Vertical Integrators: Part IV
Prize x Plan x Probability, or Why Vertical Integrators are Mispriced
Welcome to the 296 newly Not Boring people who have joined us since last week! If you haven’t subscribed, join 232,225 smart, curious folks by subscribing here:
Today’s Not Boring is brought to you by… Attention
Attention is an AI-powered call recorder trusted by sales teams at industry giants like Snowflake, Datadog, and Stripe. Simply put: you can close more deals with Attention.
Attention’s AI captures and analyzes every sales conversation, revealing how and why your deals succeed. It not only identifies winning strategies for your team but also empowers you to act on them.
With Attention, you can automate key workflows based on customer interactions: generate personalized follow-up emails after calls, populate your CRM with crucial deal information, alert stakeholders to potential churn risks, create coaching scorecards for your sales reps, send weekly executive reports on deal outcomes, and more.
Unlike other tools that merely provide insights, Attention delivers both insights and the actions needed to close more deals. Sales is hard. Attention makes it a little easier.
Hi friends 👋,
Happy Tuesday! Tough Eagles loss last night, but no better way to brush it off than with the final installment of the Vertical Integrators series.
We’ve covered a lot in this series. If you haven’t yet, check out Part I, Part II, and Part III. To summarize: I think we’re at the beginning of a new Techno-Economic Paradigm. Vertical Integrators that figure out the right combinations of new technologies to deliver better, cheaper products than incumbents at higher margins will be the most valuable companies of this period. Doing that is really hard.
Today, we’ll look at what it might take and why I think Vertical Integrators are among the most mispriced assets in the world.
This is the last theoretical piece in this series for now, but I expect I’ll be writing about and investing in Vertical Integrators for years to come.
Let’s get to it.
Vertical Integrators: Part IV
(Click 👆 to read the full piece online)
Confucius says, “Study the past if you would define the future.” Churchill adds, “The farther backward you can look, the farther forward you are likely to see.”
So let’s go back to 1878, when, from inside of his Menlo Park, NJ lab, Thomas Edison worked to light up New York City.
At the time, gas lights dominated the city. Coal gas was produced by heating coal in the absence of air, then piped from central gas works to individual lamps on streets and in homes and offices, where the gas was ignited in the burner and the ensuing, flickering flame lit up previously dark or candlelit spaces, protected by glass globes.
Gas lighting was a mini-miracle: humans were no longer bound by the sun’s movements. But gas lighting had its issues: open flames were a constant fire hazard, leaking gas could cause explosions or asphyxiation, incompletely combusted gas produced carbon monoxide, indoor air quality suffered, lights flickered, and both the protective globes and the rooms they were meant to protect accumulated soot. There had to be a better way!
And there was. Forty-seven years earlier, in 1831, the English scientist Michael Faraday invented the first electromagnetic generator, the Faraday Disc, which produced electricity from mechanical motion. It was both groundbreaking and impractical for commercial use. In 1867, German inventor Werner von Siemens developed the dynamo, a more efficient direct current generator capable of producing electricity on a commercial scale.
In the 1870s, Belgian electrical engineer Zénobe Gramme improved on Siemens’ design with his Gramme machine, the first direct current (DC) generator suitable for industrial applications. And in 1875, according to Carlota Perez, the Age of Steel, Electricity, and Heavy Engineering was born. It would be brought to life by Vertical Integrators, including Edison.
In 1878, Edison, already a famous inventor, turned his considerable talents towards electric light. In Empires of Light: Edison, Tesla, Westinghouse, and the Race to Electrify the World, author Jill Jones describes the need for vertical integration to pull this new paradigm into being:
Each component of the new lighting network had to be designed, tested, redesigned, and retested. “Everything is no new that each step is in the dark,” Edison said. “I have to make the dynamos, the lamps, the conductors, and attend to a thousand details that the world never hears of.” The truth was, this was completely new and complex technology, and no one really knew how long it would take to make it work or what it might ultimately cost.
Edison’s attempts to light up New York highlight both the challenges and the advantages of vertical integration.
Edison himself did not, contrary to popular belief, invent the lightbulb. He and his team did, however, make light bulbs commercially viable. The light bulbs they began working with, using platinum or carbonized paper filaments, lasted only an hour or so before burning out. Over the course of dozens of experiments, they discovered that a cotton fiber filament could last a workable fourteen and a half hours, and by the time New York City’s Tammany aldermen visited the lab in December 1879, Edison showed them a bamboo-filament light bulb which “should last six months with normal use.”
Of course, there was no established supply chain for bamboo-filament light bulbs, and “the directors of the Edison Light Company would not go into manufacturing. Thus forced to the wall,” Edison recalled, “I was forced to go into manufacturing myself.” Explaining further, he told one of his investors, “Since capital is timid, I will raise it and supply it… The issue is factories or death!” By the end of 1880, the Menlo Park light bulb factory, controlled and financed by Edison himself, “was turning out several hundred bulbs daily.” As Jones noted, “the nature of the new and little understood electrical science and its many unknowns were dictating these first groping corporate arrangements.”
