The numbers · 9 October 2026 · 5 min read
Three things are happening at once. The digital economy has collapsed, the people are still ahead of almost everyone, and Lebanon is quietly disappearing from the indices that would show either.
In 2018 Lebanon sold 749 million dollars of computer, telecommunications and information services to buyers outside the country. In 2023 it sold 141 million.
Those are not estimates. They are the World Bank's series on ICT service exports, built from IMF balance of payments data, and we pulled them from the Bank's own API rather than from anybody's summary. 749,443,887 dollars in 2018. 140,728,786 in 2023. A fall of 81 percent in five years, in a period when world ICT service exports grew by about three quarters.
It did not slide. It dropped twice.
2015 through 2018 sits flat between 640 and 750 million. Then 2019 comes in at 568 million, and 2020 at 241 million. That is the banking collapse arriving, and it is the fall everyone already knows about. What is less known is the second one. 2021 and 2022 hold roughly level at 245 and 252 million, which looks like a floor. Then 2023 halves again, to 141 million.
The first collapse was the currency. The second one happened after the floor was supposedly found.
Nobody we can find has published an explanation for the 2023 drop. We are not going to invent one. What we will say is that the register's own research points at part of the answer: when you trace the paperwork behind the Lebanese technology companies people cite, you arrive in Abu Dhabi, Vienna, Sharjah, Tokyo and New York. Work done by Lebanese people, invoiced from somewhere else, is not a Lebanese export. It never touches this series.
Microsoft's AI Economy Institute estimates the share of people aged 15 to 64 who used a generative AI product. In Lebanon that was 24.8 percent in early 2025 and 27.7 percent by June 2026. The global figure is 18.8 percent.
So Lebanese adoption sits nearly nine points above the world. That is the headline most people would run, and it is true.
The part worth reading is the pace. Lebanon went from 31st of 147 to 34th. Of the forty economies with the highest AI use, not one gained less ground than Lebanon's 2.9 points. Saudi Arabia went from 23.7 to 31.4 percent and overtook it. So did Slovenia and Vietnam. Jordan went from 25.4 to 31.2. The head start is real and it is being spent.
Oxford Insights scores how far a government can actually use AI to serve the public, across 69 indicators. In the 2025 Government AI Readiness Index, corrected edition, Lebanon is 117th of 195. Its best pillar is the AI sector and its talent, at 75th. Its worst are public sector adoption at 159th and resilience at 163rd.
On public sector adoption Lebanon scores 18.65 out of 100, against a regional average near 48 and a global average near 54. Saudi Arabia is 15th overall, the UAE 19th, Jordan 63rd.
Put the two measurements side by side. More than a quarter of working age Lebanese are using generative AI. The state that governs them is in the bottom fifth of the world at using it. That gap is the single most useful fact on this page.
WIPO's Global Innovation Index 2026, published on 29 September, puts Lebanon 86th overall with a score of 25.6, up from 90th the year before. We confirmed that rank on WIPO's own table.
Underneath it sits the contradiction this country keeps producing. Lebanon publishes a great deal of science for the size of its economy, and almost none of it with the companies that would build anything on it. The research recorded here puts Lebanon 4th of 139 on scientific papers per billion dollars of output, and 137th on the share of research co-authored with private firms, at 0.2 percent. Cyprus is at 4.3 percent, Armenia 1.9, the UAE 1.4. We could not re-verify those two sub-ranks against WIPO's 2026 tables ourselves, so take them as reported rather than as checked, and note that the 2025 edition showed the same shape, with public research and industry collaboration at 138th.
Papers are written. They are not written with anyone who manufactures, ships or sells.
Every number above came from somebody else's dataset. Here is what happens when you go looking for Lebanon inside them.
The ITU, the United Nations agency for digital technologies, publishes an ICT Development Index. In 2026 it scored 159 economies. Lebanon is not one of them. It was scored in 2023 at 76.1 out of 100, and it is absent from the 2024, 2025 and 2026 editions. To be scored, a country needs official data for at least five of ten indicators, or it can reject the ITU's own estimates and be left out. The report does not say which applies here. Over the same stretch Jordan went from 78.5 to 87.0.
Open Data Watch's Open Data Inventory checks what a government actually publishes across 22 topics. Lebanon scores 33 out of 100 and ranks 180th of 197. Seven of the 22 topics score zero, meaning too little was published to assess: education facilities, health facilities, food security and nutrition, international trade, agriculture and land use, energy, and pollution. In 2022 it scored 32 and ranked 167th of 195. The score barely moved. The rank fell thirteen places because other countries kept going.
A country can lose an argument about its own economy simply by not turning up to the measurement.
We did not set out to write about statistics. We set out to build a list of Lebanese AI companies, and kept running into the same wall: there is no denominator. No agreed count of companies, no reliable export figure after 2023, no connectivity score for three years running, and seven whole categories of official data that are simply blank.
That is why the register publishes its rejections next to its entries, and why every card carries the page it was read on and the day it was last opened. Not as a style choice. Because in an information environment this thin, an unsourced number is worse than no number, and the only honest thing a small record can offer is that you can check every line of it yourself.
Lebanon is not short of talent, and the adoption figures say so plainly. What it is short of is evidence about itself. You cannot make policy for an industry you cannot size, you cannot attract capital to a market you cannot describe, and you cannot argue with a ranking you declined to be measured for.
Sources
Written for Club Cedro. If something here is wrong, say so and it gets corrected with the correction left visible.
More writing