A two-year lawsuit was filed.
Profit growth is far faster than revenue growth, and the secret lies in the volatility of the business structure.
In the first half of 2026, Tostada's low gross margin Intelligent Energy and Environmental Management Systems business revenue plummeted by 79.61%, down from 36.47% of revenue in 2025 to just 5.42%.
It's been two years of Tostad's proactive push.Concentrating products, shrinking projects Concentrating strategy transformationAs a result, as early as 2024, Tostada posted its first loss since going public.Chairman of TostaThank you very much.Admittedly, the sharp shrinkage of project-type operations led to a decrease in revenues and profits, but the data bubble on the surface showed a significant improvement in business structure and the mainstream sector continued to climb steadily.These numbers are the backbone of strategic decision-making within the knife blade, and the seeds for the next decade.
In 2025, he made it clear that Tostada's insistence on focusing on reducing emissions and crossing the deadlock of transformation was at the expense of short-term scale numbers and long-term core competitiveness. 2025 was defined as the culmination of Tostada's strategic transformation.
By 2026, the strategic tone has shifted from shrinking and reshaping aluminum to growing and producing aluminum.
Total growth in the three major industries, including industrial robots and automation application systems, jumped to the top of the business segment, generating revenue of 707 million yuan, an increase of 121.62%, accounting for 54.90% of the main business revenue. In the first half of the year, the number of industrial robot orders was about 6,800 units, and shipments were about 5,500 units.
In addition, injection molding equipment generated revenues of 258 million yuan, an increase of 12.86%; the numbering machine business generated revenues of 209 million yuan, an increase of 27.60%. It is worth noting that the demand for parts related to human robots has significantly increased, with the number of licensees for numbering machines increasing by 34.34% in comparison.
The overseas market was equally prominent. In the first half of the year, Tostada achieved overseas revenue of 427 million yuan, an increase of 28.20%, accounting for 33.12% of total revenue; the gross margin reached 30.78%, a significant increase of 15.07 percentage points. At present, Tostada products have been sold in more than 50 countries and regions.
In keeping with the underlying logic of reducing waste, which is consistent with Tolsta's consistent long-termism, Tolsta maintains the growth foundation of its product business while cutting off its low-grossing business. At the same time, the shift from project-based aluminum to aluminum product-based aluminum has cleared the key barriers to Tolsta's commercialization of smart devices.
Scenery + robots + data + AI
An ongoing business closure
After the decrease in the legal capacity, the construction of a closed loop bridge became a matter of urgency for the future.
In a 2026 semi-annual report, Tostas said:ConstructedWhat's going on here?ofClosed business。
At the end of the scene,relying on the deep accumulation of injection molding equipment and machine tools to identify high-value industrial application scenarios and define product functional needs;It's like a robot.Development of a comprehensive product range of rectangular coordinate robots, multi-joint robots, collaborative robots (including modular sensory joints) and humanoid robots;It's a data center.Collecting high-quality data from real-world operations through multi-sensor fusion technologies (space, vision and perception);The end of AI.Using data to continuously train and replace AI models, consolidate the competitive position, expand the wider industrial application landscape and ultimately move towards the entire commercial landscape.
In contrast to the current industry with physical intelligence, the biggest pain is not the lack of algorithms, not the lack of hardware, but the lack of real industrial scenarios. There is a huge gap between the simulation data in the lab and the real working conditions on the production line: robots perform well in the simulation environment, and in the workshops they don't give a shit.
The barrier to Tostas is right here:It has accumulated more than 200,000 potential customers and served 150,000 manufacturing customers.Over the past two decades, the manufacturing landscape has been deep-rooted by the very scarcity of real-world resources for the training of industrial, physical and intellectual equipment.
To make the most of its scenarios and data advantages, TOSTA has made several important advances in R&D:
Toaster United Flight Co., Ltd. has released a model of the Physiverse Industrial Physics WorldIn the past, it has been possible to build a continuous evolutionary model of high-precision physics modeling - AI efficient training - that can replace real-machine landing gear.