Light bulbs were just one piece of the electric lighting system, which involved power generation, distribution, wiring, and finally, the illuminated bulb. Each piece of the system had been proven to be technologically feasible, but it was up to Edison to make the entire system commercially viable.
The whole story is one of systems design, of making trade-offs here so something might work a little better there, like the story of Base Power Company in Part I. While Edison made many such trade-offs, one passage stands out (emphasis mine):
Always hanging over Edison’s head was the cost of copper, the biggest and most daunting expense for his central station plan. The cost of copper had inspired his invention of the high-resistance light bulb. But even that huge savings would not be enough if he was to match, let alone undercut, the cost of gas lighting. In the summer of 1880, even as the “subways” were being laid out again and again, Edison had one of his profound breakthroughs, coming up with a “feeder and main” system of distribution that mimicked – under the streets – the new parallel circuits to be used within buildings. This brilliant new approach would cut copper costs to one-eighth the earlier estimates.
When asked why no one else had thought of this simple and elegant solution, Glasgow physicist Sir William Thompson replied, “The only answer I can think of is that no one else is Edison.”
Thompson was right, not because of Edison’s singular genius, but because Edison was the only person in the world who had the entire system in his head, who was thinking through distribution while having just designed the “parallel circuits to be used within buildings” and the high-resistance light bulb created to reduce the cost of copper in the system, and who was on the hook for delivering electric light at a cost competitive with gas lighting.
With the system designed, Edison turned his team to production and implementation. His general, John Kruesi, manufactured and installed “(beneath some of Manhattan’s busiest, filthiest thoroughfares) fourteen miles of underground distribution cables and wires.” Charles Dean ran Edison Machine Works, which manufactured generators by the East River. Francis Upton produced a thousand light bulbs a day back at Menlo Park. “These three enterprises,” Jones writes, “were all organized and financed by Edison or his closest associates.”
In August 1881, Edison bought a building in a grimy neighborhood at 255-57 Pearl Street for $65,000, where “home-devised and home-made” devices – “Edison dynamos – powered by coal-fired steam engines – produced an initial alternating current electricity that was then gathered from the machine by ‘commutators’ and brushes and turned into direct current.”
All the while, complications led to delays, as Edison’s investors, the newspapers, and the public became increasingly frustrated. But on September 4, 1882, after “four years of hard work of the most original, difficult sort and almost $500,000,” John Lieb threw the circuit breaker at Pearl Street and 400 incandescent light bulbs “glowed softly to light.”
The delays, costs, and grumbling were forgotten in the flip of a switch. Edison was a hero, and a rich one. In 1889, Edison consolidated his businesses into Edison General Electric Company, and in 1892, the company merged with Thomson-Houston Electric Company to form General Electric. The combined GE had an estimated market cap of $35 million, roughly $1.2 billion in today’s dollars.
That, however, was just the beginning of the War of Currents. The next year, Westinghouse won a victory for Nikola Tesla’s Alternating Current (AC) electricity over Edison’s DC when he lit up the 1893 Chicago World’s Fair. Ultimately, AC, which could travel longer distances, won out over DC.
Historians remember Edison as having lost the war. From a technology perspective, that’s true. A true Vertical Integrator, however, General Electric adopted the winning AC technology and plugged it into its system. The company leveraged its brand, size, resources, and diverse product line to invest heavily into AC and used its large customer base to introduce new AC products to the market.
In My Life and Work, Henry Ford said of Thomas Edison, “Edison is easily the world’s greatest scientist. I am not sure that he is not also the world’s worst business man. He knows almost nothing of business.”
That may be true. After merging with Thomson-Houston, Edison owned just 3-10% of the company his scientific brilliance brought to life. But the Vertical Integrator he created won. Despite Tesla and Westinghouse winning the technological war, General Electric remained the more valuable company throughout the Electric Age.
Study the past if you want to define the future.
Today, we are entering a new Techno-Economic Paradigm, like the one that Edison helped bring to life in the late 19th Century. General Electric was one of its biggest winners, because at a time when vertical integration was needed to figure out how to deliver new products to take advantage of new technological capabilities, it was a textbook Vertical Integrator:
Integrate multiple cutting-edge-but-proven technologies.
Develop significant in-house capabilities across their stack.
Modularize commoditized components while controlling overall system integration.
Compete directly with incumbents.
Offer products that are better, faster, or cheaper (often all three).
Peter Thiel argued that there were only two ways to make money coming up with new things: software and “complex, vertically integrated monopolies.”
