MagPro AeroLev 1850 Targets AI Cooling Without Evaporative Water

MagPro has introduced MagPro AeroLev 1850, a 1.85-megawatt cooling platform that targets AI data centers without relying on evaporative cooling water. Its magnetic bearings, which suspend a compressor shaft using magnetic forces, support oil-free operation. The proposition addresses a specific tension: expanding computing capacity while limiting dependence on local water supplies.

The important distinction sits inside the water-free headline. Eliminating evaporation does not mean eliminating liquid from the cooling system. It means changing how a facility releases heat outdoors. That puts MagPro’s approach against cooling towers, which reject heat partly by evaporating water.

Microsoft has already described a similar architectural direction for its future facilities, while Carrier and LG are marketing competing air-cooled equipment. MagPro enters an active market where buyers need evidence about summer performance, electricity demand, service coverage, and deployment readiness.

MagPro AeroLev 1850 Moves the Water Question Outside the Server Room

The announcement concerns heat rejection, rather than a departure from liquid cooling inside AI facilities.

The timing requires clarification. CIGU published the product announcement on September 29, 2026. The November 2 Data Centre Solutions interview brings renewed attention to the platform rather than establishing a separate launch date. That distinction matters when assessing how much deployment evidence should reasonably exist.

According to the AeroLev 1850 announcement, MagPro combines CIGU’s magnetic levitation technology with EPG’s engineering capabilities. The company describes an air-cooled platform, meaning equipment that transfers heat to outdoor air. It says the design removes cooling towers and their evaporative water requirement.

The company also advertises water supply temperatures up to 45°C and compatibility with emerging liquid cooling arrangements. Direct-to-chip cooling circulates coolant through hardware attached to processors. Immersion cooling places suitable electronics in an insulating liquid. Supporting either architecture still requires checking the complete thermal design.

These statements establish the intended application. They do not establish compatibility with every server, coolant circuit, or existing building. A prospective buyer should treat the advertised temperature range as an input to engineering review, rather than blanket approval for future hardware.

The practical issue is where heat travels after leaving a processor. A liquid circuit can carry that heat away from equipment efficiently. Another part of the facility must then release it into the surrounding environment. The water consumption question depends heavily on that final step.

This creates a common communication problem. A facility can circulate water through a closed circuit while still consuming water elsewhere in its cooling plant. Conversely, a facility can use liquid at the chips without continually evaporating water outside. The presence of pipes alone settles neither question.

For MagPro, the commercial opening follows from that distinction. A developer evaluating a constrained water supply needs an alternative to continuous evaporation. A chiller offers one possible route, but its suitability depends on electrical capacity and the temperatures the project must tolerate.

The announcement therefore changes the available procurement conversation more clearly than it changes proven operating economics. Buyers have another named platform to evaluate. Whether it improves a particular project requires evidence that connects the equipment specification to the facility’s actual operating conditions.

That is a meaningful event without declaring a universal winner. Water availability differs between locations. So do weather, electricity constraints, equipment layouts, and service requirements. The relevant question is whether removing evaporative demand solves a specific development problem without creating an unacceptable operating burden.

Water Constraints Put Developers and Cooling Suppliers Under Pressure

The immediate pressure falls on project teams that must reconcile water commitments with electrical and operational requirements.

Consider a proposed AI facility in a hot region where the available water allocation cannot support the planned cooling strategy. This is an illustrative procurement scenario, not a reported MagPro installation. Its purpose is to show where the announcement becomes relevant.

The developer faces a decision before any equipment order. It can revisit the site, reconsider the cooling architecture, or change the project’s capacity assumptions. Equipment that avoids evaporative demand deserves evaluation because it addresses the water constraint directly.

However, removing that constraint does not automatically make the project buildable. The electrical design still needs to support cooling equipment during demanding conditions. The mechanical layout needs room for equipment and maintenance access. Local approval requirements remain separate from a vendor’s product claims.

