Black Lake Technologies is concentrating its factory-AI pitch on the sequence that begins when a customer order arrives and ends when a plant must coordinate production. Fortune describes the Shanghai company’s cloud-based factory-management software as tracking orders, schedules, materials, and output. It reports that Black Lake has added AI agents intended to read an incoming order, generate a quotation, and schedule production. Founder and CEO Yuxiang Zhou’s stated objective is to put AI further into manufacturing decisions, including the connections among design, engineering, production coordination, and factory capacity.
From Incoming Orders to Factory Decisions
On its official platform page, Black Lake Technologies names four components: CAD-to-Process Agent, Quotation Agent, Coding Agent, and Data Agent. These are company-listed agents and demonstrations, rather than independently benchmarked models. The CAD-to-Process workflow is described as accepting STEP files, PDFs, and images, grouping matching filenames into a task, and importing unscheduled work orders from a work-order system. The same page labels generated material as AI-generated content that should be reviewed carefully, a qualification that makes human checking part of the advertised workflow.
Black Lake says its order-intake process can interpret drawings, spreadsheets, and customer details, identify products, quantities, processes, materials, and delivery requirements, then turn demand into production tasks. Its illustrated sequence uses factory-specific inputs, including equipment capability, quotation rules, process standards, and production constraints. The company says that sequence can match equipment and process, produce a quotation, create a process plan, revise a schedule, and issue a fulfillment alert. Those steps describe the platform’s intended operation; they do not establish that the tools execute every decision without human involvement or deliver the same result at every factory.
Risk handling is another element of the proposed decision layer. Black Lake says the system draws on equipment status, material supply, production progress, and order changes, and can flag a delivery risk before recommending actions such as plan adjustments, material substitutions, or production acceleration. This operational-data focus is relevant to EastFrontier’s AI-manufacturing coverage. Black Lake’s available materials, however, do not provide independent evidence that those recommendations improved delivery or utilization for a defined group of customers.
The company’s scale statements also require careful boundaries. Black Lake’s site claims more than 40,000 factories served, more than 2 million users, coverage across more than 30 industries, and 95% decision accuracy across core agents. It does not define the 95% metric’s test set, period, or operational meaning. A World Economic Forum organization profile identifies Black Lake as founded in 2016 and says it was selected for the WEF MINDS program in 2025. The profile also repeats the company’s factory count and customer names, so those commercial figures remain profile or company claims rather than customer-side confirmation.
Selling AI Automation Through Industrial-City Couriers
Black Lake’s route to smaller manufacturers is as notable as its product map. The South China Morning Post reported in July that Zhou recruited food-delivery workers from industrial cities to sell the company’s AI automation to family-run factories in China. SCMP said the recruits completed a five-day boot camp focused on explaining how AI could address particular factory pain points. Its reporting characterized Black Lake, which it identified as Shanghai-headquartered, as targeting a market of millions of family-run factories, not as claiming millions of Black Lake customers.
The sales tactic is tied to access rather than a published conversion rate. In Zhou’s account to SCMP, the recruits’ delivery experience gave them familiarity with reaching factory owners. Neither SCMP’s report nor Black Lake’s available materials quantify workers hired, factories approached, or revenue from the channel. Its software-workflow focus is separate from the factory trials in EastFrontier’s Dongfeng coverage.
Fortune reports Black Lake’s claim that it reached profitability in 2025 and has since grown revenue at an annual rate above 60%. Those are company claims reported by Fortune, not audited financial results. A company-provided release distributed by Kyodo News says the platform has reached nearly 40,000 factories globally and names Tesla, McDonald’s, Mixue, GAC, and Nongfu Spring among its customers. The release supplies no contracts or customer-side evidence, so the named relationships and factory figure should be read as issuer-provided assertions.
Claims, Review Steps, and the Test of Scale
Black Lake supplies illustrative outcome figures, but their limits are explicit. Its platform page says users can achieve a new-product test in as little as two weeks, 25% higher capacity utilization, 41% less capital tied up in inventory, and 8% higher order gross margin. The page provides no named customer, baseline period, methodology, or independent validation for those figures. They are examples of the results Black Lake says its platform can support, rather than general operating benchmarks for factories using AI agents.
An Alibaba Cloud Marketplace listing adds a second set of vendor claims. It presents Black Lake Smart Manufacturing as cloud-based, IoT-enabled collaboration software for digitalizing shop-floor activity and real-time coordination. The listing advertises implementation in four to six weeks, a two-day startup claim, pay-as-you-go pricing, configuration, continuous upgrades, and integration capability. It also lists claimed benefits of 40% improved scheduling efficiency, 10% improved capacity utilization, 20% lower inventory, 50% less material waste, 35% lower reject rates, and 10% less equipment downtime. Alibaba Cloud does not provide a study design or denominator for those figures, so they cannot be treated as independently verified implementation outcomes.
The same marketplace page reproduces two customer testimonials: ChinaUST Group’s IT head said replacing a traditional manufacturing-execution system saved 7,000 man-hours, while JYD Technology’s production head said its order-delivery rate improved by 5% after one month. Alibaba Cloud’s listing does not independently audit either testimonial. Black Lake also offers Black Lake Mini Worksheet, which the marketplace describes as a low-code mobile application for SME frontline workers with work-order management and digital worksheets. The stated use cases include visibility into production progress, delivery delays, tracing production problems, and worker-performance calculations.
Zhou’s thesis, reported by Fortune and echoed in the Kyodo-distributed release, is that AI’s larger opportunity lies in altering how industries operate rather than simply taking over discrete tasks. Black Lake’s own product materials make that thesis concrete through order intake, quotation, scheduling, risk signals, and reviewable process outputs. The meaningful test is narrower and more measurable: whether customers can validate the claimed workflow gains, understand the decision-accuracy measure, and maintain human review when AI-generated process content enters a live factory system. Until customer-side evidence and defined performance measures emerge, Black Lake’s factory-AI strategy is best assessed as a detailed company proposition with reported distribution efforts, not as proof that industrial decision-making has already been remade.