[PFS Blog cross-post]
I was recently talking with some friends about household robots. One of them sent around a clip of Elon Musk being asked when Optimus will be able to do household chores. Just this week Musk said Optimus would be capable of taking out the trash and doing the dishes in “about 12 to 18 months.”1 My friend offered to bet a quarter of a bitcoin that 18 months from now there will be no commercially available Musk robot that can do laundry and dishes. I wouldn’t take that bet. But I did say it seems likely that household robots will at least start coming into use within ten years or so.
My reasoning: useful robots need two things. They need enough intelligence to guide them, and bodies durable and dexterous enough to do the work. Affordability matters too, but even $50,000 robots would sell well enough to get the field going if they worked well, and then prices would come down. Cars can already drive themselves; maybe a robot needs another order of magnitude or two of intelligence, but given the pace of AI improvement it’s hard to imagine we won’t have sufficient AI to control a household robot in ten years. Battery life seems to me to be a problem, but a solvable one: even if the robot’s battery lasts only an hour during physical activity, it can go recharge itself, or swap battery packs, or maybe you have a pair of tag-team robots, and so on. As for maintenance, it could maintain itself, or two robots could maintain and repair each other, or neighbors’ robots could repair each other under some kind of repair or sharing agreement. The real question is whether household robots will be like the car—initially of limited use and reliability, but steadily better and eventually part of modern life—or like the Segway, which was mostly a flop.
So I had Claude and Grok look at this. I asked for the odds that useful household robots will be in mass production and widespread use in the US by 2037; what they’ll be able to do and what they’ll cost; whether they’ll have swappable battery packs; whether neighbors’ robots might repair each other; and who the leading makers will be. That kind of thing. Then I had Claude combine all this into an overall report, which is below. Claude wrote it in my voice, with my guidance and steering, and I asked it to give special attention to one point that interests me a lot: that one of the barriers standing in the way of robots fixing each other is intellectual property law.
1. Where Things Stand
Start with Musk, since that’s where the bet started. My friend would probably win it. On Tesla’s Q3 earnings call in October, Musk himself said he wanted to “temper expectations around Optimus.” He said the hands, durability and AI still need work, and he put deliveries to outside customers around the end of 2027, with industrial customers first and home robots later.2 This is the same Musk who said in April 2024 that Tesla might sell Optimus externally by the end of 2025.3 Musk, like many entrepreneurs, often overpromises. That’s how startup founders think, and they almost have to: they need to persuade people to invest in long shots. Often they fail. That’s the way it is. A Musk robot doing laundry and dishes in people’s homes by mid-2028 looks like a 5% shot at best.
But Musk is not the whole field. Industrial humanoids are already moving from demos into paid work, and factories and warehouses are where the technology will prove itself first: they offer structure, repetitive tasks and clear economics. J.P. Morgan projects US production of 300,000–400,000 humanoids by 2030 and about 1.5 million by 2035, mostly for manufacturing, logistics and warehousing.4 Homes are harder and will come later. Some markers of where things stand:
- Figure AI is a California humanoid startup founded in 2022 by Brett Adcock, who earlier co-founded Archer Aviation. Investors valued it at $39 billion last year.5 In September Figure sent its Figure 03 robot into 30 Bay Area homes it had never seen, with no new training, to tidy living rooms, fold towels and make beds. It succeeded 56% of the time, compared with 9% for a version trained from scratch.6 That’s a big jump, and it’s nowhere near a product. Nobody wants a maid that fails at every other task.
- In May, Figure 03 robots sorted nearly 250,000 packages over 200 continuous hours with no hardware failures. Each robot worked about four hours, then handed off to another unit and went to recharge.7 That is my tag-team idea in practice, and it shows that durability is coming along.
- 1X Technologies is a Norwegian-American startup with a new factory in Hayward, California. It will take your order for its NEO home robot now: $20,000, or $499 a month. Deliveries were promised for 2026, but as of this past summer no customer delivery had been verified. Chores NEO can’t yet do are handled in “Expert Mode,” in which a remote 1X operator drives the robot during a session the owner approves.8 So the first household robot is partly a person at a remote console somewhere.
