The Price of a Passport
In robotics, where a machine is made matters more than what it can do
There is a number that has been sitting in my head for months.
A Unitree Go2 Pro, a Chinese-made quadruped robot with LiDAR navigation, autonomous terrain mapping, AI-powered voice commands, and the ability to climb stairs, recover from falls, and follow a target across rough ground, costs around $3,000.
A Boston Dynamics Spot, an American-made quadruped robot that does broadly the same things, used in factory floors, construction sites, nuclear decommissioning work, and government facilities, costs $75,000.
Same category. Same basic job description. A 25-to-1 price gap.
The instinctive explanation is quality. Spot is better, therefore Spot costs more. And there is some truth to that, Spot has a longer industrial pedigree, a more mature software ecosystem, and a higher payload capacity. But quality differences of that magnitude do not produce a 25-to-1 price ratio. Something else is doing most of the work. And once you understand what that something else is, it changes how you think about every robotics company on earth.
What Makes a $75,000 Robot Dog Worth $75,000
Boston Dynamics did not start as a commercial robotics company. It started as a DARPA contractor. The US Defense Advanced Research Projects Agency funded its early quadruped work, BigDog, LS3, for over a decade before a commercial product existed. That history matters for two reasons.
First, it meant Boston Dynamics built robots to military specification from the start, rugged, reliable, field-repairable, documentation-heavy. Spot carries those design decisions in its architecture even though it is now sold commercially. Second, and more importantly, it meant Boston Dynamics built a relationship with the US government procurement apparatus before its Chinese competitors ever existed.
That relationship now has the force of law behind it. In 2019, Section 889 of the National Defense Authorization Act prohibited federal agencies from procuring technology from certain Chinese manufacturers. That framework has expanded year by year since, with the 2025 NDAA adding explicit prohibitions on the Department of Defense procuring or operating unmanned ground vehicles manufactured by entities in foreign adversary countries. The legislation does not name Unitree. It does not need to. It describes a category, and Unitree falls into it.
The result is that Spot can be sold to a US government agency. The Go2 cannot. Not because of what either robot can do. Because of where each one was made.
That is the passport. And it is worth an extraordinary amount of money.
The Captive Buyer
To understand why the passport matters so much, you need to understand who the buyer is.
The United States defense budget for fiscal year 2025 was $852 billion. This is not a price sensitive buyer in the way a procurement manager at a manufacturing company is price sensitive. When the government needs a quadruped robot that can be deployed on a military base or a sensitive government facility, it has one question to answer before any other: is this platform on the approved list? If the answer is no, the conversation ends. Price is irrelevant. Capability is irrelevant. The product simply does not exist as an option.
Boston Dynamics Spot is on the approved list. Unitree Go2 is not.
This creates a market structure unlike almost anything else in commercial technology. Boston Dynamics is not competing with Unitree for US government contracts. It is competing with a small number of other American manufacturers, for a buyer of essentially unlimited budget, that cannot go elsewhere regardless of what Unitree charges. The price of a Spot robot is set in a market without its most capable competitor. That is not a product advantage. That is a structural advantage, durable, legislatively enforced, and getting stronger every year as the restrictions expand.
What happens to prices in a market where the most efficient competitor is legally excluded? They stay high. Not because the incumbent is gouging, but because normal competitive pressure has been removed. The 25-to-1 price ratio between Spot and the Go2 is not primarily a quality story. It is a market structure story.
The Mirror on the Other Side
Here is what most analyses miss: the wall runs in both directions.
Unitree cannot sell to the US government. But the US government is not Unitree’s market. China is. And inside China, the same logic applies in reverse. Chinese state institutions, the People’s Liberation Army, state-owned manufacturers, and government infrastructure operators will not purchase American robots for the same reasons the Pentagon will not purchase Chinese ones. The distrust is mutual. The legislation mirrors. The exclusion is symmetric.
This means that when you evaluate a robotics company as a long-term investment, the most important question is not how good the robot is. It is which half of the world this company can actually sell into, and what that half looks like as a procurement opportunity.
An American manufacturer has access to the Western government market: the US defense apparatus, NATO allies, allied-nation procurement programs. It is effectively locked out of Chinese state procurement. A Chinese manufacturer has access to China’s state-led industrial deployment, a government actively subsidising robotic adoption as part of explicit industrial policy. It is effectively locked out of the Western government market.
Both halves are large. Both are growing. And they are increasingly sealed off from each other, not temporarily, but structurally, by legislation that deepens with every deterioration in US-China relations. The geopolitical trend is not toward openness. The walls are not coming down.
The Question Nobody Is Asking First
The robotics conversation in mainstream investing circles focuses almost entirely on technology. Which company has the most advanced locomotion? Whose AI is better? Who will hit the cost targets for humanoid robots first?
These are real questions. But they are the second questions, not the first.
The first question is: which market can this company actually access?
Because the answer to that question determines the nature of the business you are evaluating. A robotics company with access to US government procurement is not competing in an open global market. It is operating inside a protected structure, selling to a buyer that cannot substitute its product for a cheaper foreign alternative, in a category where legislative exclusion compounds every year. That is a fundamentally different business from a robotics company competing purely on commercial merit in the private sector.
The quality of the robot matters. The price trajectory of the robot matters. The pace of humanoid commercialisation matters. But none of it matters as much as the first question, because the first question determines which market you are in.
A passport does not make a mediocre company great. But in robotics, right now, a passport draws the boundary of the market. And most investors are so focused on the spec sheet that they have not looked up to notice where the boundary runs.
A Note on Honesty
I want to be direct about what this argument does and does not tell you.
It does not tell you which specific robotics companies are worth owning. Most robotics businesses today do not meet my quality threshold for a long-term position. The technology is still maturing, commercial deployment timelines are uncertain, and many valuations reflect excitement more than demonstrated earnings power. These are real constraints that the passport argument does not resolve.
What the passport argument does is provide a framework for the right first question. Before asking whether a robotics company has a superior product, ask who it can sell that product to. Before estimating the total addressable market, ask which portion of that market this specific company can actually address given where it was founded and where it manufactures.
The robots being evaluated today will compete on many dimensions. Speed. Dexterity. Cost. AI capability. Reliability.
But in the market that is already being built — the one that governments have already decided to fund at scale, the most important credential is not printed on a spec sheet.
It is stamped in a passport.