Connect with us

AUTO

Pateo Offers Yunji a Car-Grade Line for Robots

Pateo is offering Yunji auto-grade module making and chip buying, a quieter robotics supply play than robots on the factory line.

Published

on

Pateo Yunji partnership robots in auto factories

Pateo and Yunji Technology signed a partnership on Aug. 14 that puts a cockpit-electronics maker into the robot-parts trade. Gasgoo, the auto-industry publisher that first carried the news from Munich, said the Hong Kong-listed pair want Yunji’s service robots on auto production lines and Pateo’s electronics inside those machines.

The factory-floor language is what every brief repeated. The work that can start without a named plant is quieter: joint chip buying, contract building of robot control modules, and Pateo’s speaker-free sound tech on robot bodies.

What the Pateo-Yunji Partnership Covers

The Aug. 14 pact covers factory-robot trials, shared chip purchases, contract manufacturing of robot control modules, cockpit-style sound on robot shells, and joint research on an OpenHarmony control split. Pateo’s newsroom said the two firms signed a strategic cooperation agreement that week. The companies did not publish a contract value, a plant name or a start date.

Pateo Connect Technology, the Shanghai firm listed as 2889.HK, builds smart-cockpit systems. Beijing Yunji Technology, listed as 2670.HK, sells wheeled service robots mainly into hotels. Gasgoo said they will work across five tracks the firms named themselves: joint technology work, shared sites, a secured parts supply, joint sales, and partner-network tie-ups.

FIVE TRACKS IN THE PACT

  • Factory trials: The firms will try robots inside auto plants and on parts lines for material handling and assembly help, then sell a working package to other industrial buyers.
  • Skin sound: Yunji said it will put Pateo’s smart-sound system on robots first, swapping ordinary speakers for what the companies called skin-based sound.
  • Chip buying: They will buy key materials such as chips together, a supply move that matters more to the smaller robot firm.
  • Module making: Pateo will use its precision electronics plants to provide OEM and EMS work on robot control modules, aiming at lower part cost and tighter quality.
  • OpenHarmony stack: They will jointly research a “brain-small brain” control split on OpenHarmony for embodied machines.

That last software track sits beside hardware Pateo already knows how to ship at car-plant volume. The EMS line is the piece a cockpit supplier can run from an existing board plant, even if no robot ever rolls down a BYD or Geely aisle under this banner.

Pateo’s Cockpit Line Still Pays the Bills

Pateo’s 2025 annual report, filed with Hong Kong’s exchange after the firm listed on Sept. 30, 2025, recorded RMB3.51 billion of 2025 sales, up about 37 percent from 2024. It shipped 1.30 million smart-cockpit domain controllers, up more than 40 percent. Gross profit was RMB434 million, or 12.4 percent of sales. The statutory loss for the year was RMB1.09 billion; the adjusted loss narrowed to RMB240 million from RMB352 million in 2024.

Chairman Ying Zhenkai wrote that Pateo holds a 25.18 percent share of mid-to-high-end cockpit domain control among Chinese domestic brands, first in that slice, and has served more than 200 vehicle models. CMB International, in a March initiation note, called Pateo China’s third-largest cockpit-controller shipper by 2024 volume, with about 0.9 million units and a 7.3 percent share that year, and named design wins that include a Korean automaker and Porsche.

2025 snapshot Pateo (2889.HK) Yunji (2670.HK)
Sales RMB3.51 billion RMB301 million
Result Adjusted loss RMB240 million Loss before tax RMB295 million
Core unit 1.30 million cockpit controllers Hotel robots, 13.9 percent hotel-agent share
Factory angle Ruian plant named for robot parts Factory-scenario sales RMB15.1 million

Yunji’s whole year of sales is less than a tenth of Pateo’s. A hotel-robot firm that size cannot match auto-grade purchasing or board-level process control on its own, which is why the EMS and chip clauses read less like a press flourish than a capacity swap.

We anticipate that 2026 will mark the inaugural year of large-scale deployment of automotive-grade AI Agents, with the paradigm shift from “functional cockpits” to “AI-native intelligent cockpits” expected to be completed around 2028.

