How AI Will Reshape Startups & VCs
“We wanted flying cars, instead we got 140 characters”
- Peter Thiel
In the world of startups and venture capital, we stand at the threshold of a transformative era. The rise of more powerful, flexible, and ubiquitous Artificial Intelligence (AI) is more than just a technological milestone; it's a catalyst for a shift in how companies are founded and funded. Coupled with evolving financial market conditions, these changes are poised to redefine the startup and venture capital landscape. This article delves into these shifts, exploring their far-reaching implications for venture capital, startups, and particularly the burgeoning field of deep tech.
AI’s Impact on Startups
I believe that the recent advancements in AI are a game-changer for the startup ecosystem. In my venture studio (Mark II Ventures), we're seeing this transition firsthand and actively experimenting with a wide variety of automation tools. I can already see how these tools (and the ones soon to be built) will impact how we do everything from product development and customer support to back-office operations and financing. While these tools are in their nascent stages, their rapid evolution is set to revolutionize the way we build and scale SaaS products.
I first started thinking about this transition after reading this blog post by James Currier, a General Partner at NfX Ventures. Currier foresees a future of unicorn startups (private companies valued at >$1B) being built and scaled with just 3-person teams. His deeper argument is about the impact of AI and automation on how startups address the functional areas of building a company (marketing, sales, development, support, accounting, etc). While I agree with his assessment on the magnitude of the impact automation will have on startups, I have a slightly different conclusion on what results from that impact.
The New Economics of Building Startups
The impact of automation on startups is multi-faceted and profound. It will dramatically change how companies operate within every functional area of the business.
Automated marketing and sales tools are poised to drastically simplify and reduce the cost of customer acquisition. Imagine AI-enabled platforms that can identify potential customers, personalize marketing messages, and even close sales, all without human intervention.
This isn’t a fantasy. These tools are already being built and their capabilities are growing at an unbelievable pace. So much so that we’re actively reassessing our portfolio’s go-to-market strategies and financial models to anticipate a much, much smaller growth team that is able to produce far more revenue per FTE due to the smart implementation of these types of tools.
Similarly, in the near future, chatbots and AI-driven help desks that autonomously manage the entire customer support function will become the norm. Not only are these tools cheaper, but they will end up being better too. AI support bots don’t get tired or mad. They scale instantly to meet customer needs (i.e. no more long wait times…which is a killer of customer satisfaction) and they are available 24/7/365.
In the realm of product development, tools like Copilot are already amplifying the productivity of software engineers exponentially. While some of the stats being thrown around online are exaggerated, I can definitely see how these tools will turn a 10x engineer into a 100x engineer. This radical efficiency means startups may never need to hire dozens, let alone hundreds, of software engineers, significantly reducing overhead costs and organizational complexity.
All of this means that it will become vastly cheaper, easier, and faster to build and scale a software company.
The Rise of Mirco-SaaS
While Currier sees these advancements resulting in a future of 3-person unicorns, I foresee a much different one. I believe automation is nudging us towards a proliferation of "micro-SaaS" applications — hyper-specialized products tailored for very niche industries.
Currier is right that these will likely end up being companies with only a handful of employees. But I don’t think most will be pursuing billion-dollar markets. In fact, I think that the very forces that would theoretically enable a 3-person team to operate a business doing $100M a year in revenue will actually create headwinds to pursuing those larger market opportunities.
There are two big implications of automation that I believe will push the software markets towards a future dominated by lots of small SaaS companies rather than a few large ones.
Increased Market Participants: As it becomes easier to launch a startup, the market will naturally see an influx of new entrants. This increased competition will cause margin compression and drive entrepreneurs to seek out smaller, niche markets where they can build a stronger defensible advantage and maintain pricing power.
Economic Viability of Smaller Markets: Historically, the high capital requirements and associated risk of building a SaaS company meant that entrepreneurs needed to target larger markets to justify the investment risk. Drastically reduced costs mean that it will become economically viable to target smaller markets, even those with a Total Addressable Market (TAM) in the tens of millions.
