
What is a one-person startup?
It is a business founded and operated by one person using mostly software and AI in place of what previously required a whole team. This wave is real but it is being exaggerated: according to Fortune (4/2/2024), Sam Altman says he and other tech CEOs have a "bet" on the first year a billion-dollar company appears with just one person, and that this will happen thanks to AI; however according to sources we have read, no one has yet reached a one-billion-dollar valuation with just one employee. What is actually happening is that extremely small companies are generating large revenue per person. This post explains why this happens thanks to AI, provides examples along with what remains disputed, points out the limits that course sellers usually skip, and draws out what applies to a one-person company (OPC) in Vietnam.

AI reduces the cost of creating products and content, but selling and trust still need people. Diagram: OPCAIOS.
Is a "one-person startup" real?
There is no evidence of a billion-dollar company with exactly one employee. Fortune records Altman's words (in an interview with Alexis Ohanian, recorded at a conference in September) that this milestone is "unimaginable without AI but will now happen"; in the same piece, experts offer counter-arguments: James Currier (NFX) argues that a team of three is more realistic than one because of loneliness and poor decisions; Dan Sutera (Muse) says many "one-person" companies actually rely on contractors; and AI has not yet closed large deals or made aesthetic judgments. In other words, "one person" usually means "one person plus a lot of tools and outsourcing."
Examples often cited and what remains disputed
| Example | What is reported | Note |
|---|---|---|
| Base44 | Maor Shlomo built a "vibe coding" platform (see What is vibe code?); Wix bought it in June 2025 for an estimated $80 million, potentially higher if revenue targets are met through 2029 | Sources vary on headcount: one says under 10 people, another says 6 to 8; "solo" is how the press described it |
| Gamma | According to AI Magazine (19/10/2026), exceeded $100 million ARR with around 50 employees, profitable (see What is Gamma?) | Figures disclosed by the company; 50 people is not one person |
| Midjourney | One report mentions revenue above $200 million ARR with under 40 employees | Secondary source, we have not verified it |
What is common: these are very small companies, not one person, and mostly software products that can be replicated at nearly zero cost. Bloomberg (4/2026), citing PitchBook data, says the headcount of unicorns is declining; we have not fully read this article because it is behind a paywall.
Why does AI make this more feasible?
- Cost of building products drops: writing code, design, writing content is much faster with AI agents (see What is LLM? and What is Antigravity 2.0?).
- Tools do repetitive work: automate workflows, customer care, reporting (see What is n8n?).
- Cheap infrastructure and pay-as-you-go: cloud platforms, payments, email, all have free tiers or pay-per-use (see What is Vercel?, What is Cloudflare Pages?).
- Direct distribution: social media and content let you reach customers without a large sales team.
- Task-based outsourcing: work that needs people (accounting, legal, custom design) is bought on a per-project basis.
What course sellers usually skip
- Surviving is different from booming: the famous examples are outliers; there is no reliable data on failure rates for one-person companies.
- Selling and trust: AI can write sales emails but cannot replace reputation, relationships, and closing big deals.
- Products must have real demand: cheap tools make competitors cheaper too; advantage comes from expertise, data, and distribution channels.
- Hidden costs: time to check AI output, customer support, troubleshooting, legal, taxes.
- Concentrated risk: one person getting sick or burning out stops the whole system.
What applies to a one-person company in Vietnam
Legally, an individual can choose between a single-member limited liability company (Article 74 of the 2020 Enterprise Law, where the owner is only liable within the scope of the charter capital and has legal status) or a private enterprise (where the owner is liable by all assets and has no legal status). This is general information from legal analysis articles, not legal advice; please ask an accountant or lawyer about taxes and the appropriate business structure. In terms of practice, a more realistic approach is: choose a problem facing a specific group of customers, build a repeatable process, use AI to reduce the time for each delivery, and keep a reviewer at points with risk. See the specific roadmap at 30-day roadmap to building an OPC with AI; common mistakes at 5 mistakes that make the OPC model fail.
OPCAIOS's experience
OPCAIOS was built and run by one person, with a lot of work supported by AI agents: drafting articles, translations, reports, system checks. But we don't use ourselves as evidence for the "million-dollar" claim: we haven't published revenue figures, and we don't want to use our numbers to prove a trend. What we've learned from the building process is that AI helps one person work much faster on the product creation part, but the part about choosing the right customers, selling, and maintaining trust is still the hardest and most time-consuming part.
Read more: The secret of one person creating millions of USD/year without hiring any employees; The journey from freelancer with monthly salary to owner of automated OPC.
Frequently Asked Questions
What is a one-person startup?
A one-person startup is a business run by a single founder doing nearly everything, using software and AI in place of most work that previously required a whole team. According to Fortune (4/2/2024), Sam Altman told a story about a "bet" among tech CEOs over which year would see the first billion-dollar company with just one person. However, according to sources OPCAIOS read on 24/10/2026, no company has reached this level yet. For example, Gamma was mentioned by AI Magazine (19/10/2026) as surpassing $100 million ARR with about 50 employees, which is very small but not one person. Because of this, the phrase "one person" usually means one person plus many AI tools and task-based outsourcing. For one-person companies (OPC) in Vietnam, what is more realistic is learning how to use AI to reduce time spent on tasks, not chasing a billion-dollar target.
Has there ever been a billion-dollar company with just one person?
According to sources OPCAIOS read on 24/10/2026, there is no billion-dollar company with just one employee. The examples commonly cited are all small teams, not one person. Base44, built by Maor Shlomo and acquired by Wix in June 2025 for an estimated $80 million, has employee counts reported differently across sources: either under 10 or 6 to 8 people. Gamma, according to AI Magazine, surpassed $100 million ARR with about 50 employees. Midjourney was mentioned in one report as over $200 million ARR with under 40 employees, but this is secondary source material not yet verified. It should be noted that revenue and employee figures disclosed by companies or the press have not been independently verified.
What work can AI replace in a company?
AI can mainly replace repetitive and drafting work: writing, basic design, summarizing, answering frequently asked questions, and reporting. There are five reasons this is now more feasible: lower cost to create products, tools that do repetitive work, cheap and pay-as-you-go infrastructure, direct distribution through social media, and task-based outsourcing. For example, an agent might write a draft report while the business owner checks the data before sending it. However, according to a counter-argument in Fortune's piece, AI cannot yet close major deals or make aesthetic judgment calls. Work requiring trust, relationships, and risk judgment still needs people.
What business structure should I register for a one-person company?
There are two common options for an individual, and this is general information, not legal advice. First, a limited liability company with one member (Article 74, Law on Enterprises 2020): the owner is only liable within the scope of capital contributed and the company has legal status. Second, a private enterprise: the owner is liable with all assets and the enterprise does not have legal status. For example, someone wanting to separate personal assets from business risk often considers the first type. These two types also differ in taxes and procedures, so ask an accountant or lawyer before registering. To learn how to apply AI to a small business, see OPCAIOS's AI application training course.
Sources: Fortune 4/2/2024 (Altman's remarks and counter-argument), Calcalist and Globes (Base44), AI Magazine 19/10/2026 (Gamma), search results on enterprise law (Law on Enterprises 2020), read on 24/10/2026. Revenue and employee figures disclosed by companies or the press have not been independently verified; legal information is for reference only. Experience is from OPCAIOS.
Translated from Vietnamese with AI. Read the original.
Want to apply this in your business?
Book a free 30-minute call (Vietnam time, in English or Vietnamese).