While most modern founders and investors remain tied to the lessons learned over the past half-century of software development, if I’m right, they should be studying people like Rockefeller, Ford, and Edison to prepare for what’s coming. The farther backward you can look, the farther forward you are likely to see.
Edison faced many of the challenges we discussed in Part III that modern Vertical Integrators face today:
It requires building at least two companies.
Hardware by itself is just really hard.
Incumbents are good at what they do and have advantages.
Figuring out the right level of integration is not obvious.
You can’t pivot nearly as easily.
Something better might come along.
Getting to revenue takes a lot of money.
I won’t spell them all out here, but if you’d like, read back the story above and see if you can find examples of each. Even Edison’s pivot to AC, which looks smooth in hindsight, required merging with Thomson-Houston, which was further along in AC. “Plus ça change,” as someone at Edison’s Compagnie Continentale, where a young Nikola Tesla would get his start, might have said.
But as I wrote in Part II, “If the Vertical Integrator can pull all of that off – and often much more – all of those challenges become a moat, and advantages compound.”
That was certainly the case for General Electric, whose market cap peaked near $500 billion in 2000, and spent the better part of the mid-90s to the early 2000s as the most valuable company in the world.
That “if” is the rub, though. It’s the Great Filter that Meter’s Anil Varanasi told me about. If the Vertical Integrator can pull it off.
So today, in the last installment of our Vertical Integrator series (for now), we’ll discuss what it takes to build a successful Vertical Integrator, the advantages they have, and the size of the prize that awaits on the other side of the Great Filter.
What it Takes
Inspired by SpaceX, Tesla, Anduril, and, if they’re students of history, companies like GE, Standard Oil, and Ford, hundreds of startups are now building physical products, and investors are warming to the idea of funding them, at least at the early stage.
What will separate the next General Electrics, Standard Oils, and Fords from the inevitable mass of failures?
The tricky part about answering this question is that there are no playbooks or agreed-upon metrics that apply across industries as diverse as energy, defense, transportation, mining, and healthcare. This is not SaaS.
Valuing a SaaS business is a relatively straightforward exercise. The metrics are well-understood and well-trodden.
“This business is generating $75 million in ARR growing at 125% YoY with a NRR of 112% and 83% gross margins.” Compare that to other companies that have performed similarly, map their relative trajectories, maybe add a smidge here or take away a smidge there based on the founder quality or market size, apply some reasonable multiple, and come up with a number.
It’s all right there for you in the Bessemer Emerging Cloud Index and in the many proprietary datasets built up over the decades venture capitalists and public market investors have been investing in SaaS.
There is no Index for Vertical Integrators. That is a good thing. Indices are the enemy of alpha.
If valuing SaaS businesses is now a science, valuing a Vertical Integrator is still an art.
It’s a very different exercise that ultimately comes down to the size of the prize multiplied by the quality of the plan and the probability that the team can execute it.
The size of the prize is relatively obvious. The plan can be analyzed, with particular focus on potential unit economics. But probability of execution?
On that question, reasonable minds may differ greatly. Many reasonable minds are too skittish to even venture a guess.
This is why I think that early stage Vertical Integrators are some of the most mispriced assets in the world until they begin to deliver on their promise.
It’s like pricing options, but with very little data. Will a company that can produce electricity more cheaply than anyone in the world be worth at least $100 billion? Of course. Can they actually do it? Ah, that’s the question.
If you break it down, there are two pieces of the analysis:
The plan.
The team’s ability to execute against the plan.
The plan is part strategy, part Techno-Economic Analysis.
On the strategy side, there are questions to answer like: Why now? What is the bottleneck preventing incumbents from making their products better, faster, cheaper? Which technologies will you employ to make your product better, faster, cheaper? What prevents incumbents from responding? What are the string of coherent actions your company will undertake to build a system that delivers a better, cheaper, faster product in a way that others can’t mimic? Who will you need to hire to bring the plan to life?
The techno-economic analysis (TEA) is a combination technical-financial model that uses everything from component costs to efficiency curves to figure out the potential costs of a system and help make trade-offs.
If Edison had put together a TEA, for example, he would have included the cost of copper, the amount of copper needed at different light bulb resistance levels, and the amount of copper needed in wiring his distribution network, among other things. Certainly, something like a TEA helped him decide that copper costs so heavily impacted systems costs that it was worth designing a high-resistance bulb, despite trade-offs including: reduced bulb lifespan, lower electrical efficiency, and increased manufacturing complexity.
Both the strategy and the TEA boil down to this: is it possible to build a new system, starting from scratch, that can deliver a product that is better or cheaper enough than existing products that you can win a competitive market against better-resourced incumbents?
In traditional startup lore, there’s this idea that you need to build a 10x better product to displace the incumbent. With Vertical Integrators, the question is whether you can build a product with 10x better gross margins.
If you can, you have options: lower prices to steal the market, charge the same price for a significantly better product and capture the margin, invest margin back into R&D to build a much better product. The right answer will depend on the situation.