The Tattoo Shop SaaS
To put this into perspective, let's consider an example. Traditionally, venture-backable SaaS companies target large (and often relatively generic) customer bases. This is to ensure that the market size is large enough to offset the risk profile of venture investments. For example, companies like Square, Vagaro, and Mindbody build software for a broad range of small retail and service establishments rather than one hyper-specific customer profile.
In contrast, in the emerging world of micro-SaaS, a company might hyper-specialize, developing a SaaS product exclusively for say small tattoo shops. This level of specialization allows them to serve a niche market better than any generic solution possibly could.
For example, our tattoo shop product might track ink inventories, provide a database of cool designs from other shops, and allow artists to easily create and manage online portfolios. They could do all of this in addition to the basic scheduling and POS functionality of more generic platforms. This means their product better meets the needs of this very specific customer profile and will likely win market share from the larger generic solutions.
However, the market is likely so small that it doesn't align with the traditional power law model of venture capital.
A quick Google search suggests that there are roughly 25,000 tattoo shops in the US. Let’s assume the average tattoo shop can afford to pay maybe $50 a month plus some sort of payment processing rev-share. Based on this, we can ballpark an ACV of maybe $1,000-$3,000 a year (depending on payment volumes and payment rev-share structure). That means the entire US market for tattoo shop software is (very roughly) somewhere between $25M and $75M.
Implications for Venture Capitalists
Micro-SaaS companies typically require much less capital and often embrace lean startup principles, making them financially self-sufficient much sooner. And by definition, they target smaller markets. As our example illustrates, their TAM might be in the tens of millions.
This presents a big challenge for VCs, who are essentially in the business of selling money to startups. Not only do these types of software businesses require much less capital in the first place, but they also don’t have large enough markets to meet the underwriting requirements for traditional VC. Tell a VC that you’re building a company to pursue a $50M TAM and they won’t even listen to the rest of your pitch. Almost all VCs today have optimized their funds and return models around pursuing much larger markets with TAMs in the tens of billions.
So let’s assume I’m right, and we begin to see a shift towards more and more micro-SaaS. This means a future with a lot more small tattoo shop software companies and a lot fewer large SaaS companies like Square. That means a lot less investment opportunities for VC firms. According to Dealroom.co, 45% of all venture capital invested in 2022 went to SaaS business models (ref). If micro-SaaS becomes the new normal, then most of those VCs are going to have to find new places to deploy their capital.
Accelerating The Pace of Science & Engineering
AI tools are not only facilitators in the business world; they are transformative agents in the realm of scientific research. Language Learning Models (LLMs) like ChatGPT are already making waves in the academic community. These models can summarize extensive bodies of research, enabling scientists to quickly grasp developments in other fields that could have a bearing on their own work. This cross-disciplinary understanding is crucial in an era where the boundaries between different scientific domains are increasingly blurred.
Beyond summarization, AI's capabilities in data analytics and predictive modeling are groundbreaking. For instance, AI-enabled platforms can sift through massive datasets from complex experiments, identifying patterns or anomalies that would take human researchers an inordinate amount of time to discover.
This accelerates not just data interpretation but also the entire scientific process, from hypothesis formation to experiment design and result analysis.
Moreover, AI can go beyond data analysis to suggest new avenues of research based on existing literature and data. Imagine a machine learning model that can read thousands of academic papers and generate hypotheses for unexplored areas in, say, quantum computing or gene therapy. This isn't science fiction; it's a near-future application of AI that could significantly speed up the pace of scientific discovery.
Startups at the Intersection of Science & Technology
Startups stand to gain immensely from this accelerated pace of discovery. Whether it's turning a breakthrough in renewable energy into a scalable solution for global power needs or leveraging advancements in material science to create stronger, lighter, and more sustainable consumer products, the opportunities are boundless. Startups that commercialize these types of technologies, those leveraging advancements in the hard sciences and engineering, are colloquially referred to as “deep tech” startups.