- China dominates on volume. Chinese makers accounted for more than 97% of the roughly 19,100 humanoids shipped worldwide in the first half of 2026, led by AgiBot and Unitree.9 Unitree’s average selling price fell about 72% from 2023 to 2025, and its entry-level R1 sells for around $4,000 in China, though it’s more toy than maid.10
2. The Skeptic’s Case
The strongest argument against is that driving is much easier than it looks, compared with handling objects. A car has a wheel and two pedals and stays on roads with rules. A hand in a messy kitchen has dozens of joints and wet, slippery plates to deal with, and there’s no internet-scale dataset of hands doing dishes to learn from. Self-driving took 15 to 20 years to get from the DARPA challenges to real robotaxis, and the robot AI models now driving progress are only a few years old. Embodied tasks in the real world have consistently lagged progress in pure AI. Rodney Brooks (iRobot co-founder, the Roomba guy, and no fool) predicts that “deployable dexterity will remain pathetic compared to human hands beyond 2036.” He also predicts that without new kinds of mechanical systems, walking humanoids will remain too unsafe to be around people.11
My answer is that home robots don’t need human hands. They need good-enough hands and patience. A robot can do the laundry slowly at 3 a.m. while you sleep. And both of my requirements, intelligence and durable, dexterous bodies, are advancing, with the AI side moving faster than the hardware.
3. Car or Segway?
The coming household robot market looks more like the emergence of the automobile than the Segway. The Segway solved a problem nobody had, since walking works fine.12 Housework is a real burden: Americans spend about two hours a day on household activities like housework, cooking and lawn care.13 But the first wave will look like cars in 1905 or PCs in 1980: expensive and flaky, bought by the rich and by enthusiasts, with factories and warehouses going first. The Segway is still a fair warning about hype, though. Useful technology can fall short of its promised everyday impact if reliability, cost, safety or convenience disappoint.
4. The Odds
My rough odds for 2036–37:
- A robot you can buy or lease in the US that does laundry (washer, dryer, fold, put away) and dishes in an ordinary house, mostly on its own and reliably enough for early adopters: about 60–75%.
- That, plus real mass production and a million or more US homes: a bit less than even, perhaps 40–50%.
- Truly common, in 10% or more of US households: about 10%. Even Morgan Stanley, one of the more bullish forecasters, doesn’t reach 10% of US households until 2050. That would be about 15 million units, ranging from 3% of households earning $50,000–75,000 to 33% of those earning more than $200,000.14
So if “have” means everybody’s got one, my skeptical friend is probably right. If it means you can buy one that works, I probably am.
A million homes is roughly the richest 1% of American households, and that is the most realistic picture for 2037: robots as a luxury of the top 1%, the way cars were around 1910 or mobile phones in the 1980s. These households already pay housekeepers, nannies and handymen, and they can afford $30,000 or more for a machine that does some of that work. Their purchases also pay for the learning curve that makes the next model cheaper. Mises made the point about the automobile: within a couple of decades it went from a sign of a luxurious mode of living to something that “even the worker has.” As he put it, “This is the course of economic history.”15 So look ten years further out, to the late 2040s. If prices keep falling the way Chinese humanoid prices already have, a capable home robot could cost about what a used car does today. At that point something like 10–20% of US households might have one, with the affluent near saturation and robots spreading into the middle class. That is somewhat more bullish than Morgan Stanley, but it’s the normal path for a useful new technology.
5. What They’ll Do, Who Will Buy Them, and What They’ll Cost
By 2037, probably: laundry, dishes, tidying and putting things away, wiping counters, basic bathroom cleaning, taking out the trash and recycling, carrying in groceries, making beds, feeding pets, watering plants, fetching things, ordering supplies, and monitoring the house and an elderly parent (falls, reminders). Harder, or still supervised: mixed-fabric laundry with stains, ironing, cooking from scratch with knives and a hot stove, and anything delicate or truly novel. Success rates will keep improving as robots learn each household’s layout and routines and get software updates.