This shifts pressure toward cooling suppliers. A statement about water savings now needs an accompanying explanation of power demand and operating boundaries. Buyers need to understand the conditions under which a machine delivers its advertised capacity, rather than comparing headline specifications in isolation.

Microsoft’s published zero-water cooling design provides a useful buyer-side perspective. The company describes recirculating water between servers and chillers while avoiding evaporation. It also acknowledges that replacing evaporative cooling with mechanical cooling increases its power usage effectiveness.

Power usage effectiveness compares total facility energy with energy used by computing equipment. Microsoft says warmer chip-level cooling helps mitigate the additional energy requirement. Its explanation supports the broader direction while making the resource tradeoff explicit.

That disclosure should shape how buyers read MagPro’s announcement. Avoiding evaporative water use is a concrete design objective. It should not become shorthand for lower environmental impact under every operating condition. A credible assessment needs to examine water and electricity together.

The pressure also reaches organizations renting computing capacity. An enterprise comparing infrastructure providers should ask what a water-related commitment actually covers. Does it describe the cooling plant, the entire facility, or only selected new buildings? Those boundaries change the meaning of the claim.

For developers and AI product teams, the implication is indirect but practical. Cooling choices belong in discussions about infrastructure availability, project readiness, and supplier commitments. They should not be translated into unsupported promises about application speed or immediate reductions in computing costs.

Community-facing explanations also need precision. A promise to avoid cooling evaporation should identify remaining water uses and the accounting boundary. Clear language helps distinguish a measurable engineering choice from a broad claim that a data center requires no water whatsoever.

The resulting procurement pressure is constructive. Suppliers must explain what their equipment removes, what it still requires, and which conditions limit performance. Developers must carry those distinctions into project documents instead of allowing a product headline to become a facility-wide guarantee.

Avoiding Evaporation Changes the Energy Calculation

The central tradeoff concerns how a project balances water demand against the electricity required to move heat.

Cooling towers and air-cooled equipment use different routes to reject heat. Comparing them requires more than asking which machine has the more advanced compressor. The surrounding conditions and required cooling temperatures determine how useful each approach becomes.

Evaporation helps remove heat, but it consumes water. Avoiding that mechanism changes the equipment and operating choices available to a designer. The resulting energy requirement deserves explicit modeling rather than an assumption that water savings and electricity savings always move together.

An appropriate comparison starts with a common job. Each proposed system should serve the same computing load, deliver the required coolant conditions, and satisfy the same reliability expectations. Otherwise, the comparison risks rewarding a design for performing a different task.

The accounting should also cover the relevant supporting equipment. A compressor specification does not describe every electrical load in a cooling installation. Pumps, fans, and controls belong in the assessment when they form part of the system needed to deliver cooling.

Annual averages alone leave another question unanswered: what happens during the conditions that determine the electrical connection and backup strategy? A project can have an attractive yearly estimate while still facing a demanding peak operating requirement.

This is particularly relevant to a product positioned for hot locations. Buyers should request capacity and electrical demand across the expected outdoor temperature range. They should also ask which limits require reduced output, different operating settings, or additional equipment.

The warmer-water proposition deserves similar scrutiny. A higher acceptable coolant temperature can expand engineering options, but the facility cannot choose that temperature independently. Server requirements, heat transfer equipment, and operating safeguards must support the same design assumptions.

That turns compatibility into a chain of approvals. A chiller’s advertised supply temperature does not establish the temperature available at every processor. Intervening equipment and operating margins matter. Procurement teams should insist that each participant accepts the complete arrangement.

Free cooling adds another variable. The term describes using favorable outdoor conditions to reduce mechanical cooling work; it does not mean cooling without electricity. The opportunity depends on the relationship between outdoor conditions and the temperatures the equipment must maintain.

For a buyer, the useful evidence is a site-specific operating model. It should show when each cooling mode operates, the load it supports, and the associated electricity requirement. A general statement about free cooling cannot establish those annual results.