Margin gives you room to operate. Unit economics cannot be an afterthought.
On paper, credible Vertical Integrators should show significant gross margin improvements over incumbents, which really should be possible. The point of technology is to do more with less.
Plus, in large markets worth attacking, incumbents will likely have competed each others’ margins down to the bones in the industry’s modularization phase. Remember: they’re not dumb. They’re operating at the lowest cost they can without burning their whole system to the ground and starting over.
The Vertical Integrator has nothing to burn; it can simply start from scratch and choose the right ingredients from the smorgasbord of modern technology, assuming it has the capital and people to build everything it needs to build.
The margin advantage won’t always look like that chart above. Sometimes, the margin will come from being able to charge higher prices for a much better product delivered for roughly the same cost. Sometimes, a la Ford, the margin will actually be just as thin but your price will be lower, which means you’ll be able to do more volume. In any case, there needs to be a plan to produce a product in a large market at higher margins than incumbents can.
All of that is on paper, though. It comes down to the team’s ability to execute in reality.
The team’s ability to execute against the plan is the harder, and more important, thing to judge.
From Confucius and Churchill to Tyson. “Everyone has a plan until they get punched in the mouth.” And Robert Burns: “The best-laid plans of mice and men often go awry.”
Here we are roughly 22,000 words into this series on Vertical Integrators and I’m about to write something incredibly anticlimactic: it all comes down to the founders.
This is a trope in venture capital, and frankly, it sounds cheesy.
At first it seems simple, so simple that you want to dismiss it in favor of something more unique, bespoke, or complex. You’re paid to analyze and invest in businesses, so you do market maps and TAM analysis and pick apart the strategy and dive deep into the technology, compare it to other technologies, map the competitive landscape, track traffic metrics, model out a base case, bull case, and bear case, trade notes with other investors, tap your expert network for their insights, and do any number of things that makes it feel like the investment decision is based on something substantial. You pick apart everything I just wrote about in The Plan, question this assumption or that. You come up with smart-sounding reasons that This Will Never Work.
The thing is, though: a truly excellent founder has done all of this, and more.
They’ve done it to a level of depth and over a period of time that you will not be able to match. They’ve done it not to decide whether to invest 5% of their pool of other peoples’ money, but to determine how they are going to spend the next decade or more of their lives, when, as a truly excellent founder, they have a plethora of very good options competing for that prime time.
Almost everything else in venture capital follows this circular logic to some extent.
Truly excellent founders are a compression algorithm.
If you judge that a founder is truly excellent – has the potential to be an Edison, Rockefeller, Ford, Jobs, Huang, or Musk, which sounds outlandish until you remember that theirs are the magnitude of company we’re after – that tells you almost everything else.
Most importantly, it will tell you whether they’re able to hire the team of people who are the best in the world at the particular thing they do that the company needs. Team is the only determinant of ability to execute.
You will not get certainty early in the life of a Vertical Integrator. You won’t get early signs of product-market fit. You won’t get growth and retention curves. You’ll just get a plan, one that probably sounds a little crazy, and some people. It’s up to you to decide whether to bet on their ability to execute.
Which raises the obvious question: what makes a truly excellent founder?
There is no simple answer. Like porn, you know it when you see it.
Sometimes, all the things that make a founder great are present in one person, and sometimes, they’re present in a combination of co-founders. You can read the great Tablet profile on Palmer Luckey, the Isaacson books on Elon Musk or Steve Jobs, the recent Fortune article on Meter’s Varanasi Brothers, any of the profiles of the Collison Brothers, or listen to the Founders Podcast episodes on history’s greatest founders and come away with different ideas of what a truly exceptional founder looks like.
As I was writing this, I came across this paragraph in Empires of Light:
Unlike Edison, who launched his every new invention with maximum public ballyhoo, George Westinghouse introduced his lighting revolution very quietly. No press account heralded the tremendous breakthrough of AC, the bursting of the DC shackles. The AC revolution began quietly and almost invisibly.
Two truly exceptional founders competing in the same industry at the same time in the same part of the world, and on the surface, their approaches are diametrically opposed. There is no playbook or right answer. That’s the beauty.
That said, there are some things that seem to be present in most great founders:
Ability to attract and pull the best out of world-class people.
Deep curiosity and bottomlessly nuanced knowledge.
Bias towards action and iteration.
Resilience and adaptability.
Storytelling and fundraising.
Tireless work ethic and energy.
Real commitment to a larger mission and desire to solve a big puzzle.
Ability to hold the whole system in their heads, from 30,000 feet to one inch.
There’s something else about Vertical Integrator founders I’ve noticed that is counterintuitive: they often don’t come from within the industry they’re trying to reshape.