Beyond the basic chore list, other uses will follow, some within ten years and some later:
- Cooking. Chopping, prep and following a recipe come first, with the robot doing more of the actual cooking as it proves it can be trusted with a hot stove and a sharp knife.
- Helping the elderly and disabled. Steadying someone on the stairs, helping them in and out of a car, and eventually carrying a person up stairs or from a wheelchair into a car. This is enormously valuable and also the highest-stakes task on the list, since a dropped plate is one thing and a dropped grandmother is another. Expect it later, and expect specialized lifting machines to get there first.
- Riding along. Now that robotaxis are running in a number of US cities, a robot could ride in one with you or for you: loading and unloading luggage, carrying groceries from the curb to the kitchen, or helping an elderly passenger in and out of the car.
- The honey-do list and home repairs. Hanging pictures, assembling furniture, changing filters and light bulbs, caulking, and diagnosing and repairing household items and appliances. Interior painting is repetitive and forgiving, so it should come early. Re-shingling a roof is dangerous, heavy work that robots could take over from people, but a pitched roof is a hard place for a walking robot, so that will take longer. Simple plumbing and electrical fixes may come within the decade, though occupational licensing laws and building codes may forbid an unlicensed robot from doing the work. That is another state obstacle.
- Helping in a small business. Sanding, finishing and assembly in a furniture shop, stocking shelves, packing orders: the robot as an apprentice or a second pair of hands rather than a replacement for the craftsman.
The early buyers will be affluent and dual-income households, families helping aging parents stay at home, and people with disabilities. Specialized single-task machines (better Roombas, laundry folders, kitchen gadgets) will coexist with general-purpose humanoids and may beat them in some niches for a long time. And the first really good home robot might not be a humanoid at all. Wheels plus two arms is cheaper, and it doesn’t fall on your dog.
As for cost, expect $15,000–30,000 to buy, or $300–700 a month to rent, plus cheaper Chinese models at $5,000–10,000 that do less. Volume production, Chinese supply chains and the industrial learning curve should push capable units toward car-like or major-appliance prices by the mid-2030s. Renting will probably dominate at first: 1X already offers a monthly plan, and when asked whether Tesla would rent or sell Optimus, Musk said “I think probably rent them.”16
6. Batteries
Battery life is being engineered around rather than waiting on some breakthrough in energy density. Swapping already exists. UBTech, a Shenzhen company, makes an industrial humanoid, the Walker S2, that swaps its own battery in about three minutes, and Boston Dynamics’ production Atlas does it in under three.17 Home robots will mostly just dock and recharge. Figure 03 charges wirelessly through coils in its feet, so it just steps onto a stand. Figure lists a five-hour runtime, and 1X claims about four hours for NEO.18 Household chores aren’t three shifts a day, so a robot that works a few hours, naps on the pad and finishes overnight will do fine. Expect swappable packs to be standard in commercial fleets and an option at home, along with my idea of two or three robots handing off work. Between docking, swapping and tag-teaming, “runs out mid-task” stops being a real problem.
7. Robots Repairing Robots
This is mechanically plausible. If a robot can swap its own battery, it can swap its buddy’s hand or forearm. Makers will want modular parts anyway to cut service costs, and robots will be able to diagnose their own faults and order replacement parts online using the owner’s accounts. Under today’s law, though, the likelier model is a company service network, often combined with renting: when the robot breaks, the company swaps it overnight like a loaner car. You’d need one robot and a service contract, not two robots and a neighborhood co-op.
Neighbor repair agreements aren’t crazy. But the main thing standing in their way isn’t engineering. It’s intellectual property.
