On September 25, the Kazakhstan-China Investment Forum convened at Koktobe Hall in Almaty with President Kassym-Jomart Tokayev in attendance. On stage, UBTech Robotics and Kazakh firm Nero Group named a $440 million robotics localization project, sited inside the Alatau Park of Innovative Technologies special economic zone.
Educational robots first, humanoids in 2027
The project runs in two stages. Stage one is educational robotics, starting production in October, targeting 50,000 kits a year, supplying roughly 30 schools in Almaty by year-end and expanding to more than 200 schools by 2027-2028. Stage two is local assembly of service and humanoid robots, set to begin in the second quarter of 2027, targeting 1,000 units a year. The full production site covers 25,000 square meters. Nero Group has also lined up three local university partnerships, with Al-Farabi Kazakh National University, Satbayev University, and Almaty University of Power Engineering and Telecommunications, to build a pipeline for the R&D and programming roles the project needs.
Nero Group CEO Yerlan Nabiyev framed this as more than a production line landing in Kazakhstan: “This is an initial stage, which primarily involves localization not only of the production component. It is also R&D, research, programming, substantial job creation and, of course, production facilities.” Nero Group currently employs more than 200 people, and Nabiyev expects that headcount to grow materially as the business scales. The two companies started working together in 2025, and signed a formal cooperation agreement at a Kazakhstan-China forum this July, with Tokayev present. This week’s announcement turns that agreement into a physical facility.
Liuzhou runs at ten times the volume
LostJobs covered UBTech’s Liuzhou plant going live on September 13: annual capacity above 10,000 industrial humanoids, one unit rolling off the line every 10 minutes, built by humanoids doing the material handling. Put the two numbers side by side. Liuzhou produces industrial-grade humanoids at five figures a year. Almaty is starting with educational kits, with humanoid assembly not beginning until Q2 2027 and targeting 1,000 units annually, a tenth of Liuzhou’s run rate. This isn’t UBTech exporting Liuzhou. It’s UBTech exporting an earlier, lighter version of the same model, for the first time, outside China.
The timing connects to a second story. LostJobs covered China’s securities regulator freezing IPO review for humanoid makers on September 20: both AgiBot and Deep Robotics had listing applications shelved. A tighter domestic equity market hasn’t slowed China’s humanoid manufacturers down. UBTech is using a different channel, government-to-government investment forums and bilateral state agreements, to land production capacity overseas without touching public markets at all. The funding route changed. The pace and direction of expansion didn’t.
The labor read
At this stage, this is a job-creation story, not a job-replacement one. Nero Group’s headcount sits above 200, and the “substantial job creation” Nabiyev named maps to Almaty-based R&D, programming, and production-technician roles, with the three university partnerships built specifically to train the people who’ll fill them. The choice to name three specific universities, rather than gesture at “local talent” in the abstract, is worth noting on its own. A robotics venture that wants trained programmers and engineers within a few years has to start building that pipeline before the factory opens, not after. Al-Farabi Kazakh National University, Satbayev University, and the Almaty University of Power Engineering and Telecommunications are all established technical schools, not new programs created for this deal, which suggests Nero Group is drawing on Kazakhstan’s existing technical labor pool rather than importing one. That’s a different picture from Liuzhou, where humanoid robots handle their own production line. What’s being built in Almaty right now is educational kits for schools, several technology generations removed from genuine line-level automation replacing local workers.
The timeline itself is the thing worth watching. The educational-robot stage carries almost no controversy: it’s mature Chinese hardware being distributed into a Central Asian market that didn’t previously have it. The real inflection point lands in Q2 2027, when local assembly of service and humanoid robots begins. Whether the Almaty plant ends up replicating Liuzhou’s “robots building robots” model will determine how much of that 1,000-unit capacity turns into local technician jobs, and how much of it eventually displaces them. October is close enough that the first stage of this deal has almost no room to slip. If the education-robot line is running on schedule when LostJobs checks back next month, that’s the more immediate marker of whether this $440 million commitment holds up on its own timeline, before anyone has to answer the harder question about 2027.