Most haven’t studied one particular technology in school that they’ll apply to the new problem. They’re definitely technical – at least one person on the founding team has to be able to hold the whole system in their heads – but they’re more generally technical, in a “that guy can figure anything out” kind of way.
This actually makes a lot of sense. Vertical Integrators need to figure out an entirely new way to do something that’s been done a certain way for years, and they need to do it with whichever technologies and processes work best as part of the entire system. Being wedded, or even biased, to one way of doing things or one technology dulls the advantage of starting fresh.
There are exceptions to everything I’ve just written, of course. There are common traits I’m missing, and the actual effect of the combination of those characteristics in a person is hard to describe.
If I had to sum it up, it’s something like: they make world-class people want to do their best work for them. That’s a hard thing to capture in words or lists.
All of which is to say: underwriting Vertical Integrators is so difficult because to a large extent, it comes down to judging people against the Herculean tasks they’ve set for themselves based on clues but no certainties.
This is true to an extent for every startup investment. But the challenges are compounded with Vertical Integrators for a few reasons:
These are harder, more complex companies to build than typical startups.
The thing the founder is proposing has never been done.
They may require decades-long commitments to come to fruition, with no easy outs along the way.
They require more capital upfront to even have a shot, which means you need to decide before seeing how all of the pieces fit together in practice.
There are some early signs. Maybe they’ve hired incredible early talent. Maybe they have some LOIs. Maybe they’re incredibly responsive. Maybe they have sterling references. Maybe they’ve pulled off Herculean tasks before.
The further along the company gets, the more proof you have that your gut – that this might really be a truly exceptional founder – may have been right. The further along the company gets, and the more they’re able to accomplish with limited resources, the more obvious it becomes that someone people may not have thought was a truly exceptional founder actually is.
At some point, with enough proven, the valuation snaps violently into line with the size of the opportunity. It flips from deeply improbable to somewhat probable, even obvious. SpaceX and Anduril have more demand for their shares than they can meet, and enough capital to do what they need to do.
The trick is figuring it out before it becomes obvious, when the capital is really needed.
And at the beginning, you have a plan and the people behind it. That’s it.
It’s not a coincidence, then, that the most successful Vertical Integrators to date were founded by Palmer Luckey (and four specifically skilled co-founders) and Elon Musk. Each experienced previous success. Each brought a nine-figure personal fortune to bear on his new quest. Each assigned a delusionally high probability to his own success before the market believed that what they were trying to do was even possible. Neither needed outside capital to get going.
In the earliest innings of the Techno-Industrial Revolution, that’s what it’s taken. As Trae Stephens and Markie Wagner wrote in Choose Good Quests, “These future-defining problems are hard to recruit for, difficult to raise money for, and nearly impossible to build near-term businesses around, which is why they are exactly the types of problems we need the most well-resourced players pursuing.”
Now, though, encouraged by the success of those companies, and often trained within them, there are founders who are not financially well-resourced enough to put the first couple of rounds on their Black Cards. Which means investors need to figure out which people have what it takes early enough to give them the resources they need.
As you may have picked up, there’s no surefire way to do that. It’s a bet with long odds. But I think that these bets are worth taking, because if the team is able to do what it says, vertical integration has serious advantages.
The Advantages to Vertical Integration
The world faces a number of very big challenges. It always faces a number of very big challenges. We need food, energy, shelter, communication, transportation, healthcare, and materials with which to build new things.
Humans are constantly trying to figure out new and better ways to meet those needs. One generation’s solutions work for a time, until they are outgrown or atrophy. Then entirely new solutions to very old challenges are needed.
This is a never-ending game: the population grows and each person’s desires grow. More than eight billion people demand comforts not even one king could attain two centuries ago. Meanwhile, the companies built to serve those demands under one Techno-Economic Paradigm, struggling just to fight off entropy, cannot serve them in the next.
It’s up to entirely new companies to create entirely new solutions to those ever-present challenges.
In the first three parts of this series (Part I, Part II, and Part III), we’ve discussed that, after a half-century of solving low-hanging problems with software, the biggest challenges today require physical solutions. Software alone won’t solve them – either directly or by patching up leaky incumbents. That much is becoming obvious. Less clearly, selling components to incumbents won’t solve them, either, even if they are atoms-based and therefore in vogue.
In Full-Stack Deep Tech, Cantos’ Ian Rountree explained why these businesses don’t work:
And yet we see so many deep tech startups that aim to sell equipment, components, or services into slow-moving, often thin-margin industries. I don’t blame them. It’s a less intimidating, apparently more capital-efficient path to market. For these reasons I’ve backed a number of these endeavors over the years… but it’s almost always been a mistake. You raise your Seed round and sometimes a Series A on a pilot, LOIs, purchase order, or from strategic investment from a major incumbent. But then they never move as quickly as you and your investors need; they drag out the pilot, move the goalposts, or management changes and you’re forgotten altogether.
