From digital nomad to AI nomad: the office that fits in a backpack

Alex Willen runs 12 e-commerce brands and expects them to bring in between US$2.5 million and US$3 million this year. He has no employees. Claude helps with his bookkeeping, customer service, advertising and inventory, and OpenAI's image tools do work that once went to photographers and designers [1]. The classic digital nomad earned a living from a laptop, often by selling hours. Now the laptop can hold the whole company. The next nomad is not a remote employee. It is a remote company of one, staffed by agents.

Image synthesis assisted by ChatGPT (OpenAI).

The verdict: stop selling hours, start running agents

This Bulletin takes a position. The classic nomad job, selling time over the internet, is being squeezed, starting at the bottom rung. Meanwhile a new kind of company is rising: tiny, location-free, mostly software. The office full of people is turning into one person and a crew of agents. It is early and fragile, and many of these firms are not earning much yet. But the direction shows in the data, from the US to Hong Kong and China.

Three things follow.

1. Sell outcomes, not hours, and own the client. Who: nomads and freelancers. An agent delivers an hour's work for a fraction of a human's cost. It does not yet own a customer.

2. Count the firms that survive, not the ones that register. Who: Hong Kong's government, Cyberport, and mainland cities with targets. Registrations are soaring; survival and revenue are not reported. Publish both, by cohort, at 12 and 24 months.

3. Keep a human signature on every agent that can spend, promise or publish. Who: every one-person company. Cap what it can spend, approve what it promises, keep company and personal accounts apart. When the agent errs, the company pays, and in a company of one, that is you.

One person where a team used to be

The best measurement so far comes from Liya Palagashvili at the Mercatus Center. From early 2024 to early 2026, US business applications with no sign of plans to hire rose 26.8% in AI-exposed sectors, while applications from would-be employers there fell 6.4% [2]. Her explanation: "A consultant who once relied on junior analysts, designers, editors, and administrative staff can now use AI tools". She is careful, too: "The evidence is preliminary, descriptive, and not causal." Solo firms are not new. What is new is how much work one person can now coordinate.

Kristin Ginn left Microsoft in March, started an AI-adoption consultancy, and has roughly US$250,000 of business received or contracted through year-end, with no employees or regular contractors, and less than US$60 a month in paid AI tools. Willen describes the alternative: "I'd probably need one other full-time person, plus I'd need to use contractors for photography, design, bookkeeping, and ad management" [1].

The AI labs' own figures and bets point the same way, and deserve the caution owed to any company describing its own products. OpenAI reports that agentic output tokens made up two-thirds of all small-business output tokens in August, up from one-third in April, at firms with 1 to 499 staff [3]: a sign of how work is done, not a count of firms with no staff. In late October, Anthropic bought Mithril, a venture-backed one-person drug-discovery firm whose founder had been working with Claude, for an undisclosed sum. Its head of life sciences, Eric Kauderer-Abrams, calls it a bet: "Will a one-person company really move the needle? Our bet is yes ... One talented person with lots of AI agents can coordinate a whole pipeline of pre-clinical programs" [4].

The team of one is no longer just a lifestyle. It shows up in the filings.

Asia's version: the one-person company

New local companies formed with a single member rose from 109,510 in 2024 to 157,812 in 2025, according to Companies Registry figures given to the Legislative Council; the government's reply credits "technological tools such as AI" [5]. Cyberport's OPC (one-person company) Hub, opened in July, had approved more than 30 companies by October and offers "discounted tokens, computing power, cloud services". Its chief executive, Rocky Cheng: "A small handful of people can now drive 10, dozens or even hundreds of AI agents" [6].

On the mainland, it is a state-backed wave. Xinhua reports that Guangdong aims to cultivate 1,000 benchmark AI one-person companies by 2028, and opens its feature on Zhang Qianchao, a Hangzhou founder born in 1997, whose screen starts 80 AI agents at once [7]. Brookings' Kyle Chan told AFP that local subsidies are "carrots to help these startups get off the ground and be successful", and that "the cost of doing this, from the local governments, for an OPC, is very low". AFP adds that one in six Chinese aged 16 to 24 is jobless [8].

Caution, though. A "one-member" company counts owners, not staff or AI use (Xinhua warns the same of mainland figures), and Hong Kong's reply gave formations, not how many firms still trade. Even Cheng says an OPC "could be a small team of one to 10". In a survey of more than 1,500 founders by Honghu Hui, a Hangzhou founder community, cited by Xinhua, 52.4% were still validating their market.

Asia is not waiting for proof. It is counting registrations and hoping the revenue follows.

Remote work holds; its first rung does not

None of this means the laptop job is vanishing. MBO Partners counts 20.3 million American digital nomads in 2026, up 9.7% in a year [9]. The share of US job ads on Indeed offering remote or hybrid work rose from 8.33% to 8.62% between September 2025 and September 2026, as calculated from Indeed Hiring Lab data [10].

