What a failed first build taught me about bootstrapping abroad

After falling in love with Da Nang as a digital nomad from Tunisia in 2024, I made the Vietnamese city my home and eventually incorporated my startup, XWorld, in February 2025. XWorld is building Nora, an enterprise AI platform that connects a firm’s data, tools, and daily operations.

Bootstrapping a company anywhere in the world is challenging, but doing so in a foreign country with a cross-border team comes with a unique set of challenges.

Image credit: Timmy Loen

We’re still pre-revenue, so this isn’t a success story. Instead, it’s an account of what two entities – a six-person team and a first build that missed the mark – actually cost, and what I’d do differently.

The short version: Effort gets a solo founder started, but clarity about what you’re building is what stops that effort from going to waste.

Lack of clarity is expensive

I used my income from trading cryptocurrency futures to finance my efforts to build XWorld. I began Nora’s development with three freelancers I found through my Tunisian network: two developers and a designer. I paid roughly US$500 per person for their initial work.

Staffing changed with the project as I searched freelance groups, tested contributors, and moved into contracts as working relationships developed. Eventually, I assembled a team of six, including the initial three staff, whose monthly salaries ranged from about US$500 to US$800 per person.

We began developing Nora in October 2025. Since the last three of the six hires joined from Bangladesh, my coordination duties became more demanding.

See also: Why this founder killed his AI startup despite early success

I remained in Vietnam, adjusting my hours around the team and making myself available for reviews and whatever questions they had about the product. At the same time, I handled administrative tasks and earned money in my day job so I could pay people.

The team delivered an early version of the platform after a little over two months of work. What arrived was essentially a chatbot within a dashboard, but the enterprise AI product I had imagined was missing: an AI operating system that connects a company’s knowledge, tools, and daily operations and helps people make decisions.

At that point, it was costing me roughly US$6,000 to US$9,600 in payroll to keep things going. The time lost on restarting the product may have been more expensive.

Image credit: Ulla

Location, location, location

Stay scrappy