In order to solve the challenges facing humanity today, for now, we need vertically integrated companies that aim to build entirely new systems and compete head-on with incumbents.
We are entering what Carlota Perez would call a new Techno-Economic Paradigm, one defined by the combination of bits and atoms, and new Techno-Economic Paradigms both demand and reward Vertical Integrators.
While vertically integrating in an effort to topple centuries-old incumbents seems insane, it’s the only way to really get the job done and pull humanity forward.
Vertical integration allows for tighter feedback loops. Hadrian has software engineers working next to machinists in order to build better software and automate what can be automated. Radiant built a digital twin of its reactor to run simulations as it gets data from its test reactor. Earth AI drills its own targets to quickly learn which are promising. Primer can adjust its curriculum to individual students instead of waiting for Boards of Education to agree on new curricula for all students. In each case, the companies get feedback, and can iterate towards better systems, much faster than they could if they outsourced core capabilities or had to wait to sell their product to slow-moving incumbents.
Vertical integration allows for trade-offs at the system level as opposed to the product level. Base Power Company makes decisions that you might not make if you were just manufacturing battery packs to sell to customers in order to make packs easier to install. They call the whole process the Deployment Factory. That gives them more control of the overall system than they would have buying packs from third-parties or selling them to customers.
Vertical Integrators ride exogenous technology curves to build products that naturally get better over time. Base will take advantage of improvements in cell chemistries, and in the increased demand for their product from increased renewable (variable) power generation and electrification. Fuse benefits from the improved performance of the capacitors it relies on for its pulsed-power generators.
Relatedly, they don’t face existential risk from a single new technology. Their value is in the system, not in any one of the pieces. A competitor with a better technology would have to build a vertically integrated system around that technology in order to compete. Sometimes, the Vertical Integrator might simply be able to plug in that new technology to improve the performance of the whole system. Think of the distinction Sam Altman made on 20VC about building products with AI that new models might crush versus those that benefit from improved models.
Vertical Integrators can control their own supply chains and bring their own core technologies down those curves. Radiant plans to manufacture nuclear reactors in factories, hopefully bringing reactors onto the learning curves typically experienced by mass manufactured products. Fuse, while benefiting to date from improved capacitor performance, can manufacture pulsed-power generators both for itself and to sell to other fusion companies in order to produce more, bring down costs, and improve performance en route to commercial fusion. Like Amazon with AWS, these companies can be their own first and best customer.
Vertical Integrators don’t have to do anything unproven to build better products than exist. Deep tech companies often require doing something that’s never been done before, before running out of money, and then scaling that thing in order to unlock value. Vertical Integrators need to combine a number of very hard things that are at least each possible. They take on engineering and system risk, but not scientific or technical risk.
Vertically integrating means better margins, if it works, at scale. A tightly-managed manufacturing process can bring down costs. Better, there are simply fewer external parties to split margin with the more you vertically integrate, and you are in more control of your margins than if you try to sell into legacy industrials who have more power in the relationship than you do. Study Porter’s Five Forces.
Vertical Integration allows companies to deliver products that they couldn’t if they didn’t control the full-stack. SpaceX was able to launch the largest telecommunications network in the world by owning reliable launch capabilities. Meter launched Command, a product that allows customers to understand and take action on their networks using natural language, because it controls everything from hardware to the data pipeline.
Finally, vertically integrating means competing for the biggest prizes. It means tapping into the multi-trillion dollar pools of revenue available to the companies that solve whole problems for their customers, and “striking not at the margins of the profits and the outputs of the existing firms but at their foundations and their very lives.”
Simply put, while the struggle is real, the opportunity is commensurately large. Those few that succeed will have the chance to build some of the largest and most impactful companies in history.
The Size of the Prize
One of the things that first drew my attention to Vertical Integrators was that a lot of founders who seemed to be smart and (the right amount of) sane were trying to build them. That’s usually a good early signal that something is brewing. What the smartest people will do on the weekend is what everyone else will be doing in ten years, but for a category you can’t really do on the weekend.
The thing that really grabbed me from there, before I had a framework or even knew I was looking at different instances of the same phenomenon, was the sheer size of the opportunities. I remember my first conversation with Hadrian’s Chris Power, as he explained what he was trying to do, saying something like, “If you pull this off, this is a $100 billion company.”
There’s a lot of execution between here and there, but my conviction that there will be many $100 billion startups created by delivering better products than incumbents at lower costs and higher margins has only grown since.
That’s one of the most tantalizing things about Vertical Integrators: there is a clear path to tremendously large outcomes. Not easy, but clear.
With the previous generation’s winners, you had to bet that the market would change in some dramatic way to see how these startups might become giants.
Google was one of many search engines indexing a relatively tiny universe of websites, with no business model to speak of. Amazon sold books! On the internet! When Barnes & Noble existed! Facebook was a place for college kids to look at pictures of each other, with a lot of eyeballs and no ads. Even Uber and Airbnb, with their modern approaches to established industries, required investors to expand their thinking about the market in which those companies were operating.
