How a data center's power deployment timeline decreased
- Client faces long timelines, high costs, labor shortages, and extreme on-site safety risk.
- A skid-based, “plug and play” power block - the PowerCab™ 2 - moves all complex integration from the field to the factory.
- SENS and the client standardized a globally deployable system. Delivery time came down, the footprint shrank, HVAC load eased, and crane-level maintenance made swaps fast and predictable.
Background:
How did SENS and a major hyperscaler partner up to re-engineer AI power infrastructure? By moving from a high-risk, 18 month “stick-built” process to a factory-packaged solution, the collaboration solved the core challenges of AI-era development.
For one of the world's largest hyperscaler data centers and clients of ours, a visit to the SENS factory floor led to a fundamental shift. Our team's expertise, combined with the client's collaborative mindset, resulted in a standardized, globally-deployable system that reduced delivery and setup time, cut material costs, minimized the physical footprint, reduced HVAC load and ensured reliable, "crane-level" maintenance.
A new blueprint for hyperscaler power
It all began in late 2023. We had a relationship with a major hyperscale client, and a new project was being procured as a loose-piece system. The critical moment occurred when the client requested a full customer witness test for the system.
Our SENS team explained that the client hadn't bought an integrated system, but rather loose pieces for them to assemble themselves. We agreed to assemble the system in our factory to demonstrate the process.
The hazards were immediate, as crews had to muscle 300-pound sodium nickel chloride “salt” batteries into a third-party rack nicknamed “the Pillar of Salt”. The 18-hour initial charge sequence nearly failed. By putting the batteries into an integrated system they are factory tested, eliminating the risk of a bad battery during start-up.
When the client's team arrived, our SENS team was transparent about the challenging process. Then, we pivoted. We walked them to another part of the factory floor to show them the first prototype of our PowerCab™ 2, because we wanted to see if they were open to collaborate on what we thought was a better solution for them.
The contrast was profound. We explained the value of the factory-integrated model: the batteries would arrive pre-installed, pre-wired, and fully tested, eliminating all the risks they had just witnessed.
This was the turning point. We were able to get the client out of their old paradigm by showing them a new solution. In that moment, we transformed from a component vendor into a trusted expert collaborator.
Why traditional integration fails
The loose-piece, stick-built model is a key blocker for getting data centers onto the grid fast enough, as the tech inside can become outdated, requiring a pricey mid-project upgrade. This model involves procuring individual components from multiple vendors for a complex, high-risk, on-site integration and proves to be a major bottleneck. Its flaws are threefold.
First comes the Timeline Risk: stick-built projects are linear and fragile. Traditional builds take 18 to 24 months and slip easily. That cadence does not fit AI.
Second, Cost & Quality Risk: on-site work varies by crew and conditions. Skilled labor is scarce. Prefabricated systems are consistently cheaper on a total installed basis and deliver repeatable quality.
Third, Safety Risk: field integration of high-energy systems exposes teams to arc flash and electrocution. Incidents can injure people and destroy equipment.
This is why the better fit for AI is prefabricated, skid-based deployment to move the risky work into a controlled factory. That is the philosophy behind SENS PowerCab™ 2, a fully engineered DC power system that ships integrated, wired, and tested.
From loose pieces to a packaged power block
Within days, the client sent an engineer, an NDA was signed, and collaboration began. The engineer rolled out a drawing, circled the entire power section, and asked for our expert recommendation on the best way to package it.
The goal became clear: move everything outdoors and purchase under one part number that would work in all of their data centers worldwide.
This wasn't a simple order, but a fluid, back-and-forth design process that had to manage "many cooks in the kitchen," including integrators, OEMs, and end-user specs. The client drove the high-level requirements, but trusted our team of experts to find the specific solutions. We questioned the necessity of certain switches, streamlining the design, while also suggesting new reliability features.
When requirements changed, such as a new request for an additional genset connection or for direct IP-based battery communication, our team engineered smart, custom solutions to make it work.
The ultimate goal was a global "standard building block" robust enough for any weather, from the boiling desert heat of Nevada to the freezing snow of Norway. This led to innovations like raising vents to avoid snow buildup.
The result was the largest, most innovative PowerCab™ 2 ever built.
This new design clarified SENS’s critical role: located in the primary high-power substation, our system supports virtually all critical power between the utility and the data hall. This includes providing resilient 120Vdc power in the primary high-power substation and the electrical yard, supporting all power for cooling systems, and ensuring reliable genset starting.
We packaged all components for these functions into a single, resilient, "plug-and-play" system that permanently solved the "Pillar of Salt" problem."
A true partnership: trust & transparency
This partnership was immediately tested. After we had negotiated a six-month lead time, the project went dark. Then soon after the inevitable call came: the client suddenly needed the system on a new, almost impossibly short, deadline.
We accepted the challenge but stayed honest about constraints. That level of transparency moved the work from penalty clauses to shared goal, which would prove its advantages pretty quickly, as we encountered a supply bottleneck of transfer switches with long lead times. After escalating without success, we told the client the supplier’s name. This had the client using its supply chain muscle and the next day, the switches were confirmed to arrive just in time.
The client then even skipped the final witness test, trusting the factory-tested promise and shipping directly to site.
“This collaboration clearly demonstrated that packaged power systems not only speed up the deployment of individual data centers; they also enable true plug-and-play capabilities. When designed right, this allows hyperscalers to duplicate the same power infrastructure repeatedly, significantly accelerating all future deployments.”
- Sam Coleman
SENS, Chief Technology Officer
Winning the race to the grid
Commissioning in September 2025 confirmed the strength of the factory model. Any minor transit issues were handled on site by SENS technicians, but the dangerous and complex integration work - lifting tons of batteries, wiring high-voltage circuits, verifying system operation - had all been completed months earlier in the controlled environment of the factory. Where a stick-built project would just be entering its riskiest phase, this one was already running.
This approach addresses the two major constraints in getting data centers onto the grid: velocity and scalability. In a market where grid connections can take years, no company can afford another two years of site construction before going live.
Factory-built systems can be manufactured in parallel with civil works so that when the interconnection is ready, the power infrastructure is too. At the same time, the PowerCab™ 2 model establishes a repeatable, modular microgrid block that can be replicated across dozens of sites with predictable performance and minimal risk.
A template for hyperscaler infrastructure
For our client, procurement collapsed from dozens of part numbers to one and integration shifted from field labor to factory engineering. Quality became consistent, risk predictable, and timelines measurable. But more importantly, the relationship changed: instead of rigid buyer and supplier roles, both sides operated as partners solving a shared problem at the edge of industrial constraints resulting in a symbiotic partnership built on transparency and shared risk.
Featured Articles
Grid at risk: Why quick fixes won’t keep the lights on
May 28, 2025
Lessons from the Iberian blackout: why “more inertia” is not the answer
Jul 16, 2025