Bright Machines of San Francisco announced the Hybrid BRC (Bright Robotic Cell), an extension of its Bright Factory platform that lets human operators enter a sensor-monitored robotic cell and perform prescribed assembly steps without breaking the digital production record that traces each server from its first screw to its shipping label.
Not just a minor hardware update
The company positions the Hybrid BRC as a direct response to a structural weakness in high-value electronics manufacturing: manual interventions create gaps in traceability at the moments when errors are most likely. In an interview with VentureBeat, CEO Sviat Dulianinov quantified the problem: "If you assemble modern AI servers starting with manual operations, your initial yield — first-pass yield — can be as low as 20%. Then you gradually ramp up and scale, and it can reach the 60s, 65% or so."
When individual AI servers can cost hundreds of thousands of dollars, and hyperscalers are losing time and money waiting for deployable infrastructure, those percentages matter materially. Bright Machines' Hybrid BRC aims to allow human involvement while preventing human-generated breaks in the production data thread.
Why manual steps create a data black hole
Automated assembly lines produce continuous streams of production data — torque values, placement coordinates, component serial numbers, inspection images. That data thread lets a manufacturer prove a server was built correctly and, if something fails months later in the field, trace the failure back to a specific station, step, or part.
When manual intervention is needed, manufacturers historically had two poor choices: stop the line entirely (hurting throughput) or pull in-process units to a separate manual workstation outside monitored data flow (creating a break in traceability right when human error risk is highest).
How the Hybrid BRC works
The Hybrid BRC embeds guarded access doors and safety panels into the production line. When an operator opens the door, the robot arm deactivates and on-screen instructions guide the operator through each assembly step. Meanwhile the cell's sensor suite — cameras, force-feedback sensors, and tooling sensors — continues monitoring for incorrect installs, missed steps, and wrong components, applying the same quality checks used in full automation. The serial-number-level traceability record persists from start to finish.
The yield gap between humans and robots
Dulianinov contrasted manual-assembly yields with automation performance: "At robotic operations, yield-per-station level is usually more than 98% with our technology, and even at the line level, we usually get to 97.5%, 97.7% or so." First-pass yield measures the share of units coming off the line correct the first time without rework. The difference between a 20% manual ramp and a 98% automated station is not trivial — for very expensive hardware it can mean the difference between profitability and failure.
Accordingly, the Hybrid BRC treats human operators as an exception handler rather than a substitute for automation. "The more human stations you introduce, the more you increase the risk of lower yields driving the overall yield down," Dulianinov said. "That's why we prefer to start at least with 50% automation, and then move to at least 80%." He added that robot throughput can be 50–100% faster than humans at the line level.
Assembly as a hidden bottleneck in the AI infrastructure race
Public discussions of AI infrastructure often focus on chips, power, and data center construction. Dulianinov argues assembly — turning chips and motherboards into racked, tested, deployable compute — is a substantial drag on deployment timelines. When production quality falls short, testing, rework and rebuild cycles can add months. He said Bright Machines' technology can cut those delays by at least a third.
An executive added that Bright Machines' servers are "flying out into production" rather than sitting in warehouses — an indicator that assembly capacity, not only chips or power, constrains hyperscaler deployment timelines. The consequence is asymmetric: the largest hyperscalers lose millions per day when servers fail or arrive late, so customers are buying assurance as much as boxes. An unbroken data thread has therefore become a product attribute.
Existing deployments and growth figures
The Hybrid BRC is already in operation, the company says. Dulianinov stated that Bright Machines runs a number of hybrid lines in the U.S. and has "built more than 10,000 compute nodes" using the new stations. This year, the company plans to manufacture "more than half a gigawatt of compute capacity."
The firm would not name customers, citing secrecy around data-centre-related IP. It did report customer growth of "more than 3x this year" versus the prior year, driven by what the company described as the intersection of "physical AI, AI infrastructure buildout, and onshoring." Bright Machines says it has deployed more than 130 microfactories across 10+ countries, served more than 60 customers, and produced more than 300,000 servers.
How Bright Machines positions itself versus other vendors
Asked to compare the Hybrid BRC's traceability claims with operator-guidance and inspection software vendors like Tulip and Instrumental, Dulianinov drew a distinction by business model: Tulip provides operator interfaces, Instrumental focuses on inspection — "pieces of the puzzle," he said. Bright Machines positions itself as a technology-enabled manufacturer that installs and runs entire operations: lines, software, data and people under one orchestration layer called Bright Insights.
He suggested the more apt comparisons are contract manufacturing giants such as Flex, Jabil, and Foxconn, which historically built at scale using manual labor that generated little traceable data. Bright Machines' differentiation is that robot data, sensor data, and now human-station data flow into a single environment.
Company history and funding
Bright Machines was carved out of contract manufacturer Flex eight years ago. The company previously planned a 2021 SPAC public listing at a reported $1.6 billion valuation, but the deal fell through. In June 2024 Bright Machines closed a $126 million Series C round — $106 million in equity led by funds managed by BlackRock with participation from Nvidia, Microsoft, Eclipse, Jabil and Shinhan Securities, plus $20 million in venture debt from J.P. Morgan — bringing total capital raised to over $400 million, according to the company.
Data ownership and worker monitoring concerns
Dulianinov addressed two governance questions. On data ownership he said customer-related inspection data and device-specific information would be protected and owned by the customer, while process and robotics data remain with Bright Machines to support continuous improvement across its platform.
On worker surveillance he pushed back on a negative framing: high-IP electronics floors already ban personal electronics in many contexts (aerospace, defense, government workloads). He argued workers often appreciate traceability because it supports security and provenance claims — for instance, proving that American-built infrastructure was assembled under the stated conditions — and that monitoring is about proving component-by-component integrity rather than watching employees.
Modular design and the onshoring argument
The Hybrid BRC's modular, software-defined cells are intended to be retoolable in days or weeks rather than months. Dulianinov said minor design changes within a product family can be introduced "within a day," though larger shifts (for example, air cooling to liquid cooling) remain more substantial.
His broader argument is labor arithmetic: the U.S. cannot scale manufacturing the way Shenzhen-scale plants do because it does not have the equivalent millions of factory workers. "So you need to solve it with AI software and robots, and that's our thesis," he said. "It's not just robots on the floor — it's also creating jobs. All the robots, and some people on the floor." Lior Susan, founder and CEO of Eclipse and chairman and co-founder of Bright Machines, summarized the message: "The future of manufacturing isn't choosing between automation and flexibility — it's combining both in the same digital production environment."
For all the talk of gigawatts and yields, the Hybrid BRC is an admission wrapped as innovation: even the most automated factories still need humans to reach in. Bright Machines is betting that the winners in AI infrastructure will not be those who eliminate the human hand, but those who never lose sight of it.