Their early execution and growth were hard to argue with, but grasping the size of the opportunity if that growth continued, when the numbers said that growth couldn’t continue given the market sizes of the day, might have required a heavy dose of hallucinogens.
Vertical Integrators find themselves in almost exactly the opposite situation.
The size of the opportunity is relatively clear in the revenue numbers of the incumbents they seek to replace, but their ability to execute is uncertain and early growth is practically non-existent. In the place of smooth growth curves, there are discrete gates, often years in the future, that they need to pass through to prove value.
That means that none of the old playbooks, meticulously compiled over decades of watching companies like Google, Amazon, Facebook, Uber, Airbnb, or hundreds of SaaS companies, work for building or investing in Vertical Integrators.
Which is why I think they’re the most mispriced assets in venture capital. Again, it comes down to the size of the prize multiplied by the plan and the probability you assign to the team’s ability to execute it.
Throughout this series, I’ve tried to make the case that most people are underestimating the probability, at least in aggregate.
In Part I, I argued that Vertical Integrators were distinct from either Aggregators or Deep Tech companies, that the problems left over from the Aggregator era require vertically integrated solutions, and that Vertical Integrators’ business models “should look more like the business models that won the Second Industrial Revolution.”
In Part II, I showed that there are pretty compelling signs that we’re entering a new Techno-Economic Paradigm. Historically, these new beginnings have both required and rewarded Vertical Integrators, like Edison, who can build new systems with new technologies to deliver the new products that the market needs.
In Part III, I discussed that while all of that sounds great in theory, actually doing it is fraught with very real challenges. Many startups that have tried to vertically integrate too early, or in not quite the right way, have died expensive deaths.
Today, we talked about what it takes to overcome those challenges: sound plans backed by strong potential unit economics and the rare founders and teams who can execute against those plans. We also talked about some of the advantages available to the small few who can pull it off.
The biggest advantage may be this: because it’s so difficult, and because the advantages you accumulate by doing a million hard things well form barriers, there will be very little serious competition awaiting on the other side of the Great Filter.
Of course, incumbents will be there, and they will try to compete. But if these companies are able to execute on the plans that show 10x better margins than incumbents (i.e. 50% vs. 5%) in ways that would force incumbents to abandon what got them here just to try to turn their very large and unwieldy ships in time, while public market investors scream at them for burning cash, my money is on the Vertical Integrator.
Among startups, I expect we’ll see much less competition. The companies that show an early ability to execute against a big and credible enough vision will attract the top talent and the limited pool of investors willing to back such hard-to-underwrite companies, sucking the air out of the room for would-be challengers.
There may be a mini-funding bubble once a few more of these companies break out and investors come to adjust their probabilities, but the limiting factors will remain talent and the ability to design, build, and balance the complex integrated system necessary to compete.
Plus, the first startups that are able to deliver will build trusted brands that partners can count on. All else equal, would you trust your valuable payload to a launch company not named SpaceX? All else equal, would you put the second best advanced nuclear reactor in your community?
All of this points to the idea that there will be some truly massive vertically integrated companies built in the next couple of decades. I would not be surprised to see at least a dozen companies founded in the 2020s and 2030s reach trillion-dollar market caps by 2040 or 2050. Tech has barely started to touch the very big sectors of the economy.
There are too many trillion-dollar industries that are struggling to keep up or improve, and too many truly exceptional founders with too many powerful new technologies at their disposal, for the world not to look much different in a couple of decades than it does today.
Energy, manufacturing, defense, transportation, mining, construction, agriculture, and healthcare – each a multi-trillion dollar industry – is ripe for creative destruction. If anything, better products at lower prices will only expand those markets, and higher margins will make the companies that win them more valuable than the incumbents they replace.
Even labor, largely unaddressable by technology to date, is ripe for Vertical Integration. At some point, robots will perform much of the backbreaking physical labor done by humans today, and it seems clear that AI will take on a good deal of today’s knowledge work.
While we’ve explored vertical integration in the context of companies building physical products, I suspect that vertical integration – internalizing whatever necessary to fully do the job for customers – will touch all industries.
If you look at what Ramp is building towards, for example, it’s agents that do much of the work that finance teams do today. There will be agents for each department - sales, engineering, design, etc… – each a product that does a job better, faster, cheaper. Similarly, I suspect we’ll see whole services firms – accounting, legal, consulting – vertically integrated around AI-native capabilities that have significant cost advantages over incumbents, and whose advantages grow with scale and data. YC’s Jared Friedman called them “vertical agents” in a tweet last night.
If you think about what an agent or AI-first firm might be able to do in five or ten years, the companies that build the best ones look a lot like Vertical Integrators:
Integrate multiple cutting-edge-but-proven technologies.