8. IP: Why Your Robot Can’t Fix Your Neighbor’s
This is one more example of IP law hampering human progress and technology. Say I have a robot from one company and my neighbor has one from another. Under current law, here is some of what stands in the way of his robot fixing mine:
- Copyright in the software. Every robot runs on copyrighted software. Courts have held that merely booting up a machine copies its software into memory. In one case, an independent repair company was held to infringe copyright just by turning on its customers’ computers in order to fix them.19 Congress later carved out a narrow exception for computer maintenance and repair, but the point stands: copyright had turned repair itself into infringement.20 Manufacturers still use copyright this way to keep owners from fixing what they bought, as the repair technician and right-to-repair activist Louis Rossmann has been documenting for years.21
- The DMCA. Section 1201 of the Digital Millennium Copyright Act makes it illegal to circumvent a “technological measure” that controls access to copyrighted software. That holds even if you own the machine and aren’t copying anything. Manufacturers have invoked it against companies making compatible garage-door remotes and replacement printer-cartridge chips. They lost those cases, but the threat of litigation alone deters competitors. And John Deere has used software locks backed by the DMCA to keep farmers from fixing their own tractors.22 A robot that encrypts its diagnostics and pairs its parts to its own serial number is a John Deere tractor with legs.
- Patents. Patents on connectors, battery packs, actuators, hands and charging pads force every competitor to design around everyone else’s. So each maker’s battery, plug and tool ends up different. That isn’t because the differences are better for anyone; it’s because the law requires them. Printers show how this plays out. Cartridges carry chips that “handshake” with the printer, and Lexmark used its patents, along with copyright and the DMCA, to try to stop others from refilling its cartridges or making compatible ones. It took a trip to the Supreme Court in 2017 to establish that once Lexmark sold a cartridge, its patent rights in that cartridge were exhausted.23
I made this same point years ago, in a 2011 Federalist Society debate on IP at Ohio State’s law school. I pointed out that we have all these interoperability problems between technologies and standards, with companies coming up with different connectors and interlocking standards, because they have to design around patents and copyrights. Without IP, I said, “you could expect to see a lot more uniformity, interoperability, and compatibility.”24 The patent attorney on the other side, Steve Grant, thought he had a gotcha. Here’s an anarchist who wants to get rid of the state, he said, yet he wants standards and interoperability. Who’s going to impose those, if not the state?
But that misses the point. I wasn’t calling for the state to force anyone to adopt standards. I was pointing out that the state, through patent and copyright law, is a cause of incompatibility. Take away the legal barriers and interoperability tends to emerge on its own, wherever consumers value it. Industries standardize voluntarily all the time. Japan’s automakers, for example, are now discussing standardizing car parts customers never see, like wiring harnesses and hoses.25
The same goes for robots. In an IP-free world, neither my robot’s maker nor my neighbor’s could stop anyone from making compatible batteries, parts, tools or software fixes, or from reading the diagnostics on a robot he owns.26 And since consumers would prefer robots that can share parts and fix each other, makers would compete to offer that. My robot could order a part, and my neighbor’s could install it. A neighborhood robot co-op, an independent robot mechanic down the street, or my own second robot could open mine up, run diagnostics, load a fix and swap a hand. There would be no DMCA claims, no copyright infringement worries, and no patent suits.
And the whole “right to repair” debate would go away. Right-to-repair laws are a classic case of the state “fixing” a problem it created. Instead of repealing the patent, copyright and DMCA rules that let manufacturers lock down repair, legislators and regulators pile on new mandates forcing manufacturers to hand over parts, manuals and tools. Deere, for example, settled an FTC antitrust suit this past July by agreeing that for the next ten years, under federal and state supervision, it will give farmers and independent repair shops the same repair tools and information it gives its dealers.27 Libertarians shouldn’t want either one: neither the IP lockdown nor the regulatory “fix.” Abolish IP, and the “right to repair” is simply the owner’s property right in his own robot. Manufacturers would still be free to use closed designs, decline to sell parts, or void warranties for outside repairs; that’s their business, and the market would judge it. What they couldn’t do is use the state to stop the rest of us from making compatible parts or tinkering with machines we own.28
9. Who Will Make Them
Probably a split like the smartphone market’s. A few US premium makers will compete with a flood of cheaper Chinese ones. On the US side: Tesla, if it can build at scale; Figure, with its home-oriented AI and industrial experience; 1X, focused on the home; and Boston Dynamics (owned by Hyundai), Agility Robotics and Apptronik on the industrial side. On the Chinese side: Unitree, AgiBot, UBTech and Xpeng, competing on price and fast iteration. The software “brains” may come from AI labs like Google DeepMind, Nvidia and Physical Intelligence, a kind of Android for robots.