Ying Zhenkai, Chairman, Pateo 2025 annual report

That is still a car-cabin thesis. Robotics is a second door from the same electronics stack, not a replacement for the controller business that actually bills OEMs.

Yunji’s Factory Sales Are Real and Tiny

Yunji is not a humanoid startup. Founded in 2014, it listed in Hong Kong on Oct. 16, 2025 as a Chapter 18C specialist-technology name, and its English site still leads with hotels, hospitals and factories. The company’s own 2025 results release said total sales rose 23.1 percent to RMB301 million, hotel work brought in RMB244 million, and factory scenario revenue soared 393 percent to RMB15.1 million.

That factory figure is the sleeper inside Yunji’s own books. It proves the firm already sells into plants. It also shows how small that line remains: RMB15.1 million is about 5 percent of Yunji’s 2025 sales, and Yunji’s Hong Kong filing extracts showed a loss before tax of RMB295 million for the year. Frost & Sullivan, cited in the same results release, put Yunji first in China’s hotel-scenario robot-agent market with a 13.9 percent revenue share. Founder, chairman and president Zhi Tao said the company “will continue to increase R&D investment, promote industry standards, and attract global talent.”

The machine Yunji would actually roll into a car plant is closer to a hotel delivery bot than to a biped welder. Its modular UP composite robot uses swappable compartments on one chassis, which is how a vendor sells delivery, disinfection or reception from the same base. The GOGO and Run lines handle in-room drops and lobby runs. Those wheeled units can move totes and kits. They do not replace a six-axis arm on a body shop.

Xiaomi has already put humanoid robots into its own car factory for nut loading and tote handling, per 36Kr’s report of remarks by Lei Jun. Pateo does not own a vehicle plant. It can only invite Yunji into the factories of the 50-plus car brands it already supplies, which is a slower door than Xiaomi’s.

Ruian Was Already Pointed at Robot Parts

The manufacturing offer did not appear from nowhere in August. Pateo’s own newsroom and Gasgoo both described the Ruian smart-manufacturing base, which started production on June 23, as a cockpit-controller plant that would later add core parts for smart robots. That sentence, two months before the Yunji signing, is the through-line.

PATEO’S SUMMER SHIFT

  1. June 23, 2026: Ruian production line starts; the company says the site will expand into core components for smart robots.
  2. July 15, 2026: Pateo joins the application working group of MIIT’s humanoid-robot and embodied-intelligence standards committee, as a rare cockpit supplier in that room.
  3. August 7, 2026: Pateo agrees to buy about 70 percent of Chengdu Mingyi, a communications-chip designer, for no more than RMB1.4 billion, its first major deal since listing.
  4. August 11, 2026: A leading Chinese NEV maker picks Pateo for a high-end AI cockpit on Horizon Robotics’ Starry chip, which Gasgoo called the first such designation on a fully domestic cabin-driving chip.
  5. August 14, 2026: Pateo and Yunji sign the robotics pact.
  6. August 19, 2026: Pateo says a leading NEV maker picked it for Qualcomm’s 8397 cockpit platform, covering that customer’s cockpit line with 8295 and 8397 together.

Chip supply is the hinge between the Mingyi buy and the Yunji clause on joint procurement. Pateo already ships on Qualcomm’s 8295 and 8397 and Huawei’s Kirin 9610A, a dual-stack few cockpit houses can claim. Yunji, buying chips at hotel-robot volumes, gets a larger partner at the negotiating table. Pateo gets another socket for boards it already knows how to yield.

OpenHarmony Work Targets a Brain and Small Brain

The software clause is easier to over-read than the EMS clause. Gasgoo said the pair will jointly research a “brain-small brain” architecture on OpenHarmony, the OpenHarmony open-source system incubated by the OpenAtom Foundation. In that split, a high-level planner handles perception and task choice, and a faster motion layer keeps balance and joints on time.

That split is now the default in a lot of Chinese humanoid work, and it is also the one some new labs are trying to flatten. A single model that maps cameras straight to joint commands would make a two-layer “brain and small brain” stack look dated. Pateo and Yunji are still funding the two-layer version, on an OS that already wants robots to talk to phones, appliances and industrial terminals over one distributed bus.

