Following up on my Institutional Embedding thread from February. The situation on the ground has evolved past “should we engage institutions” into something more concrete and, I think, more interesting.
What happened
I went to the UNILAG AI UniPod — the facility I Photo dumped here. For those who missed it: This is the AI and Robotics hub at the University of Lagos, commissioned on April 7th 2026 by Nigeria’s Vice President. It’s part of a programme called UniPod that they plan on rolling out to 50+ universities across Nigeria.
I met with Prof. Chika Yinka-Banjo, the pioneer Director, toured the facility and hung out with students for a few hours.
Here’s what I found:
The good: The facility is genuinely impressive. Multiple Creality (and a couple massive 40x40 unidentifiable) FDM printers, a couple full stack Formlabs stations, full electronics benches (oscilloscopes, signal generators, bench PSUs, soldering stations), HP workstations, climate-controlled space, security. Everything you’d want to see in an equipped, functional, small-scale manufacturing facility. The students are smart and motivated, a few were designing UAVs with pixhawks and planning flight plans.
The reality check: Prof. Yinka-Banjo was candid with me. Beyond initial construction and outfitting, the UniPod has no official recurring funding channels. Same story for her AI and Robotics Research Lab (AIROL). No operational budget, No standing demand for the equipment, and volunteer/student staffing. The government basically built the thing, filled it with goodies, then F-ed off. Like a pinata salesman. So, the 3D printers and oscilloscopes etc are basically gonna be sitting there waiting for something to do.
This means the model I originally thought about — universities issuing purchase orders for Quiver units — probably wouldn’t work short to mid term for Nigeria, at least not as a primary pathway. They don’t have the procurement budget. Most of the other institutions on those spreadsheets I shared are likely in a similar if not worse position, UNILAG is the country’s top insitution afterall.
What I’m proposing instead
Rather than selling to the university, what if we manufacture at the university?
The UniPod has the infrastructure. It has the student workforce. What it doesn’t have is operational demand and a programme that justifies its existence to UNDP and TETFund (which run and sponsor the unipod scheme). Pan Robotics could provide that demand by running a Quiver, amongst other things, manufacturing line out of the UniPod facility.
This mirrors what Julius plans to do in Europe. Julius is forming a GmbH and building dev-kits as a bonded manufacturer. I’m proposing the same thing under AIP-009, but the manufacturing happens inside a university facility with student assembly labour — which has some specific advantages:
Labour costs. A skilled assembler in Lagos costs $400–800/month fully loaded. At 15–25 hours per unit, that’s around $40–125 assembly labour per Quiver, versus the $500–2,000+ it would cost in the US or Europe. This is a structural advantage, not a one-off saving.
Facility costs. The UniPod equipment already exists. Pan Robotics pays the UniPod something like a percentage of each unit sold as a facility fee. The more we build and sell, the more flows back to the UniPod programme. The PETG-CF enclosure parts come off their 3D printers at material cost (~$15) instead of the $50/unit we’re paying externally.
Market proximity. West Africa is a greenfield. Agriculture, infrastructure inspection, oil & gas, environmental monitoring, aerial nacho cheese experiments,— real demand, minimal supply. No open-source UAV / eVTOL manufacturer exists anywhere on the continent.
Contributor pipeline. Nigerian engineering students building Quivers become Arrow contributors, I’ve already sent a couple to joined the discord already. They’ll encounter conditions we’ve never tested, and that environmental data feeds back into the ecosystem.
The BOM
Going through Thomas’s cost-per-drone report, and as the supply chain is overwhelmingly Chinese (Hobbywing, Holybro, SIYI, Nanoradar, JLCPCB), it could mean shorter and cheaper shipping to Lagos than to the US. Frame metal components and fasteners could be source and worked locally for cheaper aswell.
Base Kit — ~$3,560 per unit
| Subsystem | Cost/unit (10-unit batch) |
|---|---|
| Airframe (CNC 6061 + CF tubes + landing gear + connectors + shock absorber) | $579 |
| Propulsion (Hobbywing XRotor X6 Plus G2 ×4 + spare props) | $594 |
| Avionics (Holybro Pix32 v6, RPi5, H-RTK F9P, Matek M9N, Nanoradar NRA15, SIYI A8 mini + HM30 + spare antennas, GigaBlox Nano ×2, CubeNode ETH, DB201 Remote ID) | $1,727 |
| PCBs — Main + FC + BC (JLCPCB assembled) | $507 |
| Harness / connectors (XT60, Phoenix terminals, Molex housings, attachment interface cables ×36, AMXL-150 fuse, JST GH pre-made, LED power switch) | $115 |
| Fasteners / hardware (rivets, screws, washers, inserts, nuts, Loctite, hinges, latches) | $160 |
| 3D-printed parts (PETG-CF, ~1.5 kg) | $50 → ~$15 at UniPod |
Optional Add-Ons — ~$2,795
| Add-on | Cost |
|---|---|
| RPLIDAR S2L (obstacle avoidance LiDAR) | $282 |
| Nanoradar MR82 (forward-facing radar) | $291 |
| Tattu 14S 30000mAh battery | $735 |
| SKYRC PC3000H charger | $609 |
| RadioMaster Boxer TX + battery + receiver | $268 |
| Nanuk 976 transport case | $610 |
Totals
| Configuration | Cost per unit |
|---|---|
| Base kit | ~$3,560 |
| Fully loaded (base + all add-ons) | ~$6,355 |
| Landed Lagos estimate (inc. 10–15% customs) | ~$4,100–7,300 depending on config |
Assembled and tested units could sell at say, $8,000–11,000, higher for fully loaded builds with battery, charger, TX, and transport case. That gives margin for assembly labour, facility fees, protocol fees, overhead, and profit.
The AIP-009 structure
Same framework as the European dev-kit, adapted:
-
Pan Robotics operates as the manufacturer entity
-
10% protocol fee on hardware sales settled as $ARROW token purchase, per AIP-009
-
5% on service engagements (training, payload development, support contracts)
-
Pilot period fee waiver requested for the first 12 months, per AIP-009 incentive provisions
-
Quality standards, build logs, and transaction reporting per Arrow standards
-
Happy to discuss the bond structure — same concept as Julius’s, adapted for the context
The ask
1. Node designation. Formal approval as a licensed manufacturer node under AIP-009, covering West Africa.
2. Pioneer units. I’d like to request 1–2 Quiver units or completed sub-assemblies be sent to the UniPod. The purpose is twofold:
(a) I want one so we can disassemble, document, and reassemble to validate and train the manufacturing process
(b) The Unipod/AIROL could use them to run experiments, build out payloads, run tests and generally use the system, and
(C) Have units that stay at UniPod as a semi-permanent demo — visible and presentable to every university administrator, government official, and researcher who walks through the door of Nigeria’s most prominent AI facility. That’s market seeding we can’t buy frankly.
What the DAO gets back
-
Protocol fee revenue from a new geographic market
-
Environmental telemetry from tropical operating conditions
-
Contributor base expansion — students and researchers in Nigeria becoming part of the community
-
Proof that AIP-009 can scale beyond Europe and the USA — a new node, different continent, different regulatory environment, same protocol
What the UniPod gets
-
Standing operational demand for equipment that’s currently idle
-
Paid student workforce pipeline (Pan Robotics pays, not the university)
-
Revenue from a percentage of each unit sold — directly tied to production, not a fixed cost
-
Research outputs from production and drone use data (publishable, novel, Africa-first)
-
The narrative they need for UNDP: “Our UniPod is manufacturing open-source eVTOL drones”
Happy to answer any questions, share more detail on the facility, or walk through it in on a meeting call.
What do folks think ?