As artificial intelligence reshapes the demands placed on global infrastructure, the question of how to cool the machines that power it has become one of the defining engineering challenges of the era. LONGWELL, a cooling technology company, has answered with its LWBE3G FanWall system — developed in just 90 days — which reduces fan energy consumption by 38 percent and meaningfully improves the efficiency ratios that determine whether a data center is sustainable or merely functional. In a moment when power grids strain under the weight of AI's appetite, innovations like this one carry implicat
LONGWELL's FanWall tech cuts AI data center cooling energy by 38%
Efficiency innovations will likely drive the next wave of data center development.
Mark
So this is a cooling fan system. Why does that matter enough to write about?
Mimi
Because data centers use enormous amounts of electricity just to keep servers from overheating. If you cut that by 38 percent, you're talking about real money and real environmental impact at scale.
Luke
But we should be clear—this is one company's claim about one system. Do we have independent verification of that 38 percent figure?
Mimi
The source says it was developed with a top global OEM and improved their PUE from 1.42 to 1.28. That's a measurable metric.
Luke
Right, but we don't know which OEM, and we don't know if those numbers come from a controlled test or real-world deployment. The orders through 2027 are real, but we don't know how many units or what the actual installed base looks like.
Mark
What makes this different from other cooling solutions?
Mimi
It's designed specifically for AI workloads, which generate 60 kilowatts or more per rack. That's a different beast than traditional server cooling.
Luke
And it was built in 90 days, which is fast. But fast development can mean different things. We don't know if that's a strength or a risk.
Mark
So the market is adopting it. What does that tell us?
Mimi
That data center operators believe it works and that efficiency is becoming a competitive factor. Power constraints are real, especially as AI grows.
Luke
Orders are a good sign, but we should watch for actual performance data once these systems are running in production. That's when the real story emerges.
Le Pouls
- AI processors now demand more than 60 kilowatts per rack, and legacy cooling systems were never designed to absorb that kind of heat density without hemorrhaging energy.
- The gap between what data centers consume and what they actually need for computing has become a financial and logistical crisis, with PUE ratios exposing billions in wasted overhead.
- LONGWELL compressed what might have been a years-long development cycle into 90 days, racing to meet an industry that cannot afford to wait for incremental progress.
- The FanWall system pushes PUE from 1.42 to 1.28 while cutting noise by 6.5 dB(A), turning cooling from a tolerated burden into a measurable competitive asset.
- Secured orders running through 2027 confirm this is not a prototype moment — operators are buying, installing, and betting their infrastructure roadmaps on this approach.
As artificial intelligence reshapes the demands placed on global infrastructure, the question of how to cool the machines that power it has become one of the defining engineering challenges of the era. LONGWELL, a cooling technology company, has answered with its LWBE3G FanWall system — developed in just 90 days — which reduces fan energy consumption by 38 percent and meaningfully improves the efficiency ratios that determine whether a data center is sustainable or merely functional. In a moment when power grids strain under the weight of AI's appetite, innovations like this one carry implications far beyond any single facility's electricity bill.
LONGWELL has introduced a cooling system built for the specific pressures that artificial intelligence places on data center infrastructure. The LWBE3G FanWall reduces the energy consumed by cooling fans by 38 percent — a meaningful cut in one of the largest sources of waste in modern computing facilities. Developed in just 90 days alongside a major equipment manufacturer, the system reflects the compressed timelines the AI industry now demands.
The efficiency gains are measurable in the metric that matters most to operators: Power Usage Effectiveness, or PUE. LONGWELL's technology improved this ratio from 1.42 to 1.28, meaning facilities now lose less energy to overhead like cooling and power distribution. At the scale of large AI data centers, that improvement translates directly into millions of dollars in annual savings. The system also reduces operational noise by 6.5 decibels — a quieter facility being an unexpected but welcome byproduct of better engineering.
The urgency behind the technology is real. Modern AI processors can exceed 60 kilowatts per rack, and that figure continues to climb as models grow more complex. Older cooling architectures were not designed for this density. LONGWELL built FanWall specifically for this environment, not as a retrofit but as a purpose-designed response to a new reality.
Market reception has been swift and concrete. Orders already extend through 2027, signaling that data center operators view efficient cooling not as a secondary concern but as a strategic priority. The broader implication is significant: as power grids in some regions approach their limits, cooling systems that consume less electricity may determine where and how quickly the next generation of AI infrastructure can be built.
A company called LONGWELL has released a new cooling system designed specifically for the data centers that power artificial intelligence. The system, called LWBE3G FanWall, cuts the energy used by cooling fans by 38 percent—a significant reduction in one of the largest power drains in modern computing facilities.
The technology emerged from a compressed development cycle. LONGWELL built it in just 90 days working alongside a major equipment manufacturer. That speed matters because the AI industry moves fast, and data center operators need solutions that can keep pace with rapidly evolving hardware demands. The result is a system that doesn't just use less electricity; it also makes the data center quieter, reducing noise by 6.5 decibels.
The real measure of efficiency in data centers is something called PUE, or Power Usage Effectiveness. It's a ratio that compares total facility power to the power actually used by computing equipment. A lower number is better. LONGWELL's FanWall improved this metric from 1.42 to 1.28—meaning the facility now wastes less energy on overhead like cooling and power distribution. For operators running massive facilities with thousands of servers, that difference translates directly to millions of dollars in annual operating costs.
AI workloads generate enormous amounts of heat. Modern processors can demand more than 60 kilowatts per rack, and that number keeps climbing as models grow larger and more complex. Older cooling systems struggle with this density. LONGWELL's technology is built to handle it, which is why it matters now rather than five years ago. The company designed FanWall specifically for this new reality, not as an afterthought retrofit.
Market response has been swift. LONGWELL has already secured orders that extend through 2027, suggesting that data center operators—the companies building the infrastructure for AI services—see real value in the approach. This isn't theoretical efficiency; it's a technology people are buying and installing. The orders signal a broader shift in how the industry thinks about cooling: no longer as a necessary evil to be tolerated, but as a competitive advantage to be engineered carefully.
The timing reflects a genuine constraint. As AI adoption accelerates, energy consumption at data centers has become a bottleneck. Power grids in some regions are already strained. Cooling systems that use less electricity don't just save money; they make it possible to build larger facilities in places where power availability is limited. For a company like LONGWELL, that's a market opening. For the industry, it's a sign that efficiency innovations will likely drive the next wave of data center development.
Citations marquantes
The technology was developed in just 90 days working alongside a major equipment manufacturer.— LONGWELL development timeline
Envie de l'histoire complète ?
Lire l'original sur pluang.com ↗
Technology