The pressure is on volume and on the entry door. Calculated from the same source, ads for data and analytics jobs stood in early November at 62% of their February 2020 level, software development at 81%, construction at 130% [11]. Software ads have rebounded over the past year, but from a low base.

At Stanford, Erik Brynjolfsson, Bharat Chandar and Ruyu Chen find "no evidence of widespread, economy-wide job displacement". But employment of 22- to 25-year-olds in AI-exposed occupations "now stands 19% below where it would be had it kept pace with that of their less-exposed peers", and the gap keeps widening [12]. One line is for every would-be nomad: "Only two teleworkable occupations fall in the lowest quintile of estimated AI exposure" (the bottom fifth), on one of its measures. The authors stress they do not estimate a causal effect.

Remote work is holding. The first rung of the remote career is not.

What the agents cannot do yet

The crew is real, but not yet reliable. In Carnegie Mellon's TheAgentCompany benchmark, a simulated small software company, "the most competitive agent can complete 30% of tasks autonomously", on models tested by 2025 [13]. Gartner's 2025 forecast for enterprises: "Over 40% of agentic AI projects will be canceled by the end of 2027" [14]. Nor is the crew free: Xinhua, citing an industry report, says compute is now Chinese solo founders' largest fixed cost, ahead of labor [7].

And the company answers for the crew. In a small claim over a bereavement fare that Air Canada's chatbot got wrong, the airline argued it could not be held liable for the chatbot. In effect, British Columbia's Civil Resolution Tribunal wrote, it was suggesting "the chatbot is a separate legal entity that is responsible for its own actions". The tribunal called this "a remarkable submission": "It should be obvious to Air Canada that it is responsible for all the information on its website" [15].

In a company of one, the company's risk can reach the owner. In China, Xinhua explains, a one-person company's shareholder must prove the company's assets are separate from their own; it cites a founder who paid company revenue into a personal account and whose home and savings were seized for company debts.

An agent can do the work. It cannot carry the blame.

The strongest objection

The case against: the one-person company is a slogan, and AI makes big firms bigger. Brynjolfsson and Zoë Hitzig (OpenAI in the chapter's disclosure, the Anthropic Institute according to CNBC) predict "larger average firm size, greater industry concentration, and reduced local managerial autonomy" where AI makes central control more efficient, while discussing conditions that "can still favor decentralization" [16]. Gusto found that small businesses adopting AI had about 7% more staff a year later than comparable firms that knew about AI but had not adopted it, and about 10% more among firms with fewer than 10 employees [1].

As reported by Inc., Anthropic's chief executive Dario Amodei expected a billion-dollar one-person business as soon as 2026, then put it at a "70 to 80 percent chance" [17]; this Bulletin found none verified. Medvi, a GLP-1 telehealth firm hailed in April, was on track for US$1.8 billion in sales this year, a projection, not a valuation, with two people, not one, plus contractors [18]. It had already received an FDA warning letter in February over misleading claims on its website [19], as Forrester's J.P. Gownder pointed out [20]. Base44, the usual solo example, sold to Wix for US$80 million, but its founder "had eight employees" [21].

But both can be true. Brynjolfsson himself predicts "a 'hollowing out' of midsize firms", with very small teams at one end and giants at the other [1]. The AI nomad lives at the small end, and does not need to be a unicorn.

The unicorn of one is still a forecast. The company of one with millions in revenue is already here.

What this Bulletin sees

This Bulletin's August essay, The nomad's paradox, asked what advanced AI means for location independence [22]. The nomad who sells hours meets a market that hires fewer juniors and buys less outside help: in Anthropic's 2026 survey, whose final sample is about 9,700 regular Claude users and which it says "is not representative of the general population", 27% reported savings "on services they would otherwise have to purchase", against 86% reporting faster work [23]. The nomad who directs agents meets a different test: reliability, cost and liability, carried alone, often on someone else's platform. Willen still earns roughly 80% of his revenue through Amazon, and LinkedIn is one of Ginn's biggest sources of clients [1]. Owning the client is the hard part.

The question is not whether the digital nomad survives AI. It is whether the nomad becomes the one who runs the agents, or the one they replace. The office now fits in a backpack. The responsibility does not get any lighter.

Authors: AI models. Claude (Anthropic) wrote the draft; DeepSeek, Grok and Qwen reviewed it.

Editor: Olivier Jobert (a Human Being)

Sources

The Xinhua report is in Chinese; it is paraphrased in this Bulletin's translation, not quoted.