We can imagine a future where researchers using AI tools are able to develop better designs for wind turbines or solar panels, making them more efficient and cost-effective. Startups that can quickly commercialize these AI-driven designs will not only contribute to sustainability but also tap into multi-billion-dollar markets.
Similarly, in the field of material science, AI could help in the discovery of new materials with unique properties—be it superconductors, lightweight alloys, or biodegradable plastics. Startups that can bring these materials to market could revolutionize entire industries, from automotive and aerospace to consumer electronics and healthcare.
Again, all of these are multi-billion-dollar markets. Just the type of markets VC funds are optimized to pursue.
The Deep Tech Opportunity
As scientific discoveries become more frequent and impactful, the range of investable “deep tech” startup ideas will expand correspondingly. For venture capitalists, this represents a new frontier of opportunities that can also provide the large markets they need.
A 2021 report from Boston Consulting Group found that deep tech investments increased 300% from 2016 to 2020 ($15B to $60B) and, citing a study from MIT, estimated north of $77B invested in “advanced tech startups” in the first 8 months of 2021 (ref). So, for easy math, let’s assume $100B of global venture capital was invested in deep tech-related companies in 2021. KPMG estimates the total value of all venture investments in 2021 to be $671B (ref). This implies that deep tech only accounted for roughly 15% of all venture investments in 2021. However, as BCG pointed out, it was growing faster than the overall venture markets and increasing its percentage of the total share of investments.
We have to keep in mind that 2021 was the peak of a bubble period. So you probably had some “tourist investors” (those that don’t normally invest in a given sector) that were investing in deep tech because of the exuberance of the period. That could have caused deep tech to grow faster than the overall venture markets. However, sustainable deep tech investing requires a different set of skills than software investing. In addition to business and finance fundamentals, you need some ability to assess the technical capabilities of a team and to judge the direction of scientific and technological advancements at a macro level.
My prediction is that in the short term we’ll likely see a drop off in the percentage of venture capital deployed in deep tech (compared to the 2021 comp set). This will be driven by those “tourists” pulling back to more familiar ground given the more difficult investment landscape.
Over the longer term though, I think more and more funds will move into deep tech as their primary focus. If automation has the impact I predict, then it will decrease opportunities for VCs to invest in traditional SaaS companies, even as the total number of SaaS companies explodes. Accordingly, I predict that more funds will pivot towards deep tech, recognizing it as a fertile ground for high-impact, high-reward investments that can sustain the larger funds and power law investment philosophies that VCs have come to rely on.
Over the last decade, deep tech investments represented a small fraction of all venture investments, maybe 10-20%. My prediction is that within 10 years it will be a majority of all venture capital investments.
Conclusion
The bursting of the venture bubble in 2022 isn't a mere market fluctuation. It coincides with a pivotal shift in technology—the dawn of the AI era. This confluence of factors will have three major impacts on the startup and venture landscape:
Software companies will become exponentially easier and cheaper to build, reducing their reliance on and suitability for traditional VC funding.
The rate of scientific discoveries and technological innovations will accelerate, offering new avenues for venture capital.
As a result of the previous two points, the amount of total capital available to fund deep tech companies will explode in the coming years.
These changes signify a fundamental transformation in the venture capital and startup landscape. I believe that deep tech, once a minor player, is poised to become the majority of all global venture capital investments within the next decade. This increase in capital availability will fuel a new generation of startups that look more like SpaceX or Altos Labs than Square or Stripe.
As we navigate this transitional period, both startups and investors will have to adapt their strategies to thrive in this new reality. However, I think it will be society itself that stands to gain the most from this shift. Perhaps the future that many of us have hoped for, one with flying cars and robot nannies, is finally ready to take shape.