10. Conclusion
So I’ll stick with my hunch. By the mid-2030s, household robots will very likely be in real use: pricey, mostly rented, imperfect, and mostly in the homes of the affluent, but useful. That’s the car around 1910, not the Segway. Industrial scaling through the late 2020s and early 2030s is the proving ground; if it succeeds and costs fall, useful home robots become a reasonable expectation rather than science fiction, and a decade after that they should be spreading well into the middle class.
The main risks are reliability in messy real homes; safety, regulation and insurance; privacy (a robot with cameras wandering your house, sometimes driven by a stranger); and whether the whole thing is convenient enough that people prefer it to hiring help or buying a specialized appliance. Then there’s the state. IP is the main problem, but there are others: product-liability law, safety regulators, privacy rules, occupational licensing, and maybe tariffs or “national security” bans aimed at Chinese robots, which are already the cheapest and most numerous. Every one of these will make robots more expensive, more closed and slower to arrive than they need to be. The engineers will probably solve the hands. The bigger question is whether the state will let us get the full benefit.
A fair way to settle the ten-year question would be a bet with terms like these, and I’d happily take the “yes” side: by January 1, 2037, at least one robot is sold or leased to US consumers for no more than $30,000 or $1,000 a month (in 2026 dollars) that does laundry and loads and unloads a dishwasher in ordinary homes with no remote human driving it, as confirmed by independent reviews, with at least 100,000 units in US homes.
Related
- Stephan Kinsella, Against Intellectual Property (Auburn, Ala.: Ludwig von Mises Institute, 2008)
- Stephan Kinsella, “The Problem with Intellectual Property,” in Christoph Luetge & Marianne Thejls Ziegler, eds., Handbook of the Philosophical Foundations of Business Ethics, 2nd ed. (Dordrecht: Springer, 2026)
- Stephan Kinsella, “Right to Repair,” C4SIF (Dec. 5, 2025)
- Stephan Kinsella, “IP as Contract,” C4SIF (Apr. 20, 2025)
- “KOL079 | ‘Federalist Society IP Debate (Ohio State)’ (2011),” Kinsella on Liberty Podcast (Sep. 14, 2013)
- Stephan Kinsella, “LIBERTARIAN ANSWER MAN: Smart Contracts,” StephanKinsella.com (Feb. 16, 2022)
- Stephan Kinsella, “The Title-Transfer Theory of Contract,” Papinian Press Working Paper #1 (Sep. 7, 2024)
- Stephan Kinsella, “Kinsella’s Libertarian Case Against Intellectual Property (Grok and ChatGPT),” C4SIF (July 3, 2025)
- Stephan Kinsella, “Whereupon Grok admits it (and AI) is severely gimped by copyright law,” C4SIF (Mar. 19, 2025)
- “KOL505 | The Conflict Between AI and Copyright: World Liberty Summit 2026–Bolivia,” Kinsella on Liberty Podcast (Nov. 11, 2026)
- “KOL492 | Menger Institute Podcast #6: Property Rights, Patents, Anarchy, Technology, Long-Term Hope for Freedom and the Technological Death of the State,” Kinsella on Liberty Podcast (June 13, 2026)
- “KOL458 | Patent and Copyright versus Innovation, Competition, and Property Rights (APEE Guatemala 2025),” Kinsella on Liberty Podcast (Apr. 8, 2025)
- Chris Arnold, “Musk Says Optimus Will Do the Dishes ‘Within 12-18 Months.’ Here’s Where Tesla’s Robot Stands,” TIKR (Nov. 20, 2026) (remarks at the TomorrowX conference, Nov. 18, 2026). [↩]
