From Breakthrough to Baseline: AI and the Self-Driving Illusion
This essay discusses the rise and unraveling of robotaxi dreams and how the same myth of lasting edge is playing out in AI.
We’ve all seen the graph.
The smooth bell curve of technology adoption. Innovators, early adopters, the eager majority. It promises a clean arc from breakthrough to ubiquity. What it hides is what happens to the competitive advantage.
What I love most about Uber is the mythology of its founding. Imagine being a billionaire investor, eager to back the next big thing.
Investors poured money into Uber for nearly a decade. During that expansion phase, investors were quite literally subsidizing random people’s car rides.
“Uber passengers were only paying 41% of the actual cost of their trips.”
- Source
That scene plays in my head like a satire of modern venture capital.
Think about that. You raise billions of dollars, and use it to pay for half of a stranger’s ride across town. You spend someone else’s money to subsidize car rides for entire cities, millions of trips, all bankrolled by venture capital.
The Bills Come Due
Uber Technologies, formerly known as Uber Cab, has long championed a vision of a driverless future. At various points, the company openly stated its intention to replace human drivers with autonomous vehicles.
Uber CEO Dara Khosrowshahi said the following;
“I think the human displacement here, while it's not something that is going to happen tomorrow, is going to happen eventually.”
- Source
We don’t disagree that eventually, autonomous cars will replace human drivers.
On paper, this makes a lot of financial sense. Labor is one of Uber’s largest cost centers. Replacing drivers with machines promises massive margin expansion and operational scalability.
But there is a fundamental flaw in the profitability of this theory.
Capitalism does not reward exclusivity for long. Markets operate under the assumption of open access to capital and technology. So even if Uber partners with or acquires the first company to build a commercially viable self-driving car (let’s call it Company A), two key problems emerge.
Company A is unlikely to sell exclusively to Uber.
In pursuit of profit, Company A will want to maximize its addressable market. Limiting its product to one client would cap its upside.
If Company A does restrict sales to Uber, Company B will inevitably emerge.
Another firm, perhaps better, faster, or cheaper, will produce similar autonomous vehicle technology and sell it broadly. This erodes Uber’s self driving moat almost immediately.
The result? Market pricing for rides collapses.
Once autonomous vehicle technology becomes commoditized, it will not just be taxi fleets that gain access. Regular homeowners could own self-driving cars outright. And if everyone has a robotaxi in their garage, why rent from anyone at all?
Margins collapse across the board. Uber loses its pricing power. So do its competitors. Labor may be eliminated, but with it goes the scarcity that once justified profit. The economics of automation don’t guarantee dominance. They flatten the field.
In other words, the robotaxi revolution doesn’t guarantee any single player dominance. It guarantees a new cost baseline for everyone in the industry. A level playing field rarely favors the first mover. It favors the one who best adapts to commoditization.
There could be a short transition period, remember those internet machines in Airports?
This exact dynamic is now playing out with artificial intelligence.
Firms racing to integrate AI, whether for content, customer service, logistics, or code, are hoping for durable competitive advantages. They tout efficiency gains and claim AI-first status as a moat. But just like the robotaxi fantasy, the real story is not about who uses AI first. It is about what happens after everyone has it.
AI models are not static. What is advanced today is open-source tomorrow.
Cost advantages erode as access becomes ubiquitous.
Custom GPTs, automated workflows, and machine-written reports will all follow the same curve: scarce, then standard, then expected.
AI is not the edge. Process, design, and decision-making built on top of AI might be.
But once the tools are broadly available, your advantage is not in having them. It is in what you do with them that no one else thought of.
Technology scales. Strategy does not.
Conclusion:
You cannot build a durable moat on a capability that becomes universally available. Whether it is robotaxis or AI, the diffusion curve ensures that every edge built on exclusive access will erode. Eventually, the question shifts from “Do you have it?” to “What are you doing with it that others are not?”
The winners will not be the first to adopt, but the first to design systems that evolve after the curve flattens.
Disclaimer: Results may include real trades from my personal account guided by the Stratum Algorithm but executed at my discretion. Differences in execution timing, transaction costs, and position sizing can lead to performance variations. Modeled results are derived from historical simulations. All content is provided for educational and research purposes only, not financial advice. Past performance is not indicative of future results.
Thank you for the restack!