Building an AI Business from Lebanon (Where the Grid Fails Daily)

You can build an AI business from Lebanon. I do it every day. The grid fails, the banks froze, and the work still ships, because a modern AI business runs in the cloud, gets paid through the right entity, and sells to markets far bigger than the street outside.

This is not a sob story. It's a playbook for anyone building from the edge of the map.

What does building from Lebanon actually look like?

Start with the conditions, stated flat.

State electricity has been rationed for years. Most homes and offices run on private generators, solar, or both. Internet quality depends on where you sit and what hour it is.

Then the money. In 2019 the banking system seized. Deposits were locked and the lira lost most of its value. The World Bank's Lebanon Economic Monitor (Spring 2021) said the crisis was likely to rank among the ten, possibly the three, most severe crisis episodes globally since the mid-nineteenth century.

Then the people. A lot of talent left.

Nobody would design these conditions on purpose. But here is what they do: they strip out everything that isn't essential. When power, money and people are scarce, you stop decorating and start building systems.

Constraint forces clarity.

Why build an AI business from Lebanon at all?

Because the case is stronger than it looks from outside.

• Talent. The engineers, designers and operators who stayed are sharp, multilingual, and used to solving problems with no safety net.
• Cost. Running a team here costs far less than in Dubai, London or Riyadh. That buys runway, and runway buys time to get it right.
• Timezone. Lebanon sits between Europe and the Gulf. The working day overlaps with both, so one base can serve clients on either side.
• Proximity. The MENA market is hungry for AI and underserved. Speaking Arabic and living in the region means understanding the buyer before the first call.

I stayed because the edge is where the asymmetric bets are. The comfortable markets are crowded. The hard ones are wide open.

Ignis Ignotum Initium. The fire begins in the unknown.

How do you keep an AI business running when the power goes out?

You design so the outage doesn't matter.

The rule I build by: nothing that matters lives on a machine that needs the grid.

Layer	The failure	The workaround
Power	State cuts and generator switchovers	Solar with battery storage, plus a UPS on the router and workstation so a switchover never drops a call
Internet	Local outages and slow peak hours	Two independent links from different providers or technologies, such as a fixed line plus mobile data or satellite
Compute	A local machine dies or goes dark	Cloud-first everything: code in GitHub, apps on Vercel, data in managed databases
Team	People in different towns on different power schedules	Async by default: written specs, tracked tasks, logged decisions
Workflow	Work lost mid-task	Autosave, frequent commits, nothing important kept only on one laptop

The stack is boring on purpose. Boring survives.

99Visibility, our GEO platform, runs on Next.js, Supabase and Vercel. None of it cares whether my street has power this afternoon.

Our client builds follow the same rule. Top Tools Lebanon, the bilingual tools store we built, runs entirely in the cloud. A power cut in an office never takes the store down. Customers order, the order is saved, the owner gets an email.

How do you sell globally and get paid from Lebanon?

The banking walls are real. So you route around them, legally and cleanly.

1. Put each business in the right entity. Virtual Minds, the AI product studio I co-founded, is registered under ADGM in Abu Dhabi. 99Visibility runs through its own UK company. Each entity sits where its customers and payment rails are.
2. Use payment rails that actually work. Lebanon is not on Stripe's list of supported countries. So the entity you choose decides whether you can take a card payment at all. Business accounts on platforms like Wise make cross-border invoicing workable.
3. Sell locally the way locals buy. Inside Lebanon, cash on delivery and WhatsApp still close a lot of sales. Our e-commerce builds start there, with card payments as an add-on later.
4. Position for the buyer, not the address. A client in the Gulf isn't buying "someone in Lebanon." They're buying a system that works, with proof they can click.

Location stops mattering the moment the work is visible and the invoice is payable.

What did the constraint teach me about building products?

Three lessons. All of them transfer to any market.

Distribution first. Before Torch99, I grew 40+ mobile apps. The lesson held every time: a product nobody finds doesn't exist. I wrote about that in From 40+ Apps to AI Strategy.

Lean or dead. When money can't sit safely in a local bank, every subscription and every feature has to earn its place. Waste doesn't survive here, and it shouldn't survive anywhere.

Build for failure. When the grid fails daily, you stop assuming conditions will be good. You build things that degrade gracefully: cloud instead of local, queues instead of fragile live dependencies, written process instead of knowledge locked in one head.

Products built that way don't break when conditions do. Not the power, not the market, not the team.

Why is the MENA AI gap an opportunity, not a handicap?

MENA businesses are underserved on AI. Most global tools are built in English, priced for Western budgets, and blind to how this market works:

• Arabic and right-to-left interfaces
• Cash on delivery as the default checkout
• WhatsApp as the real sales channel
• Infrastructure that fails without warning

That gap is the moat. A consultant in another timezone can sell an automation. Someone who builds here knows it has to survive a power cut, render Arabic correctly, and close the sale in a chat.

A second gap is opening fast: AI search. When a customer asks ChatGPT or Perplexity for the best supplier in Beirut, the answer comes from whatever the models can find and trust. Most regional brands have given them very little to work with. That is exactly the problem 99Visibility was built to fix.

Being here isn't the handicap. It's the edge.

If you run a business in the region and want to know what a serious AI engagement should include, read what you should actually get from an AI consultant in Lebanon.

FAQ

Can you build an AI startup in Lebanon?

Yes. The product runs in the cloud, the company can be registered where the customers and payment rails are (ADGM, the UK, or elsewhere), and the team works async. Lebanon becomes the operating base, not the bottleneck.

Can AI automation work for a business with an unreliable power grid?

Yes, if the automation runs in the cloud and not on an office computer. The business only needs a phone or laptop with a connection to see the results. Add a UPS on the router and a backup internet link, and an outage costs minutes, not orders.

How do Lebanese founders get paid by international clients?

Through an entity and payment rails outside the local banking system. That usually means a company registered abroad and a business account built for cross-border payments. Set it up before you sell, not after the first invoice bounces.

Is the MENA market ready for AI?

The demand is there and the supply is thin. Businesses that move now, with automation built for Arabic, cash on delivery and local infrastructure, get a head start that is hard to copy.

Building from a hard place, or running a MENA business that's ready to put AI to work? Book a call and let's map it.

---

About the Author

Liwaa Beaini is the founder of Torch99 and co-founder of Virtual Minds, an AI product studio registered under ADGM (Abu Dhabi Global Market), UAE. He is an AI strategist and product builder operating from Lebanon, specialising in GEO (Generative Engine Optimization), mobile app growth, and AI-first systems. He has grown 40+ mobile apps, advised enterprise clients across MENA, and built products across AI visibility, local intelligence, and digital transformation.

View full profile · All articles · Connect on LinkedIn