MagPro AeroLev 1850 should therefore be evaluated as part of a facility design, rather than as an isolated water-saving component. Its proposed role is clear. The unresolved calculation is whether its operating profile matches the constraints of the project considering it.

This framing also prevents a misleading environmental shortcut. Reduced onsite evaporation is one outcome with a defined boundary. Overall resource performance requires additional information. Keeping those questions separate allows buyers to recognize the water benefit without inventing a broader result.

Carrier and LG Make Oil-Free Air Cooling a Competitive Category

MagPro faces competitors offering similar architectural ingredients, so differentiation requires more than magnetic bearings and an air-cooled label.

Carrier introduced its air-cooled centrifugal chiller, AquaEdge 30CF, on February 26, 2026. A centrifugal compressor uses a rotating impeller to increase refrigerant pressure. Carrier says its design uses magnetic bearings and an oil-free compressor architecture.

Carrier also emphasizes recovery following a power interruption. It advertises restoration of full cooling capacity in under three minutes and capacity exceeding 3 megawatts, depending on ambient conditions. Those remain manufacturer claims, but they identify concrete dimensions for competitive evaluation.

LG announced an AI cooling chiller on September 21, 2026. Its announcement likewise highlights an oil-free magnetic bearing compressor and refrigerant free cooling. LG also stresses recovery after power restoration and specifies conditions behind its capacity figures.

These announcements do not prove that the products perform identically. They show that the category already contains overlapping design choices and competing reliability claims. MagPro cannot establish a unique market position simply by combining terms that other suppliers also use.

For buyers, that is useful competitive context. The shortlist can include alternative implementations of broadly similar ideas. The evaluation then moves toward operating performance, integration, support, and the evidence each supplier makes available.

A fair comparison needs consistent conditions. Nameplate capacity, meaning a product’s stated rated output, should not become a ranking without its associated temperatures and assumptions. An equipment rating only helps when the buyer understands the conditions attached to it.

Recovery claims require equivalent care. Returning to operation after power restoration is different from maintaining uninterrupted cooling throughout an outage. A project team should examine the entire sequence, including electrical transfer, control restart, and the behavior of connected equipment.

Service arrangements deserve attention alongside these technical comparisons. A purchaser should establish who performs commissioning, who diagnoses faults, and how replacement parts reach the site. These are due diligence questions, not evidence that any named supplier lacks those capabilities.

The same applies to expansion. A modular product can simplify some planning decisions, but expansion still needs electrical capacity, piping connections, controls integration, and space. Buyers should test the installation plan rather than assuming that adding another machine completes the task.

MagPro’s strongest available positioning is consequently specific: another air-cooled option aimed at projects concerned about evaporative water demand. Its competitive strength will depend on whether its evidence and delivery arrangements meet the requirements of those projects.

The broader market signal is that water-conscious cooling has become a product competition rather than an abstract sustainability discussion. Multiple suppliers now ask buyers to evaluate related architectures. That increases the importance of comparable documentation and makes unsupported uniqueness claims less persuasive.

Efficiency Claims Need Comparable Tests and Clear Boundaries

The public materials establish the product’s intended capabilities, but they do not independently establish its operating advantage.

The published product release carries a press-release label and appears in a sponsored section. It repeats company performance claims. That makes it useful for documenting what MagPro announced, while limiting its value as independent validation.

MagPro’s release claims up to 40% higher overall energy efficiency, up to 10% lower maintenance costs, and up to 8% lower failure rates. Readers should retain the qualifiers. These are company comparisons, not independently verified results for a named operating facility.

The missing context matters more than the size of the percentages. A comparison needs an identified baseline, matching test conditions, and a defined measurement boundary. Without those details, a buyer cannot reliably transfer a claimed improvement into its own project model.

Efficiency and energy consumption also require careful language. A percentage improvement in an efficiency measure is not automatically the same percentage reduction in electricity. The mathematical relationship depends on the metric and on whether the comparison holds delivered cooling constant.