- Asif Khan, “Tesla (TSLA) Q3 2026 earnings results and conference call transcript,” Shacknews (Oct. 21, 2026); Fred Lambert, “Tesla pushes Optimus deliveries to late 2027, says hands still a problem,” Electrek (Oct. 22, 2026); Nehal Malik, “Tesla Officially Begins Optimus Robot Production; Demo Coming Soon,” Not a Tesla App (Oct. 22, 2026). [↩]
- Scarlett Evans, “Tesla Humanoid Robot Optimus Could Launch in 2025,” IoT World Today (Apr. 24, 2024). [↩]
- Pontus Armini, “J.P. Morgan Forecasts 300,000 to 400,000 US-Made Humanoids by 2030,” Humanoids Daily (Nov. 12, 2026). [↩]
- “Figure,” Contrary Research (Mar. 27, 2025); Eugene Demaitre, “Figure AI passes $1B with Series C funding toward humanoid robot development,” The Robot Report (Sep. 16, 2025). [↩]
- Greg Bock, “Figure Unveils Helix 2.5 With Zero-Shot Humanoid Generalization Across 30 Homes,” The AI Insider (Sep. 17, 2026). [↩]
- “Figure AI Pops Champagne as Autonomous Marathon Crosses 200 Hours Without Hardware Failure,” Humanoids Daily (May 22, 2026). [↩]
- 1X Technologies, “1X Opens NEO Factory in Hayward, CA – America’s First Vertically Integrated Humanoid Robot Factory, with Consumer Shipments Planned for 2026,” GlobeNewswire (Apr. 30, 2026); Radica Maneva, “1X NEO Review: Price, Pre-Order & Release Date (2026),” RoboZaps (Mar. 10, 2026; fact-checked July 16, 2026). [↩]
- “Chinese Makers Accounted for More Than 97% of Global Humanoid Robot Shipments in H1 2026,” TechNode (Aug. 11, 2026). [↩]
- “China humanoid prices plunge as Unitree cites 72% drop,” Humanoid Guide (June 15, 2026). [↩]
- Rodney Brooks, “Predictions Scorecard, 2026 January 01,” RodneyBrooks.com (Jan. 1, 2026). [↩]
- I’ve made a similar point about money and technology. Money solves real problems, those of barter, and Bitcoin may solve a real problem too: the state’s corruption of the money that emerged naturally on the market. By contrast, “smart contracts” and putting property deeds on a blockchain look like solutions in search of a problem. Most contracts deal with uncertain future events and obligations that can’t simply be escrowed or automated in advance. Under the title-transfer theory of contract, “damages” and performance bonds are just pre-agreed conditional title transfers, and someone still has to decide whether the conditions have been met. AI and arbitration may automate some of this, but only at the margins. See Stephan Kinsella, “Am I a Bitcoin Maximalist?,” StephanKinsella.com (Oct. 11, 2024) (“Smart contracts make no sense, keeping records in a blockchain is stupid”); idem, “LIBERTARIAN ANSWER MAN: Smart Contracts,” StephanKinsella.com (Feb. 16, 2022); “KOL463 | Contracts, Usury, Fractional-Reserve Banking with André Simoni,” Kinsella on Liberty Podcast (Apr. 28, 2025); Stephan Kinsella, “The Title-Transfer Theory of Contract,” ch. 16 in David Howden, ed., The Palgrave Handbook of Misesian Austrian Economics (Palgrave, forthcoming 2027), based on idem, “The Title-Transfer Theory of Contract,” Papinian Press Working Paper #1 (Sep. 7, 2024); idem, “Libertarian Answer Man: Future and Conditional Title Transfers Under the Title-Transfer Theory of Contract,” StephanKinsella.com (Sep. 27, 2024). [↩]
- U.S. Bureau of Labor Statistics, “American Time Use Survey – 2025 Results” (June 25, 2026). [↩]
- “Humanoids: A $5 Trillion Market,” Morgan Stanley (May 14, 2025). [↩]
- Ludwig von Mises, Liberalism: In the Classical Tradition (1927), ch. 1, § 5, “The Inequality of Wealth and Income.” [↩]
- Nehal Malik, “Musk: Tesla Will Rent Optimus Robots Instead of Selling Them,” Not a Tesla App (Nov. 6, 2026). [↩]
- Seth Price, “UBTech Walker S2 Swaps Batteries Autonomously,” Control.com (Aug. 6, 2025); Jijo Malayil, “Boston Dynamics’ humanoid auto swaps its batteries in under three mins to keep working,” Interesting Engineering (Aug. 10, 2026). [↩]