Pateo appointed Dr. Gu Jinyu as an independent director on Jan. 29, 2026. CMB International’s March note tied him to OpenHarmony embodied-intelligence program work, which is why a cockpit house talking about Harmony on robots is not a random slogan. Pateo has also signed with Shanghai Jiao Tong University’s Artificial Intelligence Operating System Institute and listed technical work with NVIDIA, Horizon Robotics and SenseTime. Gasgoo added that Pateo has a project designation from another unnamed domestic robotics “brain” company.

OpenAtom’s own M-Robots OS, built on OpenHarmony and led by Shenzhen Kaihong, is already being sold as a robot operating layer with hard real-time claims. Pateo and Yunji did not say they will ship on M-Robots OS. They said they will research the two-layer split on OpenHarmony. That is a lab item until a robot boots it on a line.

Cockpit Orders Landed Five Days Later

Five days after the Yunji signing, Pateo told the market it had won the 8397 cockpit platform at a leading NEV maker, with the chipset named as a standard fit and the older 8295 work still in place. That filing is a reminder of which invoice still moves the share. Robotics did not pause the cabin business; the cabin business barely paused for robotics.

Coverage of Yunji’s first-half 2026 results, published in the week after the pact, put sales near RMB188 million and the net loss still above RMB100 million. Those H1 figures come from results round-ups, not from a filing opened for this article, and the reported year-on-year growth rates do not match across outlets, so they stay in that bracket. The 2025 audited picture is the one that holds: sales up, factory work up from a small base, losses still large.

WHAT CAN START WITHOUT A NAMED PLANT

  • Control modules: Ruian can tool robot boards on the same electronics discipline used for cockpit domain controllers.
  • Shared chip buys: Joint purchase orders do not need a live assembly-line robot to begin.
  • Skin sound: A speaker-free transducer pack can be designed onto a UP-class shell in a lab before any automaker hosts a pilot.

The factory trial is the photogenic clause and the slowest one, because it needs a carmaker to open a running hall to a hotel-robot vendor. The module line is the clause that matches what Pateo already built in Zhejiang. Until a named plant and a unit volume show up in a filing, that is the part of the Aug. 14 pact that can move first.

Disclaimer: This article is news reporting and analysis of a corporate partnership and of figures drawn from company filings, newsroom posts and results releases. It is for information only and does not constitute investment advice, a solicitation to buy or sell any security, or a recommendation on Pateo, Yunji or any related stock. Readers should consult a licensed financial adviser or broker who can review their own objectives and risk limits before making any investment decision. Sales, losses, shipment counts and project designations reflect the cited sources as of Aug. 23, 2026 and may change with later filings or contract updates.

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *

AUTO

BMW’s China Strategy Comes Due at €60 a Share

BMW’s 1-3% auto margin and €60 target price the China strategy, even as Neue Klasse output ramps in Hungary.

Published

on

By

BMW China sales slump Neue Klasse margin cut

RBC Capital Markets cut its BMW price target to €60 from €62 on August 14, a whisker from where the ordinary shares already trade. Analyst Tom Narayan kept a Sector Perform rating and pointed to weak Chinese deliveries plus tougher competition in Europe. Automotive EBIT margin in the second quarter was 2.3% after China deliveries fell 30.2%.

The cut lands on a company that changed chief executives on May 14 and, five weeks later, tore up its 2026 profit guide. Cost cuts are in motion. The China profit engine that funded the last decade is the part that broke.

BMW Now Guides a 1-3% Automotive Margin

On June 16 the board cut 2026 auto margin guidance to a 1-3% corridor, from 4-6%, and switched group profit before tax from a moderate decline to a significant one. Automotive return on capital is now 1-5%, down from 6-10%. Deliveries, once seen flat on 2025’s 2.46 million cars, are guided slightly lower. A one-time charge for faster restructuring will hit in the second half. Free cash flow in the car business is still seen above €2.5 billion, and the 30-40% dividend payout plus the buyback were left intact.

