The VC-Funded Company Is an Obsolete Organizational Form
Here’s a story you’ve heard before. A smart motherfucker has an idea and a pitch deck. Some equally (?) smart VCs hand over $5 million for 20% of the company. Thirty equally (?) smart engineers get hired. An office in SoMa gets leased. A product gets built. Eighteen months later, there is a Series A. Then Series B. Then either an IPO that makes everyone rich, or an acquisition that makes some people rich, or a flameout that the fund writes off as one of nine expected failures out of ten.
Good story.
Great story.
Made sense for fifty years.
But the story is now mostly wrong, and the people in the room know it.
The specific economic conditions that made VC-funded companies the dominant organizational form are eroding fast enough that clinging to the model is starting to look like a category error. A lot of GPs already suspect this. The conversations at LP meetings have changed. The conversations at partner offsites have changed. What hasn’t changed, yet, is the deployment.
The old justification for the excesses and cultural idiosyncracies of venture was that software companies have enormous fixed costs and near-zero marginal costs. You spend $10 million building the thing, but once it exists, the 10,000th copy costs nothing to deliver. The company needs a large pile of money before it has any revenue, and banks won’t touch it because there’s nothing to collateralize. You can’t repossess a SaaS app. So the equity investor steps in, willing to eat a high failure rate in exchange for the occasional 100x return.
And the fixed costs were real. In 2005, building a web application meant 5-15 engineers at $150-250K each, servers racked in-house, an office (nobody admitted to remote work yet), a year or two of runway, and a sales team because distribution meant humans on the phone with humans. Add it up: $2-5 million before the idea was even known to work. That’s real capital spent on a hypothesis.
VC solved that problem. High upfront costs, massive uncertainty, no collateral, need for risk-tolerant money. To a degree, it made sense.
But every single one of those line items has been cratering. All of them.
Run the numbers on what it actually costs to build a serious SaaS product right now, in 2026, compared to 2015.
The big one is labor. In 2015, the team needed maybe 8 full-stack developers at $180K/year loaded. That’s $1.44 million in engineering salaries for a single year of building, before anyone has used the thing. Today? Conservative surveys say 2-3x productivity gains for experienced developers on greenfield work using AI coding tools. Call it 2 engineers instead of 8. Maybe 3 if the team is cautious. (My bet is 2.) Servers that used to run $50,000 upfront, before a line of code was written, are now serverless: zero cost until someone uses the product. Offices? Remote won that argument. These are no longer interesting conversations.
Distribution might matter more than the engineering costs. Getting the first 10,000 users used to mean a 5-person marketing team with a real budget. The “build it and they will come” line is still wrong. They will not come. But a small team with good instincts and a Twitter account can now reach that scale through content, TikTok, open source, and social. Not guaranteed. Getting attention is a different skill than building products, and that skill gap is part of why distribution still feels hard. But the cost floor has dropped to where a competent team can reach a real audience without spending money.
Add it up. Two people, AI tools, serverless, organic distribution. Burn rate: two salaries. $30K/month if they pay themselves well. Six months to build and launch. Total: $180K.
That’s a seed round from friends and family. It’s what a mid-career operator has in savings. None of it requires giving up 20% of a company to a fund on Sand Hill Road.
The retort from the GP side: venture isn’t just about money. VCs provide mentorship, connections, credibility, recruiting help, and strategic guidance.
Depending on the fund and its degree of LARP, this may or may not be true. Some funds genuinely provide these things. The harder question, for an LP allocating capital across the asset class, is whether they provide $5 million worth of these things, net of dilution, across a portfolio.
Start with mentorship. It’s the strongest claim and the hardest to evaluate. Some founders credit their investors with advice that saved the company. There’s a massive survivorship bias problem here. The fund hears from founders whose companies succeeded, and successful founders tend to credit everyone who helped them. The fund almost never hears from the founders who followed the advice and it destroyed the company, because those founders are off doing something else and don’t write Medium posts about it.
Connections? Made more sense when getting a meeting with an enterprise customer or a key hire required a warm introduction from a known investor. LinkedIn, Twitter, and the general flattening of professional networks has weakened this. Not eliminated. Knowing Marc Andreessen still opens doors that are closed to most people. But weakened to the point where it’s worth a dinner, not a fifth of the equity.
Credibility is just circular. VCs provide credibility because other people believe VCs provide credibility. If bootstrapped companies come to be seen as legitimate, and that shift is already well underway, the credibility premium evaporates.
If those are the benefits, what are they meant to outweigh?