Maintenance claims introduce a different evidence requirement. Removing oil-related tasks can change a maintenance schedule, but total maintenance depends on the complete machine and operating arrangement. Buyers should request the actual service schedule and identify the responsibilities retained by the operator.

Failure-rate comparisons need still more context. Relevant questions include the number of machines observed, operating duration, failure definition, and equipment used as the baseline. Without that information, a precise percentage can suggest more certainty than the disclosed evidence supports.

Factory testing and operational evidence serve different purposes. An acceptance test can check whether equipment meets specified conditions before delivery. Field records can reveal behavior over changing weather and workloads. Neither should be described as a substitute for the other.

The reviewed materials do not establish an independently audited MagPro deployment with published seasonal performance. That is a limit on the available evidence, rather than proof that the product fails. The responsible conclusion is that buyers still need project-level validation.

The water claim also needs a documented boundary. A vendor should distinguish avoided evaporation from initial filling, servicing requirements, and unrelated building consumption. A facility operator should then explain how its own water reporting accounts for those categories.

Retrofit claims require similar discipline. An existing facility might have equipment, temperatures, and controls that constrain a replacement project. Buyers should request an integration assessment before interpreting future compatibility as a promise of minimal infrastructure work.

This is where technical teams benefit from preserving the underlying documents. Keeping test reports, assumptions, and supplier responses in a searchable knowledge base can help reviewers identify when comparisons use different boundaries.

The goal is a traceable decision. Every claimed advantage should connect to a defined condition and a relevant project requirement. That approach gives promising equipment a fair assessment while preventing promotional percentages from becoming unsupported engineering commitments.

Watch Deployment Evidence, Operating Curves, and Delivery Readiness

The next useful developments are measurable disclosures that connect the announcement to an operating project.

The first signal is a named customer deployment with a clearly described cooling arrangement. Strong evidence would identify the application, commissioning status, and scope of the water claim. An announcement of commercial interest would provide less assurance than an installation that has entered service.

A customer reference would strengthen the adoption case without settling every performance question. Buyers would still need to know whether the site resembles their proposed project. Climate, operating temperature, workload, and reliability requirements affect how transferable a reference becomes.

The second signal is a detailed operating envelope, meaning the conditions within which the supplier specifies acceptable performance. Useful disclosures would connect cooling output and electrical demand to outdoor temperature and water conditions. Independent testing would add confidence if its scope were clearly stated.

Those curves would strengthen the technical case when they support the intended deployment conditions. They would narrow the addressable application if they revealed limitations incompatible with a buyer’s site. Either result would improve the market’s understanding of where the platform belongs.

The third signal is delivery and service readiness for the market being targeted. Buyers should look for concrete commissioning arrangements, support responsibilities, spare-parts provisions, and applicable product documentation. These disclosures would help distinguish a marketable design from an equipment package ready for a particular project.

Competitor announcements will remain relevant, but the most useful comparison concerns equivalent evidence. More capacity on a product page does not resolve a service question. A faster restart claim does not replace an annual electricity assessment. Each disclosure should answer its own procurement requirement.

Over the coming months, these signals offer a better framework than counting repeated launch coverage. The underlying announcement dates to September. Additional articles matter most when they add customer evidence, operating detail, or commercial information that was previously unavailable.

For AI infrastructure buyers, the immediate action is to define the project’s constraints before requesting comparisons. Establish the required cooling temperatures, acceptable water consumption, electrical limits, and recovery expectations. Then ask suppliers to demonstrate how their proposals satisfy that same set of requirements.

Before treating MagPro AeroLev 1850 as the answer for a planned facility, request a site-specific assessment with documented assumptions. Ask which water uses disappear, how electrical demand changes, and what happens during the project’s hottest operating conditions. Compare those answers with alternative equipment using the same load and reliability requirements. Preserve the supporting reports so facilities, procurement, and computing teams can review the same evidence. The announcement gives buyers a concrete option to investigate; the purchasing decision should follow from the operating case. Which supplier can document the cooling performance, resource requirements, and support arrangements your project actually needs?