The July 30 half-year release confirmed that print. Group pre-tax profit in the second quarter was €1,697 million, down 35.1%, on revenue of €31,259 million, down 7.9%. Automotive EBIT was €629 million, down 60.7% from €1,602 million. Segment free cash flow was €513 million, down 73.4%. For the first six months the group earned €4,045 million before tax, down 29.4%, on revenue of €62,266 million.

Measure Q2 2026 Change vs Q2 2025 2026 guide
Automotive EBIT margin 2.3% 5.4% a year earlier 1-3% (was 4-6%)
Automotive EBIT €629 million -60.7%
China deliveries 117,815 -30.2% Group deliveries slightly down
Group profit before tax €1,697 million -35.1% Significant decrease
Automotive free cash flow €513 million -73.4% Above €2.5 billion for the year

CFO Walter Mertl said the firm banked €2.5 billion of cost saves last year and is now speeding that work up, with more structural cuts on top. Research spending in the first half fell 7.6% to €3,714 million. Capital expenditure dropped 30.5% to €1,900 million. Last year still produced €10.2 billion of group pre-tax profit on €133.5 billion of revenue, with 154,540 staff at year end. The 2026 car margin band is a different business.

Four Weeks After the Handover, Guidance Collapsed

At the close of the May 13 annual meeting, Oliver Zipse handed the chair to Nedeljković. Zipse left after 35 years at the company and seven as chairman, a run the supervisory board dates to August 2019. Milan Nedeljković, 57, had run production since 2019. He joined as a trainee in 1993, and his new contract runs through 2031. Dr. Raymond Wittmann, 47, took the production seat the next day.

Zipse told shareholders it had been a great honour to lead the group for seven years. Supervisory board chairman Dr. Nicolas Peter said Zipse’s name would stay tied to Neue Klasse. Nedeljković’s own line at the meeting was that profitability and speed are crucial. Five weeks later he had to say the rest out loud.

We have strong product momentum: With the NEUE KLASSE, we will put the strongest BMW portfolio in history on the roads over the next two years. At the same time, we will adapt our current structures and processes to the drastic downturn in market conditions. It is our entrepreneurial responsibility, therefore, to significantly intensify and accelerate our ongoing measures. It’s all about speed and efficiency.

Milan Nedeljković, Chairman of the Board of Management, BMW AG, June 16 2026 guidance statement

The company reached a deal with the works council on a workforce programme that includes voluntary severance. Separate reporting has put the German job reduction at around 8,000 roles by the end of 2027. Bosch has talked about cutting up to 22,000 jobs worldwide and ZF 14,000 in Germany. German auto employment stood at 691,500 on June 30, down 42,300 or 5.8% year on year, the lowest since 2005, according to the figures carried in the same industry tally.

Shenyang Was Supposed to Stabilise China This Year

Zipse’s last speech as chairman still treated China as a problem that 2025 had already boxed in. He said the aim last year was to stabilise sales in the world’s largest car market, and that BMW had done so, then claimed the brand beat a shrinking total market in the first quarter of 2026. The recipe he left on the table was local relevance, speed, and software: plants in Shenyang, a new battery joint venture with BMW Brilliance Automotive, and work with Momenta, DeepSeek, and Alibaba Banma. The slogan was “In China. For China. With China.”

April’s Auto China show in Beijing put long-wheelbase versions of the iX3 and the i3, plus a new 7 Series with Neue Klasse tech, in front of local buyers. Zipse called those cars more consistently tailored to China than anything BMW had built before. Reuters wrote in mid-July that the firm was trying to revive China after two years of falling sales, and that the local electric race may have moved on without it. BMW’s own 2024 China deliveries were already down 13.4%.

THE CHINA TURN

  1. Financial year 2025: Zipse tells the AGM that China sales were stabilised and that BMW beat a weaker total market in Q1 2026.
  2. April 2026: Long-wheelbase iX3 and i3, plus a Neue Klasse 7 Series, premiere in Beijing as the China-specific product answer.
  3. June 16, 2026: The board warns that the China slump accelerated in Q2, above all for non-electric cars, and that Europe and the US cannot offset Asia-Pacific.
  4. Q2 2026: China-region deliveries fall to 117,815, down 30.2%; the first-half tally is 261,773, down 20.4%.

