In 5 years, there will be no such thing as an “AI company”

There will be no such thing as an “AI company.”
Just like there are no longer “big data companies.”
Just like there are no longer “internet companies.”

AI has been everywhere as of late. Every other startup is an “AI company” (with an instant 10x valuation bump), and every other fund is now an “AI fund.” As if “AI” is a whole new thing.

But AI is not a new “what” – it’s a new “how.”

What does that mean? Let’s step back in time for a second.

Remember when “big data” became a term? As if it’s a whole new “what.”

We witnessed the same fervor: So many big data companies; so many big data firms.

For the first time, we could save more customer information than ever before. We could store so much more inventory detail. We could pull so many more KPIs for our C-suite executives. Maybe we could even track where our packages are in real time?

This early wave saw big data as a new “what.” A new self-contained category of business. Folks invested in all different types of data infrastructure, data visualization software, etc. And “big data companies” sprouted up left and right that were just variants of existing companies that added “big data” to their tech stack.

But that was just the tip of the iceberg.

What this first wave failed to realize was that “big data” wasn’t a new “what” — Big data was the new “how.” It wasn’t just another standalone category – it was going to become the fundamental foundation for all companies, old and new.

Big data not only revolutionized the way we do things, but unlocked entire categories of businesses that no one ever imagined to be possible. Sectors that became the hottest sectors of the last decade and made unicorns real.
Social networks that changed the way we interact with others and the democratization of information.
Marketplaces that allowed everyone to run their own business and be their own boss, from craftmaking to driving cars.
Shopping platforms that are personalized to not only what you like but what you should like.
Marketing technologies that can test advertisement copies and audiences to sell everything from shark onesies to our next President.
Fintech, digital health, all variants of enterprise and vertical SaaS.
All the sectors VCs now know and love.
All the sectors that didn’t exist before.

As “big data” became part of the fabric of everything, there were no more “big data companies” (besides the few that stood as the architecture/infrastructure giants) because every company is fundamentally a “big data company”
And there were no more “big data VCs” because that would just make you a “generalist."

What we saw happen with the evolution of “big data” wasn’t unique.

Remember “the internet”? Or “the Internet” with a capital “I”?
The internet that has become the backbone of culture, community, and global society. We are online all day, every day.

The internet wasn’t just a “what” either, but a “how.”
Remember the wave of “internet companies” and “internet funds”? Or even better the “dot coms” as if only a small subset of companies will have web addresses?

So, let’s now cut back to the present.

AI is another revolutionary “how” — another leveling up of data & computational power.

We’re seeing the same pattern of behavior: Talking about AI like it’s the new shiny thing (much like crypto a couple of years back — for comparison, crypto is a new “what”).

Companies self-identifying as “AI companies.” Applying “what we see AI can do” to familiar sectors for iterative improvements. Faster development, smarter personalization, etc. We can now write social media posts faster, render animation with more realism, generate better UI designs, script better sales flows. But are they “AI companies” or the next evolution of a marketing software/media/design/enterprise SaaS company?

Every new firm wants to be an “AI firm.” Every existing firm is trying to start an “AI practice.” Investing in variants of LLM and infrastructure alongside every company with an “AI” spin, sometimes by sacrificing established home-grown sector expertise. But what do you think AI applied to healthcare will look like? It won’t be an “AI company” but the next evolution of a “healthcare” company.

But just like big data, AI isn’t a “what” – AI is a “how” that will terraform the startup landscape.

Every company will be an AI company, and every AI fund will just be a generalist fund.

So, how should we think about the impact of AI?

Press a button, and AI will do it for you.
This is a double-edged sword.

AI’s computational power and reasoning allows companies to be built faster and with less capital, fundamentally changing the sectors of companies that are venture-backable.

On one hand, entire existing sectors will cease to be venture-backable as the time and capital need decreases. To draw the analogy to big data again: In the early days of big data, we saw data visualization SaaS companies charging thousands per month with consulting fees on top. These companies are now replaceable by a single junior data scientist running open source software. We’ll see the same trend here. Entire sectors that have made us f-tons of money could no longer be even viable as standalone companies, becoming just mere widgets or tools buildable over a weekend. Think sales script optimization, think advertisement copy optimization. Other existing sectors will have to continue to leverage AI creatively to evolve.

On the other hand, entire sectors will emerge as new venture-backable opportunities — sectors that we have only dreamt of as being possible. AI will make possible businesses that we’ve always assumed would take too much effort, capital, and time to bring to life. This is where I get most excited: making the unfundable fundable.

Imagine EdTech, but make it AI.
Imagine Supply Chain & Complex Logistics, but make it AI.

And, now, imagine Science, but make it AI.

I’m going to call this category ScienceTech (SciTech for short, if you will)
ScienceTech companies are the software, platforms, or tech-enabled services that accelerate scientific discovery or assist in bringing science to life.
(Much more on this in coming issues)

Imagine…
Not the drones, but the operating system that can coordinate swarms of drones.
Not the drug, but the platform that accelerated its discovery.
Not the polymer, but the framework for generating the most function-appropriate formulation.
Not the satellites, but the software that powers real-time fleet trajectory coordination.

Much like fintech isn’t the old school banks, but the rails that power the banks. SciTech is the category of companies that are the rails for doing science, powering and accelerating our race to new discoveries. And these companies are now only able to come to life because of AI.

ScienceTech won’t build George Jetson’s Rosie. But it will help develop the energy for her mainframe, and give her a memory bank that tells her what his favorite breakfast is.

Most excitingly, this new type of company will have the best of two worlds:
The intellectual property defensibility and the vast market potential of Deep Tech
And the revenue paths and the growth profile of SaaS, the darlings of the last decade
And so at last – SciTech makes science venture-backable.

And we’ll talk about the potential for “f-tons of money” to be created in this sector next week. 😉