we build useful, usable products for 8 billion people—not just the 80 million tech-forward users?”
I call this: AI for the Real World
Specifically, our current thesis today centers around AI’s emerging ability to perform entire functions of work, heretofore done only by humans. Historically, software has provided humans with tools to help them work. Now, the software is doing the whole job that a human would do, often using software tools itself!
Therefore, we are not interested in companies that simply add AI features to existing software or automate current tools. Instead, we're seeking founders who can create entirely new software categories, ones that handle complete workflows from start to finish, just like a human employee would tackle their daily responsibilities.
The basis of this current thesis is the research by OpenAI and Microsoft published this year that shows AI beginning to match human performance on a variety of intellectual tasks AND that AI also excelled at demonstrating human traits like patience and empathy.
The opportunity is immense and expansionary. This is not about AI replacing humans at work. It's about democratizing access to skills that were previously too expensive or inaccessible. In the future, everyone should have access to the equivalent of an executive assistant, data scientist, on-call nurse or software developer - areas where there are actually shortages today.
We believe that these "labor layer" companies have a higher potential to be transformational because they are not merely disrupting previous generations of software (and so disrupting the ~$1 trillion software TAM), but also actually eating into the entire labor market TAM, which is likely 10-100x the opportunity. At minimum, we estimate the addressable TAM to include:
- $2.5T in global salaries in sales, marketing, engineering, HR, and security
- $2.5T more in outsourced IT + business services
And there are no incumbent large tech players in these areas!
At Axiom, our thesis is simple. We back founders that:
- Look beyond the obvious, "hot" categories and invest in AI for the Real World
- Tap into labor budgets and build the AI "labor layer" - software that does the work of humans themselves
- Stay relentlessly focused on what’s possible in 2–5 years, not just today
If that sounds like you, come find me.