Pakistan's farms are top 10 in output, bottom 50 in yield

From British Telecom GM to mango farmer

The episode opens with Muhammed Bukhari having just landed in Islamabad — ten days out of London after twenty-two years. He moved lock, stock and barrel with his wife and children, entirely on account of Farmdar. The timing, he admits, was more luck than plan. “I wish that there was so much planning in this that I could tell you we planned it a lot — but there wasn’t. I think luck was very involved.”

Muzamil traces the arc of Bukhari’s career before getting to the farming question. Bukhari grew up in Lahore, attended Beaconhouse Garden Town, and left for Queen Mary, University of London around the turn of the millennium to study economics. He never finished that degree cleanly — by his second year he was working part-time at an early-stage startup, and by his third year he had largely abandoned lectures for it. A banking internship in Pakistan had already put him off the conventional path. “It was one of the most boring things I had ever taken. It really put me off for a very, very long time.”

From the startup he moved to McKinsey’s London office, spending three years mostly in healthcare, improving NHS hospital organisations. Travel demands on a young family pushed him toward corporate life, and he joined British Telecom’s Future Leaders Programme. He eventually became GM Commercial for Broadband in the UK — responsible for the products that connected homes across the country. “It was good clean work. I never had any burden on my conscience that I was involved in something that was not completely clean and had a good impact.”

The mango problem that started everything

The pivot to agriculture did not begin with a startup idea. It began with the family’s own farms — mango orchards in Rahim Yar Khan and wheat land in Sheikhupura — that were sitting underutilised, leased out to contractors (artis) who paid in instalments and took the fruit to mandi. Around 2015, Bukhari and his family decided to try exporting their mangoes to the UK directly.

What he found when he got to London’s wholesale markets was humbling. “We used to think our mangoes are the best in the world. When I actually tried to market them in London, my eyes were opened. Our quality was really poor. The way we were growing them didn’t do justice to the varieties we have.”

Quality, he explains to Muzamil, is not just about taste. It has three dimensions: appearance (bruised or blemished fruit gets pulled from the shelf immediately, creating 20–25% waste), consistency of size (which depends on how the tree is managed, not just the variety), and texture — mangoes going soft before they reach the consumer. The family installed a blast chiller at the farm, tried reefer containers, modified atmosphere packaging sourced from Turkey, and extended shelf life from roughly two weeks to longer. None of it solved the core problem. “If your product’s underlying quality is poor, you are just elongating the shelf life of a bad quality thing.”

That realisation brought them back to the farm itself.

What satellite imagery actually sees

Over the next eighteen months, Bukhari’s team ran experiments with drones, multispectral cameras, satellite imagery, sensors, and weather stations. The results were a step change: consistency improved, fruit drop reduced, water use fell by roughly 50%, and fertiliser use dropped by around 30%.

The mechanism behind the water saving is worth understanding in detail. Pakistan’s dominant irrigation method is flood irrigation, which loses approximately 75% of water to evapotranspiration. The conventional trigger for watering is the concept of wattar — when the topsoil looks dry, you flood. But mango tree roots sit deep underground, and topsoil dryness tells you almost nothing about what the plant actually needs.

Farmdar’s solution is a plant stress report derived from satellite imagery. “We look at plant moisture levels. You can segment your land very easily — this acre needs water, this acre does not. You are still flood irrigating, but you are intelligently flood irrigating rather than blanket flood irrigation.”

The satellite data is not ordinary photography. Each image contains up to 21 bands of light — near infrared, red edge, and others invisible to the human eye. “Think of it like a CT scan. What the human eye is not capable of seeing, we can pick out from the information contained within a satellite image.” From those bands, Farmdar’s models can identify plant moisture levels, nitrogen ranges (excessive, adequate, or deficient), overall plant health, and early stress signals that correlate with incoming disease — giving farmers advance warning before visible symptoms appear.

Muzamil pushes on how this works without any physical device in the soil. Bukhari’s answer is direct: “The best device you can put in a soil is a plant. It can tell you a lot.”

Two channels, one data strategy

Farmdar operates two channels. The direct-to-farmer channel offers five products — water stress reports, general health reports, nitrogen level reports, and others — all free to individual farmers. A farmer in Wazirabad with two acres registers, marks their field boundary on the app, and begins receiving weekly personalised reports. No subscription, no fee.