Develop significant in-house capabilities across their stack.
Modularize commoditized components while controlling overall system integration.
Compete directly with incumbents.
Offer products that are better, faster, or cheaper (often all three).
In short, in a new Techno-Economic Paradigm, every industry will be reshaped, and they will be reshaped by Vertical Integrators. Right now, at the end of an old TEP and the beginning of a new one, things are uncertain, even scary.
But history says that humanity will experience wonders that even I, as optimistic as I am, can’t imagine. The period between the dawn of Perez’s first TEP – the Industrial Revolution – through the beginning of the last – the Information Age – saw the invention of miracles that would have made an early 18th century person’s head explode.
If I’m right, after half-a century of incredible progress in the world of bits, the combination of bits and atoms, brought about by Vertical Integrators, will blow our minds, too.
We will see cheap, abundant energy, cures for our most deadly diseases, supersonic flight, robots that do our chores, cars that drive us, homes people can afford, adventures in space, and much more besides.
But are these companies that create these miracles venture backable? I think they’re among the only companies that are.
That may be a controversial thing to say, when new AI-powered SaaS companies are hitting $10s of millions of revenue in the first year while Vertical Integrators take years and $10s of millions of dollars to get to first revenue, but it’s important to remember that it takes 7-10 years, at a minimum, to get to IPO.
One question you need to ask is: with everything changing as fast as it is, do I see this company surviving and continuing to grow a decade from now?
That’s a hard question to answer for companies that catch fire early and attract waves of competitors with the tools to build products in days that would have taken months.
Vertical Integrators are the opposite bet. They’re more uncertain and capital intensive early, but those things flip with time. As Ian Rountree wrote in Full-Stack Deep Tech.
Though full-stack startups may initially be more capital intensive we believe that they are ultimately more capital efficient.
Once Vertical Integrators prove that they can do what they say they’re going to do, they unlock access to non-dilutive financing and practically bottomless markets with very little competition. It may take 7-10 years to start generating meaningful revenue, but it’s much easier to imagine these companies being around in a century once they do.
Their longevity means clearer paths to big outcomes, and the dependability to become platforms themselves.
Just as Edison and Westinghouse laid the wiring for a surge of businesses newly possible and improved by reliable electricity, the Vertical Integrators that bring this new Techno-Economic Paradigm into being will become a platform for entirely new businesses.
This is already happening with SpaceX – a big part of the “why now?” for every space startup – and it’s not hard to imagine that there will be thousands of businesses made possible by cheaper energy, faster construction, faster travel, longer, healthier lives, better food, and automated manufacturing and labor, even if it’s hard to imagine exactly what those businesses will look like.
These businesses – the Vertical Integrators and the new businesses they make possible – will benefit humanity as much or more than the Industrial Revolution and the Second Industrial did. With better tools, and faster progress, we can solve ever-harder challenges.
But if history is a guide, none will capture more value than the intrepid Vertical Integrators that figure out how these new tools fit together and set the stage for the half-century to come.
That doesn’t mean it will be easy. In fact, it will be very, very hard. There will be billion-dollar flameouts and promising technologies that get punched in the mouth by manufacturing.
But some will make it through, and they will be very valuable, both financially and societally. Stay tuned for the Deep Dives to hear the stories of the ones to which I assign the highest probabilities in real-time.
Ultimately, the companies that bring these miracles into existence will be rightly rewarded for doing so, for they will have defined and improved the future. It’s time to vertically integrate.
Thanks to Claude for editing.
That’s all for today. We’ll be back in your inbox on Friday with a Weekly Dose.
Thanks for reading,
Packy
Packy this is a really excellent series. I love how you frame it with electricity as the example. It took decades for inventors and entrepreneurs to realize the potential for reorganizing and building the systems around electricity. I think the same could be said for software. Peter Thiels quote "we wanted flying cars, but got 140 characters" rings true. We were never going to get flying cars until we moved through the social media wave because application layer and infrastructure layer feed each other in staggered innovation cycles; each feeding off the needs, capabilities and limitations of the other. Software and the infrastructure that undergirds it may have finally arrived at the point where new organizing principals around our systems and the businesses that can be built around them will be discovered an enabled. It's the idea of exponential change - the next couple of doublings of software and infrastructure impact and potential could be massive. The next 10-20 years could be very interesting.
Glad that you mentioned Thomas Edison, who I believe is greatly underestimated. So many people today think that Edison invented the lightbulb, which does not sound very impressive. More knowledgeable people know that Edison only perfected the light bulb and made it econonomically viable, which seems even less impressive. Because of this misunderstanding, many people today underestimate his accomplishments.
Thomas Edison (and his team) conceived of, designed, tested, constructed and then scaled across the nation the ELECTRICAL GRID, one the most important technological innovations in world history.