Atal Agarwal: With Patience, One Can Build a Self-Sustaining Business Generating $1 Million to $5 Million in Annual Recurring Revenue
Immigrant entrepreneur Atal Agarwal claims that with patience, it is now possible to create a self-sustaining business generating $1 million to $5 million in annual recurring revenue, leveraging artificial intelligence to achieve equivalent value with less capital.
This statement was directed at venture capital buyers. Helium Ventures recently pointed out that companies stuck in the same revenue band find it difficult to raise another round, as a batch of targets is surging at about 10 times year-on-year. Atal Agarwal offers an alternative path: instead of seeking the next check, use models to thin out manpower and startup capital, turning $1 million to $5 million into a sustainable bootstrap range, rather than waiting to be deemed a "non-venture capital story."
His own public resume is that of a product manager turned immigrant entrepreneur: hailing from a small town in Uttar Pradesh, India, he holds a Master's in Technology Management from the Indian Institute of Technology Kharagpur and the University of California, Santa Barbara. He is currently based in San Francisco, developing AI products aimed at immigrants, with parallel projects listed as OpenSphere, OpenStars, LegalBridge, among others. He has not disclosed which of these companies have reached that revenue band, but he assesses that the leverage changes occurring in the industry are rewriting coding, customer service, customer acquisition, and content production into replicable capacities for small teams.
There are already verifiable bootstrap examples. SubMagic reached $1 million in annual recurring revenue within three months of its first paying user, achieving $8 million in about two years with 13 people and zero external funding, averaging $700,000 in revenue per person, with daily new registrations between 5,000 and 10,000. Chatbase crossed $1 million annualized in five months, later reaching about $8 million while still bootstrapped. HeadshotPro achieved approximately $360,000 in monthly recurring revenue with about 40,000 paying users. Fibbler, with two people and two years, surpassed $1 million with zero funding. Formula Bot was bootstrapped by a solo founder to over $1 million. ProjectionLab built a side business to $1 million in annual recurring revenue in four years. Pieter Levels' portfolio of products publicly runs over $3 million in annual recurring revenue, with projects like Photo AI and Interior AI generating over $250,000 monthly.
Vertical AI is also being compressed into shorter windows: some projects with one to two founders in hard-to-sell industries like law and life sciences have achieved over $1 million in annual recurring revenue from idea to execution in six to eight months, then reaching $3 million in four months, with the option to be acquired in 12 to 18 months. Base44 reached $1 million in annual recurring revenue in about three weeks, with around 250,000 users and nearly $200,000 in monthly profit within six months, with the founder retaining 100% equity. In contrast, venture capital-backed AI targets have raised the speed benchmark: Cursor reached $100 million in annual recurring revenue in about a year, while Lovable went from zero to tens of millions to about $100 million in a few months. Both sides use the same set of models, with the difference being whether to sell growth to funds.
The buyers are small and medium-sized enterprises and creators willing to pay monthly, not the next round partners; the sellers are founders replacing engineering, customer service, and deployment roles with model calls. Funds remain in subscription fees and inference bills, rather than in preferred stock liquidation stacks. Beneficiaries are founders who maintain high equity and can exit with qualified small business stock tax structures after five years, as well as model companies selling computing power to these small teams; those under pressure are mid-level software companies still structured with over 10 people, preparing for Series A with a "look like it can hit $100 million" narrative—similarly achieving $1 million to $5 million, but with investor-backed versions finding it harder to raise funds, while non-investor-backed versions have a chance to survive for the first time.
Source: Public Information
ABAB AI Insight
Atal Agarwal's judgment counters the previous narrative that companies stuck at the million-dollar revenue mark can only sell or shut down. He is not an operator like those behind Mailchimp or Basecamp, who have already built self-sustaining giants, but rather treats immigrant status, visa uncertainty, and parallel projects as an experimental field: first working on product discovery at Castlight Health, then applying the same method of turning institutional friction into software to visas, fundraising matching, and legal documents. His historical behavior is closer to that of immigrant founders engaged in continuous trial and error, rather than publishing capital theories after building a profitable platform.
As a result, capital pathways have split into two streams. One continues to pour money into AI-native targets that can scale from zero to $100 million in eight to eleven quarters; the other transforms OpenAI, Anthropic, and open-source code assistants into public infrastructure, allowing two-person companies to sustain themselves through subscriptions. The motivation for the bootstrap path is not to avoid growth, but to avoid preferred liquidation rights: once a company does not raise a Series A, $1 million to $5 million shifts from a "failed venture capital story" to an asset that is "dividend-paying, sellable to strategic buyers, and can be held for five years to enjoy tax exemptions." The model rewrites the capacity of engineers that originally required $5 million to $8 million to hire into variable costs billed by token.
Benchmark cases have already split into two columns. One column includes Mailchimp, which reached about $800 million in annual revenue with zero venture capital and was acquired by Intuit for about $12 billion, and Tuple, which achieved over $5 million in annual recurring revenue in a narrow market with a team of fewer than 15 people. The other column includes Cursor, ElevenLabs, and Midjourney, which have pushed the same leverage to the limit with venture capital. The industry position is not in early exploration but in the bifurcation after tool maturity: vertical niches with a total addressable market of only $25 million to $50 million can now be consumed by two-person teams with high margins, without needing to first prove they can become unicorns.
This is technological substitution. What is being replaced is not the end customer, but the intermediate organizational costs that "must first finance expansion to prove scalability." The mechanism is: large models drive the marginal costs of coding, analysis, and content close to the calling price, while customer acquisition can still go through alliances, search, and vertical communities, thus decoupling revenue bands from financing bands—$1 million to $5 million is a death valley on the venture capital ledger but a target range on the bootstrap ledger. Patience here is not a virtue; it is because there is no longer a need to exchange four quarters of growth for the next check.
ABAB News · Cognitive Law
- When leverage is cheap enough, million-dollar revenue shifts from financing failure to bootstrap target.
- The same model, funds use it to scale tenfold, founders use it to hire fewer people.
- Where liquidation rights disappear, patience finally has financial significance.