The Headless Company
The Headless Company
First published on X, October 2026.
Every October, the Headless Horseman rides back into the culture. He rides hard and gets where he’s going without the thing everyone assumed he’d need.
He’s an apropos visual of a phenomenon that, if not quite here yet, is beginning to gather momentum: the headless company. We’ve seen modern versions of this in recent memory: ghost kitchens (restaurants without a physical restaurant space), retailers without a physical footprint, and banks without a single branch. And now, today, the headless company. It’s a business that, until recently, wouldn’t have been possible, because the head and the headcount were prerequisites to scale, let alone existence.
As we approach Halloween, let’s take a turn thinking about something the Horseman rides without, and your business might one day leverage (as both trick and treat).
Act I: No Head
For most of business history the rule was simple: an idea can start with one person, but it can’t scale with one person. Somewhere between prototype and real revenue you needed people for marketing, sales, operations and the back office. You owned the idea, but along the way, you paid for the capacity.
Shark Tank is where that trade is most visible. Across 18 seasons, about 62% of pitches land an on-air deal, averaging roughly $284K for 23% of the company (Sharkary).
What does that investment get? Yes, it’s often help replenishing inventory and access to the Shark’s networks and distribution channels. Heck, a lot of pitches (even those that go without a deal) use the appearance as a savvy bit of guerrilla marketing. But often, what they’re really looking for is capacity. So many of these founders have been toiling in basements and garages, working nights and weekends to bring an idea to life, and frankly, many are fried. They need help, they know they need help, and $284K for a quarter of their company feels like the only trade to make.
But what if it wasn’t?
What if scale were possible through spooky, untraditional means? Even the most mission-driven founders, who make a priority of hiring, staffing and supporting amazing teams, share an aim: to build something of value, scale and sustainability. The formula for that is changing in real time.
Let’s do the math. The average US worker costs an employer about $47 an hour all-in, per the BLS. Here’s the problem: the equation we once understood has fundamentally changed.
In the past, the equation was: do I pay $47 an hour (or any amount, really, it’s inconsequential at this point) to get a job done, or does the job simply not get done? Or does the job stay with me, and I keep shouldering the burden, like our stereotypical beleaguered Shark Tank builder?
That math is breaking down, because now the real comparison is agentic tools versus the work you would have paid an employee for, shouldered, or skipped: the polished eCommerce storefront you’d have put off, the social presence you’d half-run on weekends, the app that stays a someday idea.
That’s what $30, $300 or $3,000 a month in tools buys me. Not fewer employees. More of the work actually getting done. Shoot, even on a free plan, agentic tools exist for marketing, operations, strategy, sales and supply chain management, and the arms race among these platforms has meant more compute, better models and greater capabilities for builders.
And here’s the part I think people miss. The bots (especially those that run on their own virtual machine, like the recently released Claude Duet and the fast-growing elephant in the room, Grok Bot, which Elon Musk called “the fastest growth of a product that I’ve ever seen in my life”) work 24 hours a day. As long as there’s compute, I can keep handing them tasks, push them to do more, and have several working at the same time on different things. A team of people needs breaks. They go to bed, take vacation, get sick, and they should. So I’m not just renting cheaper help. I’m renting capacity that can do things people often can’t: always on, running in parallel, and scaled up or down in minutes. To be fair, people still bring judgment the bots don’t have, which is exactly why Act IV exists.
There’s a personal side to this too. I don’t know a founder or builder or creative who hasn’t had an idea hit at dinner, late in the evening, or halfway through a morning jog, and the first thing you want to do is get cranking on it while it’s fresh. Those out-of-hours moments used to be dead time, and the idea either waited or faded. Now I can hand it off on the spot and wake up to a first version, which honestly might be my favorite part of all this.
Act II: No Map
The Horseman never needed directions. Distribution has always been the part of a business you had to hire or partner your way into, and it’s the part the bots haven’t fully covered yet. I think that’s next.
✱Not good at building slides or a pitch deck? Enlist Gamma.
✱Need audio or a podcast about your business? There’s NotebookLM or ElevenLabs.
✱Want to create video content for your brand? There’s HeyGen and Higgsfield.
✱Need real-time dashboards to oversee your distribution results? Claude can build them as artifacts.