Muzamil asks the obvious question: what’s the catch? Bukhari is candid. “The farmer’s data is actually valuable. We are creating a repository of data and we know that at some stage we will be able to monetize that farmer’s data somehow.” The more farms on the platform, the better the predictive models, and the more valuable the aggregate dataset becomes — to input suppliers, food processors, insurers, and eventually policymakers.

The B2B channel serves agri-businesses directly. The clearest example is sugar mills. Farmdar can scan 50,000 to 100,000 acres and, with over 90% precision, classify every crop type growing in that area — sugarcane, wheat, corn, chilies, grasses. It can identify which sugarcane variety is growing where (important because recovery rates differ by variety), when it was planted, and whether it is a ratoon crop. Mills currently do this manually, deploying hundreds of field agents in summer heat with registers. “Human nature kind of gets the better of most people. We can do this remotely, digitally, at a much smaller cost — no comparison.”

The fragmentation problem

Muzamil raises something he has noticed across the agriculture startup space: many companies are trying to do everything at once — B2B marketplace, analytics, input supply, advisory, sensor manufacturing. Bukhari agrees without hesitation. “If you try to boil the ocean or reinvent the wheel, you will simply not be able to benefit the farmers.”

Farmdar’s answer is deliberate focus. They describe themselves as a test-and-diagnostic, digital-first company. Where partners already exist — agronomists, specialty fertiliser companies like Vital Agrinutrients, irrigation innovators — Farmdar connects farmers to them rather than building those capabilities itself. “We are a very open, collaborative organisation. We like partnerships.”

The B2B sugarcane focus is not arbitrary. Bukhari’s family has been growing and supplying sugarcane to mills for fifteen years. “We understand that segment really, really well. We have a huge amount of base knowledge as farmers of sugarcane.” That ground-level credibility shapes the product in ways that a purely analytical team could not replicate.

The data layer beneath everything

Later in the discussion, Muzamil pushes toward the larger ambition. He notes that Pakistan’s water crisis is often framed as a supply problem — build more dams, pray for more rain — when the data tells a different story. 80% of Pakistan’s water goes into agriculture. 75% of that is wasted through flood irrigation. “It’s not a metric of getting more water. It’s all a matter of just using that water for the right thing.”

Bukhari agrees and extends the point to food production globally. By 2050, the world will need 60 to 100% more food to feed a population approaching ten billion. Producing that extra food would require additional land eight times the size of Pakistan. “Food production is inefficient everywhere. It’s just that these problems are massively accentuated in Pakistan.”

Farmdar’s stated ambition is to be the base data layer on which the rest of the ecosystem builds. “We see ourselves as the base layer of data creation, on which multiple layers will come — some that we will add ourselves, and some that we would like the rest of the ecosystem to use.” The investor pitch analogy Bukhari uses is precise: “We want to be a Bloomberg for agriculture.”

The company recently closed a $1.3 million seed round led by Indus Valley Capital, with participation from Atif Mian, Shehryar Hydari, and Dev Sai Ventures, as well as LMKR — chosen specifically for their deep knowledge of both tech and agri-tech. Farmdar is also the only Pakistani company in Microsoft’s Growth X accelerator for the MENA region, where they ran two successful pilots against a challenge set by Koch Holdings, a Turkish conglomerate whose agriculture app serves roughly a million farmers. Koch has since asked Farmdar to quote for services in Turkey.

Pakistan’s unfair advantage

By the end of the conversation, Muzamil asks Bukhari to give two images of Pakistan’s agriculture twenty years from now — a best case and a realistic case. Bukhari’s answer is measured but genuinely optimistic.

Pakistan is already a top-10 global producer of sugarcane, wheat, rice, mangoes, dates, and oranges. The problem is that on yield per acre, it sits near the bottom 50 for all of those same crops. “If we can move from bottom 50 to twentieth, we will probably jump two or three places further in the global rankings. We have the ability to export all our food. We have been given the gift that we can feed the world’s people — and that is actually true, not just drama.”

He flags one structural issue that rarely enters the conversation: 60% of Pakistan’s agricultural workforce is women. “If we want to do something in agriculture, we have to influence them. We have to give them a stake in it. Put smartphones in their hands.” Without that, any productivity agenda is working with less than half the available human capital.

His overall optimism rating: 70 out of 100. “I would take that,” Muzamil replies. So would most people watching Pakistan’s agriculture from the outside.