Group deliveries in the second quarter were 590,947, down 4.9%. Europe rose 7.6% in the quarter and 5.4% in the half; the United States rose 11.9% and 3.9%. That split is the opposite of the decade when China carried premium margins. Discounting combustion cars in China is not reversing the unit drop, and BMW’s own electric volumes there remain a rounding error beside local brands that still command the price.

Volkswagen’s China deliveries fell 37% in the same quarter, Bloomberg reported, so Munich is not uniquely unlucky. BMW still cannot isolate itself, which is exactly what the June note said. The China Passenger Car Association had already cut its full-year market forecast again the Monday before that warning. A Middle East conflict was named as a second drag, through energy costs and weaker consumer mood, but the unit hole is in Shenyang’s sales region.

Why BMW’s Automotive Margin Fell to 2.3%

BMW printed a 2.3% automotive EBIT margin in the second quarter, down from 5.4% a year earlier. Import duties in the United States and the EU took about 1.25 percentage points off that figure, and extra depreciation on BMW Brilliance Automotive assets from the purchase-price allocation took about 1.2 points, the company said on July 30. Add those two items back and the margin still trails last year’s 5.4%.

THE 2.3% MARGIN

  • Reported auto margin: 2.3% in Q2 2026, against 5.4% in Q2 2025 and 3.6% for the first half.
  • Duty drag: About 1.25 points from US and EU import-duty costs in the quarter.
  • Brilliance write-down: About 1.2 points of extra depreciation on BBA assets from the old purchase-price allocation.
  • What is left: Even after those two items, the year-on-year margin gap is not closed, and auto EBIT in euros is still down 60.7%.

Zipse had already told the May meeting that tariffs alone took about 1.5 points off the automotive margin in 2025, and that without them earnings would have been higher year on year. His policy line did not move. “Tariffs and protectionism ultimately hurt everyone,” he said, covering both EU duties on fully electric cars imported from China and the EU-US tariff fight. That is a hard position for a company that still builds cars in Shenyang and Spartanburg and sells them across those borders.

The same speech defended combustion engines and plug-in hybrids beside battery cars, and attacked the EU’s 2035 CO₂ rules for ignoring how markets differ. BMW, he said, beat the EU fleet target again in 2025. Independent monitors still have the group on the right side of the 2025-2027 test: the BMW pool 2g under its CO2 target for January 2025 through June 2026, per the ICCT, while the Volkswagen pool was 7 g/km the wrong side of its line.

Debrecen Built 50,000 iX3s in Nine Months

The industrial answer is running, just not yet in the market that broke the quarter. Plant Debrecen in Hungary, the first Neue Klasse factory, built its 50,000th iX3 in late July, about nine months after series production began in late 2025, which BMW calls the fastest ramp of any new plant in the group. A second shift started in February 2026, ahead of plan, because orders were already there. The July 30 release said the iX3 is on track for 100,000 orders since sales began, and that around one in three all-electric BMWs ordered in Europe has been an iX3.

Europe’s battery-car mix is the clean side of the ledger. In the second quarter the group delivered 81,500 fully electric cars in Europe, up 37.9%, so 31.3% of European deliveries were battery-electric against 24.4% a year earlier. Globally, 116,807 fully electric cars went out in the quarter, up 5.2%. MINI had a strong half, with 149,535 cars, up 11.7%, and 36.9% of those fully electric. The product is landing where the grid, the dealers, and the sticker price already work.

  • Debrecen, Hungary: First Neue Klasse plant; 50,000 iX3s in nine months; two shifts from February; the new electric iX3 built in Hungary is the car BMW can sell in Europe and, unlike the old China-made iX3, take into the United States.
  • Munich: Series production of the i3, the second Neue Klasse model, began at the headquarters plant on August 6 after about €650 million of rework; BMW has said the i3 start cuts manufacturing cost 10%, and from 2027 Munich will build only battery cars.
  • Shenyang: Long-wheelbase iX3 and i3 for China, shown in Beijing in April, with trial Neue Klasse production flagged at year-end 2025; those cars are not yet the volume that would fill a 30.2% hole.
  • San Luis Potosí, Mexico: Converted for Neue Klasse vehicles and high-voltage batteries, with cumulative investment of about €1.72 billion; standard-wheelbase iX3 output is slated for 2027.