- Figure AI, “Introducing Figure 03” (Oct. 9, 2025); “Figure 03,” Figure AI (accessed Nov. 22, 2026); Maneva, “1X NEO Review,” supra. [↩]
- MAI Systems Corp. v. Peak Computer, Inc., 991 F.2d 511, 518–19 (9th Cir. 1993). [↩]
- 17 U.S.C. § 117(c), added by the Computer Maintenance Competition Assurance Act, Title III of the Digital Millennium Copyright Act, Pub. L. No. 105-304 (1998). [↩]
- See Stephan Kinsella, “Right to Repair,” C4SIF (Dec. 5, 2025) (embedding a Rossmann video and noting a recent Rossmann interview that mentions my Against Intellectual Property); “You Don’t Own What You Bought | with Louis Rossmann,” Schweizer Monat, YouTube (May 22, 2026); Louis Rossmann, “What does authorized repair do? Let’s find out!,” YouTube (Feb. 22, 2017); and generally Rossmann’s YouTube channel. [↩]
- 17 U.S.C. § 1201; Chamberlain Group, Inc. v. Skylink Technologies, Inc., 381 F.3d 1178 (Fed. Cir. 2004); Lexmark International, Inc. v. Static Control Components, Inc., 387 F.3d 522 (6th Cir. 2004); Luke Hogg, “How John Deere Hijacked Copyright Law To Keep You From Tinkering With Your Tractor,” Reason (Jan. 8, 2024); see also Brian X. Chen, “Why One Man Is Fighting for Our Right to Control Our Garage Door Openers,” New York Times (Dec. 4, 2025). [↩]
- Impression Products, Inc. v. Lexmark International, Inc., 581 U.S. 360 (2017); Lexmark, 387 F.3d 522, supra. [↩]
- “KOL079 | ‘Federalist Society IP Debate (Ohio State)’ (2011),” Kinsella on Liberty Podcast (Sep. 14, 2013) (debate of Mar. 3, 2011, at the Moritz College of Law; the exchange is at about 49:40–52:00). [↩]
- Byron Hurd, “Japan Wants to Standardize Car Parts to Protect Its Auto Industry,” The Drive (July 16, 2026). This is linked in the update to KOL079, supra. [↩]
- Could a robot maker get the same result by contract, say by making buyers agree never to let another company’s robot or an independent mechanic work on theirs? Not really, as I’ve argued for years. A contract binds only the parties to it, not third parties like the neighbor, his robot, or an independent parts maker, so it can’t replicate IP’s rights against the world. And such terms would be commercially self-defeating. Sellers rarely try them today precisely because they can rely on IP law instead. A robot maker that demanded buyers give up the freedom to repair or modify their own robots, on pain of hefty penalties, would lose customers to competitors that didn’t. It would be like a bookseller requiring you to pay for a book and also agree to pay large fines if you ever copied it: buyers would go elsewhere, or to piracy. When you pay good money for a car or a tractor, you expect to use it as you see fit, and the same would go for a robot. See Stephan Kinsella, Against Intellectual Property (Auburn, Ala.: Ludwig von Mises Institute, 2008), “IP as Contract”; idem, “IP as Contract,” C4SIF (Apr. 20, 2025); idem, “The Problem with Intellectual Property,” in Christoph Luetge & Marianne Thejls Ziegler, eds., Handbook of the Philosophical Foundations of Business Ethics, 2nd ed. (Dordrecht: Springer, 2026), Part III.C.2, “IP by Contract” (errata at idem, “‘The Problem with Intellectual Property,’ Springer 2026,” C4SIF [Sep. 2026]). [↩]
- Federal Trade Commission, “FTC, States Secure Settlement with Deere & Company, Advancing Farmers’ Right to Repair” (July 8, 2026). [↩]
- See Kinsella, “The Problem with Intellectual Property,” supra; idem, Against Intellectual Property, supra; and the other Related links below. [↩]