✱Don’t like managing a bunch of social media accounts? Ditch it. Spin up a CMO with Grok Bot and let it do research, find topics, build a content strategy and then actually run your YouTube, X, LinkedIn, Snapchat, Instagram and TikTok accounts.
The shocking part of writing this is that you not only don’t need employees for these things, you barely need to understand them. You can be running mapless and hapless and still get most of what you need through agentic collaboration. And businesses are beginning to build for this future.
One great example: ad platforms are starting to be built for agents. Neural Ads, for one, plans and buys advertising across Amazon, Meta, TikTok and X, and lets AI agents operate it directly (MCP details). Picture the bot that writes your posts handing a goal and a budget to yet another bot that buys the audience for them, then reporting back to you. This isn’t just headless, it’s interactionless.
If that vision holds up (and I think it will), the map stops being something you buy with equity or headcount. It becomes something you configure.
Act III: No Mount
Here’s where the equity math flips. If capacity is cheap and getting cheaper, the 20% you’d hand an investor for help you could rent is the most expensive thing in the room. And what you’re renting isn’t a cut-rate version of a team. It’s capacity that never clocks out and can run ten jobs at once.
That’s not a knock on investors. Money, judgment and a Rolodex are still worth a lot, and bots can’t replace the Rolodex (yet). But “I need people to do the work” used to be the main reason to give up a big slice early. That reason is disappearing like the jumbo-size candy bars in the Halloween bowl.
Solo isn’t new, either. There are already 30.4 million one-person businesses in the US (Census). Solo was never the hard part (although, let’s be real, it’s ridiculously hard). The real hard part was scale. How does a solopreneur scale when the ideas, the vision, the operations, the connections, the legwork and the relationships are all routed through one discrete and infinitely important chokepoint? It’s been the limiting factor...but now? This is the golden age of micro-entrepreneurship, and what was once the bottleneck can be the unleashing of a singular vision.
Sam Altman has talked about a one-person billion-dollar company, and not only is that real, it’s becoming real now. The horsepower of a headless vision is probably a founder, a vision, a bit of market pressure testing, and then a wildly impactful ramp-up leveraging a fleet of bots. A billion-dollar company will happen, but more impactful at the macro level is that a lot of $300K or $3M companies can now 10x without 10x’ing their employees and expenses.
Act IV: Who Holds the Reins
So what the heck are people doing? In this picture, two jobs are still mine: taste and approvals. We now live in a time where knowing how to build is less important than knowing what to build. That’s half of the equation: the human judgment of “is this idea worthy of my effort and attention?”
After that, much of it becomes a question of taste. Although nearly everything could be outsourced to agentic tools, a lot still comes down to taste, vision and preference. You can hand off nearly all the work, but there are still going to be elements, nuances and concepts that require refinement, editing and interception. Knowing which work can be greenlit, and which still requires a yes, a no or a “not quite,” is key. The bots can make a hundred versions of something. They can’t tell you which one feels like us.
The fascinating part is this: as those refinements take hold, as the playbooks get stronger, and as agentic tools build more persistent memory of what has worked, what the builder has preferred and what’s driven success, the hypothesis is that the need for refinement, editing and interception shrinks dramatically. Over time, we might actually see the automation of entire businesses. It’s a concept radical enough that it feels weird to write, and perhaps a bit spooky, but it’s entirely plausible: the chief role of the founder, creative and builder becomes getting the right idea at the right time and setting up the tools correctly to run at that idea, in a potentially headless ecosystem.
This, of course, comes with a number of potentially ghostly sociocultural ramifications. I’d be kidding myself not to say it: if one person can do this, some of the roles that used to get hired won’t be. That deserves a serious conversation, not a victory lap. And although I wake up daily feeling like I’m behind, rationally, many businesses weren’t built on this kind of chassis or for these kinds of harnesses. Change will come, but it’s likely to come with some pacing caveats.
In the end, the Horseman never seemed to need a head. The org chart, the floors of people, the partner who takes 20% for capacity you couldn’t build alone: all of it is quietly becoming optional. What you still need is a point of view, and the discipline to tell your agentic tools and 24/7 bots what it is. And maybe, in the end, there’s a vision here that’s not as spooky as some might think.
Happy Halloween.