Toast Unity Without QuestionsConstruction of industrial transport scenarios physical AI data infrastructure closed-loop platformIt covers the entire chain from virtual simulation, data marking, data compilation, model training to real machine measurement, rapid replication of technical programs, cross-industry batch replication.
In technical architecture, Tostar uses the VLM semantic analysis + robotic task planning language + atomic skills to execute the puzzle. Through visual models, the task planning module converts natural language instructions into executable sequences of actions, ultimately achieving stable and reliable motion execution from the atomic skills library, allowing the large model to truly understand the instructions in the workshop, rather than just chatter.
Tostar is being rebuiltFrom machine-to-machine acquisition to model iteration, integrated computing platformIt covers the five major phases of AI data acquisition, data processing, model training, robot deployment and feedback to build a data base for the large-scale deployment of industrial intelligence.
Being physically intelligentIt's a three-way street.Point of view
The value of the closed loop will ultimately fall on the product. In the first half of 2026, Tostada's smart product matrix is moving from lithium verification lithium to lithium delivery lithium.
A humanoid robot with a small tooth，That's it.It is China's first artificial intelligence robot to be used to inject scenesAll processes have been automated in the molding scene, covering the process of removal, testing, transportation, etc.
CurrentlyToastedIt's moving.“It's small.”Small-scale verification of scenes such as 3C movements and landfills。In the 3C industry transportation scenario, STOP can solve site space constraints, automated transformation of landing difficulties and variation in the demands of the scenario, and the inadequacy of traditional schemes; in the landfill platform scenario, STOP can be used for precision components, various types of metal parts, plastic hardware, and other landfill processes, allowing for disordered sorting and solving traditional robotic flexibility problems.
The four-legged robot is wearing star-crossed jeans.，Ability to perform multiple tasks in different application scenarios, including autonomous inspections and specialized tasks in complex environments.
Tostada's body-smart robot workstation has been implementedCommercial DeliveryIn the meantime, we will continue to contribute to achievements.For high frequency manufacturing industries such as injection molding, 3C, flexible sorting and overlay workstations have integrated AI-based recognition and adaptation control capabilities. They are able to quickly switch lines under limited artificial intervention, and can be used for electronic components, hardware and plastic products sorting scaffolding.
Toaster's entry into the field of physical intelligence, not because of the wind tunnel, but because it saw the need for a large number of flexible installations and complex scenery selections in areas such as injection molding, 3C.First, identify one or two scenarios that can actually generate value, allowing robots to run in specific environments, and continue to iterate in the process.。
Involved in the electricity sector
AdditionallyUp there.A puzzle about an AI strategy
In addition to the product, Tostada is also working on a bigger chess set:The hardware layer of the AI strategy.
Four days after the publication of the semi-annual report, Tostada issued a foreign investment announcement: it plans to invest $147 million in the acquisition of equipment manufacturers in the field of optical communications to acquire 49% of the shares in Electroelectric, with preferential purchasing rights for the remaining 51% of the shares.
The company was founded in 2015 with a focus on fully automated light module assembly, chip testing and other equipment, and its products have been applied in bulk to the 400G/800G/1.6T high-speed light module mass production line.
Toaster industrial robots can be superimposed with laser-electric combined equipment to form differentiated solutions in the field of optical communications; Toaster acquired accumulated micron precision motion technology and advanced anti-breastfeeding robot precision execution technology; Toaster electronics can achieve expansion and reduction through Toaster's supply chain and scale manufacturing capabilities.
In other words, at the hardware level of AI infrastructure, high-end precision equipment, such as five-axis connected numerical control machines, light module precision coupling equipment, is the foundation for the AI infrastructure shelf.
What is the meaning of this verse?
There are a lot of players on the board, with different paths, on industrial robots and smart cards.
Tostar is not a champion of the wind, but a farmer; not a conceptualist, but a real warrior; not a lone warrior, but an eco-builder.
Looking back over the past two years, from the proactive shrinkage to profitability, from the decrease in sales to the decrease in sales, every step has proven that ups and downs are a normal part of growth.