VCs need 10x returns on the winners to cover the losers. Which means every company in the portfolio has to pursue a strategy that’s either worth $1 billion or worth $0. The billion-or-bust thing is not founders being reckless. It’s what the math actually requires.
It produces some weird pathologies. Companies raise more money than they need (funds want to deploy large checks; founders feel pressure to “raise as much as you can”) and then spend it on premature scaling. Fifty employees when ten would do. Super Bowl ads before product-market fit. Three new markets at once.
Everyone chases growth metrics that look good in fundraising decks rather than economics that work. The “we lose money on every transaction but we’ll make it up in volume” thing sounds like a joke. It described the actual strategy of multiple billion-dollar startups from the last decade.
There’s a founder I know who built a very nice $8 million/year B2B tool. His investors told him it wasn’t big enough. He tried to scale it into something it wasn’t, burned through his runway, and shut down. The $8 million/year business was right there. Companies that could have been profitable, healthy $20 million/year businesses get pushed to “swing for the fences” and destroy themselves. The VC model has no room for a $20 million/year business. A $200 million fund can’t move the needle on something that tops out at $20 million in revenue. The fund needs unicorns.
Which means there’s a whole category of businesses the model either ignores or damages. Real businesses, real profits, real employees. Just no plausible path to a billion-dollar valuation. The missing middle of the cap stack.
Too bad for them.
The counterargument I take most seriously: some things require massive upfront capital. You can’t build a semiconductor fab with $180K.
Right. Physical infrastructure, drug discovery, satellites, fusion. Those still need large pools of patient capital, and VC is a reasonable source for that. Deep tech and frontier science are arguably where the model still earns its place.
But most VC dollars over the past fifteen years have not gone to fabs. Not to drug discovery. Not to anything hard. They’ve gone to the exact categories where costs have collapsed: SaaS and marketplaces and consumer social, over and over. Every Crunchbase report on capital deployment shows it. VC has been a software financing vehicle for decades, and software stopped needing it.
“Even in software, some ideas need large teams. You can’t build an enterprise CRM with two people.” That was true until about 2024. AI coding tools, pre-built components, APIs, cloud infrastructure: the minimum team for building serious software is shrinking fast. Not to one person for everything. But the trend line is clear, and it’s accelerating.
So what fills the gap?
Small self-funded teams, for one. Two to five people building profitable software from day one, using AI tools, keeping costs low, owning 100% of what they build. The “indie hacker” and bootstrapper communities have been doing this for years and getting condescended to by the VC world the entire time. They are now the most attractive M&A target many funds have access to.
Revenue-based financing is another piece. Lend $500K, get repaid as a percentage of revenue, no board seat, no pressure to blitzscale. Pipe, Clearco, and a dozen others do this already. The asset class will grow, and a few smart funds will reposition into it before the rest figure out their pitch decks need a rewrite.
There are weirder, more interesting arrangements too. Fluid teams that form around a project, ship it, and dissolve. This is how most creative work already happens. AI tools make it feasible for software too. Open source projects that give the core product away and charge for hosting and enterprise features. Dozens of companies have proven this works when development costs are low enough for community contributions to carry the load.
This is not a “venture capital is a scam” argument; those essays write themselves. For the era in which it emerged, the model was a brilliant financial innovation that funded Google and Amazon and plenty of other companies that put a dent in the universe (for better or worse). Some businesses today still benefit from a large upfront capital infusion. Some funds add real value beyond the check.
But the default assumption, that a software startup needs venture capital, is an artifact of cost structures that no longer exist. The economics have shifted underneath the industry. Organizational forms will follow. They always do, even if it takes longer than you’d expect.
The East India Company lasted for centuries and then it was gone. Steamships replaced sailing ships. The telegraph replaced packet boats. An integrated trade-and-governance monopoly stopped making sense once transaction costs dropped a hundredfold.
Sound familiar?
The fixed costs of building software are dropping by orders of magnitude. The standard takes claim this means SaaS is doomed. The model that’s actually doomed is the one that exists to finance fixed costs that don’t exist anymore. The organizational forms will be replaced by forms better suited to the new economics. Nobody yet knows what those forms look like in detail. Probably messier and more varied than the model they’re replacing.
The funds that will outperform over the next decade are the ones running this analysis now and adjusting fund construction, check size, and thesis accordingly. The funds that lose are the ones still pitching the 2015 deck to LPs in 2027.
Selfonomics is designed, built, and backed by Studio Self
We make tech legible.
Reach out: hello@thisisstudioself.com