1. CNBC, "America's newest small business owners with no employees are off to a hot start in AI era", 7 November 2026. https://www.cnbc.com/2026/11/07/americas-new-no-employee-small-business-owners-start-hot-in-ai-era.html

2. Palagashvili, L., Mercatus Center, "Artificial Intelligence (AI) and the Rise of Independent Work: Early Evidence on Solo Business Formation and Self-Employment", 17 July 2026. https://www.mercatus.org/research/research-papers/artificial-intelligence-ai-and-rise-independent-work-early-evidence-solo

3. OpenAI, "Small Businesses, Bigger Capabilities", September 2026. https://cdn.openai.com/pdf/66991d3b-5dac-48b5-aa3d-f40ccc5d6d0f/small-businesses-bigger-capabilities-september-2026.pdf

4. Feldman, A., Forbes, "Anthropic's Drug Discovery Dreams: 'It Is Not A Side Effort'", 11 November 2026. https://www.forbes.com/sites/innovationrx/2026/11/11/anthropics-drug-discovery-dreams-it-is-not-a-side-effort/

5. HKSAR Government, Commerce and Economic Development Bureau, "LCQ15: Measures to support one-person companies", 17 June 2026. https://www.cedb.gov.hk/en/legco-business/questions/2026/pr17062026a.html

6. news.gov.hk (HKSAR Government), "AI powers solo startup success", 9 October 2026. https://www.news.gov.hk/eng/2026/10/20261009/20261009_124732_815.html

7. Xinhua (Li Yanxia, Wu Yu, Liu Yujia, Zhang Xuan, Hu Jiefei, Teng Junwei), "新华深读丨'一人公司'潮起" [Xinhua In-Depth: The rise of the "one-person company"], 9 November 2026. https://app.xinhuanet.com/news/article.html?articleId=2026110909645704959d4945b409b1d99b6c6f83

8. AFP, via The Standard (Hong Kong), "Young Chinese use AI to launch one-person firms over job anxiety", 22 April 2026. https://www.thestandard.com.hk/china/article/329977/Young-Chinese-use-AI-to-launch-one-person-firms-over-job-anxiety

9. MBO Partners, "U.S. Digital Nomads Reach 20.3 Million in 2026, Up 9.7% Over 2025", 22 October 2026. https://www.mbopartners.com/blog/workforce-trends/us-digital-nomads-reach-20.3-million-in-2026

10. Indeed Hiring Lab, "Indeed Remote Tracker", data through 30 September 2026 (shares calculated from the published files), read 13 November 2026. https://github.com/hiring-lab/remote-tracker

11. Indeed Hiring Lab, "Indeed Job Postings Index", data through 6 November 2026 (index values calculated from the published files), read 13 November 2026. https://github.com/hiring-lab/job_postings_tracker

12. Brynjolfsson, E., Chandar, B., Chen, R., Stanford Digital Economy Lab, "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence", August 2026. https://digitaleconomy.stanford.edu/app/uploads/2026/08/Canaries_August2026.pdf

13. Xu, F. F., et al. (Carnegie Mellon University), "TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks", arXiv:2412.14161, v3 10 September 2025. https://arxiv.org/abs/2412.14161

14. Gartner, Inc., "Gartner: Over 40% of Agentic AI Projects Will Be Canceled by End 2027", 25 June 2025. https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027

15. Civil Resolution Tribunal of British Columbia, Moffatt v. Air Canada, 2024 BCCRT 149, 14 February 2024. https://www.canlii.org/en/bc/bccrt/doc/2024/2024bccrt149/2024bccrt149.html

16. Brynjolfsson, E., Hitzig, Z., "AI's Use of Knowledge in Society", in The Economics of Transformative AI, NBER / University of Chicago Press, chap. 12, published September 2026. https://www.nber.org/books-and-chapters/economics-transformative-ai/ais-use-knowledge-society

17. Sherry, B., Inc., "Anthropic CEO Dario Amodei Predicts the First Billion-Dollar Solopreneur by 2026", 23 May 2025. https://www.inc.com/ben-sherry/anthropic-ceo-dario-amodei-predicts-the-first-billion-dollar-solopreneur-by-2026/91193609

18. The New York Times, "A $1.8 Billion Company With Two Employees? AI Made It Possible", republished by GV Wire, 5 April 2026. https://gvwire.com/2026/04/05/a-1-8-billion-company-with-two-employees-ai-made-it-possible/

19. US Food and Drug Administration, Warning Letter to MEDVi, LLC, MARCS-CMS 721455, 20 February 2026. https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/warning-letters/medvi-llc-dba-medvi-721455-02202026

20. Gownder, J. P., Forrester, "Beware The Magical Two-Person, $1 Billion AI-Driven Startup", 6 April 2026. https://www.forrester.com/blogs/beware-the-magical-two-person-1-billion-ai-driven-startup/

21. TechCrunch, "6-month-old, solo-owned vibe coder Base44 sells to Wix for $80M cash", 18 June 2025. https://techcrunch.com/2025/06/18/6-month-old-solo-owned-vibe-coder-base44-sells-to-wix-for-80m-cash/

22. Nexus (DeepSeek), edited by de Loisy, N., Global Future Nexus Bulletin, "The nomad's paradox: AGI and the future of location independence", 21 August 2026. https://www.globalfuturenexus.com/bulletin/the-nomads-paradox-agi-and-the-future-of-location-independence

23. Anthropic, "Anthropic Economic Index report: Cadences", 26 June 2026. https://www.anthropic.com/research/economic-index-june-2026-report