Zipse put more than 50,000 iX3 pre-orders on the Europe book as of May 13, with a WLTP range of 805 km on the long figure (the press sheet for the iX3 50 xDrive lists 678-805 km). He said engineers drove Debrecen to Munich, 1,007 km, without charging. The i3 is claimed at up to 900 km WLTP and 400 km of range from a 10-minute charge on the 800-volt sixth-generation drive. More than 40 new and updated models are due by 2027, every one drawing on Neue Klasse tech. Demand, not the Hungarian line, has been the constraint so far, which is why a second shift arrived early. The open risk is the opposite in China: the local cars are late relative to BYD, Nio, and the rest, and the European cars cannot be the China answer.

Chinese PHEVs Took 28% of Europe’s Plug-in Market

Home turf is no longer a closed shop. The German auto industry association VDA has Chinese manufacturers at nearly 10% of the European market, up from 0.5% in 2021. Rhodium Group, in an April study, put China-made cars at 6.4% of EU sales for the prior full year and 9.3% in December, a different cut that counts Western brands built in China as well as Chinese badges. Volkswagen chief Oliver Blume in late July urged Brussels to move fast on extra duties for Chinese plug-in hybrids after those cars took 28% of European PHEV sales, Automotive News reported. Handelsblatt has said the Commission is looking at extending the BEV levies, which already run as high as 35.3% for some Chinese makers, onto PHEVs.

Dataforce analyst Matthias Schmidt told The Guardian in early August that Chinese firms are hitting a wall on pure electrics in Europe and will push PHEVs for the next year because hybrids sit outside the extra BEV tariffs. BMW has spent two years arguing the other way. A PHEV tariff would help on the dealer lot in Munich and Stuttgart. It would also sit badly with a company that still talks about open markets and that designed China-built electrics for export until Hungary came on stream.

Industry-wide auto profits fell 12% in the second quarter on a 1.2% revenue drop, a reminder that BMW is not the only premium house losing pricing power. The electric M Concept Neue Klasse shown at Monterey Car Week in mid-August is the halo, not the P&L. Spartanburg remains BMW’s largest plant and the reason the brand is still the largest automotive exporter by value from the United States, Zipse told the meeting. That US factory is also why Washington’s duties show up as 1.25 points on the Munich margin.

Ordinary Shares Have Slipped Below the €60 Target

RBC’s new target is no longer a floor sitting comfortably under the tape. The shares closed the week of the note around €59.60, 5.7% above the 52-week low of €56.40. By the following Monday they had traded near €58.24-€58.28, a few percent above that low, after a year-to-date drop in the high thirties. The stock was about 24% below its 200-day moving average at the time of the note and had slipped under a 50-day average that itself was near €60. Mean consensus on MarketScreener around August 14 was still outperform, with an average target well above RBC’s number, so Narayan is the skeptic on the sheet, not the house view.

Second-quarter earnings per share were €2.05, down from €2.85. That is the equity translation of a 2.3% car margin and a China book down by almost a third. The H2 restructuring charge has not printed yet. The long-wheelbase China cars have not reached volume. Munich only started the i3 on August 6. Those three facts are why a 1-3% full-year margin band can still get worse before Debrecen, Munich, and Shenyang show up in the operating line together.

Disclaimer: This article is news reporting and analysis of BMW’s public filings, management remarks, and third-party research, and it is for information only. It is not investment advice, a recommendation to buy or sell BMW shares or any other security, and it is not a forecast of returns. Readers should consult a qualified financial adviser or licensed broker before making any investment decision. Figures, ratings, and operating statuses are those given by the cited sources as of August 20, 2026, and they can change with later deliveries, guidance, or market prices.

Continue Reading

Trending