Full stack startups, revisited
About a decade ago, Chris Dixon and Balaji Srinivasan popularized the idea of “full stack startups.” The premise was that as software moved from an independent sector to natively integrated in the entire economy, the best opportunities were to build end-to-end experiences and bypass incumbents. Don’t just sell in, but compete head on.
Coming out of that era, there were wild successes like Tesla, Bitcoin, Uber, and Airbnb. But there was also a whole host of companies and ideas attempted that didn’t pan out. Some of those cases were pure mismanagement. Some failed to escape the base rates of the sector when it came to growth or valuation expectations. Competitive advantage in sectors like insurance or logistics often stems from traditional economies of scale, rather than software or technology leverage. The venture capital blitzscaling model was often a square peg meeting a round hole. I understand why investors and founders have skirted away from some of these opportunities in recent years.
But, maybe we were still just too early. I loved the sentiment behind this recent tweet from Immad Akhund, founder of Mercury. In 2024, we funded several ambitious projects at Haystack – founders re-thinking industries like biomanufacturing to funeral homes to residential housing from the ground up and from first principles.
What continues to change? What does the ingredient set look like today for founders interested in going full stack? What quivers in their bow do founders have today they didn’t have 10 years ago?
AI: There is incredible interest in how AI will transform the product experience for the end user, but I’m intrigued by how it changes the operating structure of large organizations, acting “under the hood.” Whether it’s using LLMs to generate internal documentation to audio assistants helping with customer support to multi-agent prompting acting as a substitute for market research, there is a lot of room for experimentation. We already have some consumer services businesses in our portfolio experimenting with LLM-based workflows behind the scenes.
Robots: We’re further away from widespread robotics adoption, but the encompassing advancements in LLMs acting as “robot brains,” world understanding models, and work from companies like Physical Intelligence and Skild wildly surpass any state of the art of a decade ago. Robots truly help us tackle the physical world, and I’m most interested in areas where they act as a substitute for dangerous and dirty work.
Regulatory muscle: Silicon Valley is *significantly* more politically engaged in DC and elsewhere than it was a decade ago. Much of this was in recent years, but the growing engagement with regulatory and political bodies mirrors the development of any American industry. See automakers, banks, airlines. SpaceX and Anduril’s success partially stems from their in-house human capital that understands how to work with and navigate government agencies and regulatory bodies. The knowledge base is building up in the ecosystem to then disseminate to the next generation of full stack companies.
Capital: There was bubble talk 10 years ago. Sam Altman famously made a bet that there wasn’t. And ironically, the size and scope of the venture capital industry back then pales in comparison to today’s landscape. There are large (large!) venture capital firms with deep pockets continuing to put billions of dollars to work every year. Not just equity either. This could be semantics, but many of these traditional venture capital firms are turning into “technology and innovation companies” with a host of financial products like debt, buyouts, revenue-based financing, etc. They are eager to fund capital-consumptive opportunities with innovation at the core. Much of their newfound business model depends on it.
The right combination of these factors give founders increasing opportunity to find leverage in their business models, build product and distribute in formerly-permissioned environments, and accumulate market share economically for their teams and investors.
Why am I personally interested in all this? Why do I want more full stack startups to exist? The cost of software development continues to come down (thanks Chris!), and the odds a YC team pivots into one’s space and competes over the weekend increase. The minimum marketable product in financial services, energy, agriculture, space, etc. simply cannot be built over a weekend. I like businesses that avoid competition, and the full stack approach also self selects for founders with a certain level of authenticity, insight, and aspiration.
Tomorrow’s technology companies have the opportunity to contribute to growth. Not growth in terms of MoM, YoY, ARR, etc., but in terms of real GDP growth. True economic growth. And that will only come from attempts to reshape large parts of the economy.
If you are dissatisfied with the status quo in some way or want to revisit a full stack startup idea, I would love to speak with you.