From Air-Only to Liquid-Ready: A Practical Roadmap for Your Data Center
For most operators, the liquid cooling question isn’t “if” anymore; it’s “how fast can we get there without breaking anything?”
AI and GPU racks are driving densities well beyond what traditional air systems were built to handle. Recent market studies expect the global data center liquid cooling segment to grow at more than 20%+ CAGR this decade as operators chase efficiency and AI capacity. Grand View Research
At the same time, cooling systems still consume roughly a third of total facility power on average, so every decision you make about cooling shows up in your power bill, ESG report, and risk register. ScienceDirect
This post lays out a practical, no-hype roadmap for moving an existing, air-cooled environment toward hybrid or fully liquid-cooled operation—with clear roles for your internal team, OEMs, and a field-execution partner like Guardian.
Step 1: Define the “Why Now?” In Business Terms
Liquid cooling conversations often start with chip roadmaps and rack kW numbers. That’s important, but it’s not enough to get budget or executive backing.
Anchor the “why” to three business drivers:
- Capacity and time-to-market
JLL’s latest global data center outlook notes that AI demand is now a primary driver of new build and retrofit projects and that GPU advancements require a shift to liquid cooling to sustain growth. jll.com
Put simply: without a liquid plan, AI projects will stall when you hit power and cooling ceilings. - Energy and sustainability pressure
Research from Vertiv and others shows that introducing liquid cooling into high-density environments can reduce total data center power by ~10% and significantly improve cooling efficiency metrics. Vertiv
With projections that data centers could account for more than 14% of U.S. power demand by 2030, every percentage point matters. The Guardian - Risk control at higher densities
As new AI servers push tens or even hundreds of kilowatts per rack, relying on air alone amplifies hot spots, throttling, and failure risk. ASHRAE’s recent liquid cooling guidance is clear: to stay within safe thermal envelopes at these power levels, liquid is no longer “experimental”; it’s a mainstream requirement. Datacenter Dynamics
Document these drivers on one page your leadership can sign off on. That page becomes the north star for every cooling decision that follows.
Step 2: Map Your Starting Point “Cooling Reality Check”
Before you talk about immersion tanks or direct-to-chip loops, you need an honest picture of where you are today.
Key questions to answer:
- What are your current and forecast rack densities by room and by row?
- Where are the worst hot spots today, and what quick fixes are masking deeper issues?
- What is your true power headroom at each site once you account for safety margins, utility constraints, and growth?
- Which rooms are most strategic for AI and GPU deployments over the next 24–36 months?
Combine this with a walk-through of the facility by a team that understands both cooling and project execution. Guardian’s data center and enterprise services team already manages complex installations, migrations, and decommissions of “any scope, complexity, size, or location”; the same project discipline applies to liquid cooling retrofits. guardiandatadestruction.com
Out of that assessment, designate:
- 1–2 pilot rooms or rows
- 2–3 “liquid later” rooms where you’ll stay air-only but enforce better best practices
- Any “do not invest further” spaces that will be sunset as part of your longer-term site strategy
Step 3: Choose the Right Liquid Cooling Patterns, Not Just Products
There’s no single “right” liquid technology. Instead, think in patterns:
- Direct-to-Chip (D2C) Cold Plates
- Best for: High-density GPU/CPU servers where OEMs support cold plates
- Pros: Highest heat removal at the chip, often integrated with existing form factors
- Considerations: Facility water quality, materials compatibility, and manifolds/wet connections per ASHRAE guidance. ashrae.org
- Rear-Door Heat Exchangers (RDHx)
- Best for: Mixed-density rooms where you need a fast, incremental upgrade path
- Pros: Easier retrofit in existing rows, can significantly cut load on room-level CRAC/CRAH
- Considerations: Floor loading, door access, consistent water supply temperatures
- Immersion Cooling (Single- or Two-Phase)
- Best for: Extremely high densities, edge sites with space constraints, or greenfield designs
- Pros: Excellent heat transfer, potential for significant fan and chiller energy savings Dash Harvard
- Considerations: Fluid management, OEM warranties, and operational culture change
Your goal in an existing facility should usually be a portfolio of patterns, not a one-size-fits-all bet. For example:
- RDHx in mixed-use aisles
- D2C for the main AI training cluster
- Continued air-only for low-density or near-retirement workloads
This is where Guardian can collaborate with your OEMs, ITAD partners, VARs, and MSPs to design a plan that fits your hardware mix and refresh cycles rather than forcing a forklift upgrade.
Step 4: Build a Liquid Cooling Runbook Before the First Rack Arrives
Too many projects focus on hardware, then scramble to figure out who does what once equipment lands in the dock.
An effective liquid cooling runbook should cover:
- Roles and responsibilities
- Who owns the facility loop, CDUs, and leak detection?
- Who touches manifolds, quick disconnects, and fluid changes?
- Where does your field-execution partner (Guardian) step in onsite?
- Standard operating procedures (SOPs)
- Bringing new liquid-cooled racks into production
- Performing maintenance that requires wet connections to be opened
- Switching between redundant loops or cooling paths
- Incident response
- Detection, isolation, and cleanup for leaks or line damage
- Data protection and chain-of-custody if systems are impacted
- Escalation paths that include your partners across ITAD, logistics and OEM support
Guardian already operates with strict chain-of-custody, auditing, verification, and reporting standards for data center projects nationwide.
Extending those disciplines into liquid-cooled environments helps ensure that every hose, manifold, and CDU is treated with the same rigor as a high-value server or storage array.
Step 5: Plan for Lifecycle and Decommissioning from Day One
Liquid cooling is not just an installation project; it’s a full lifecycle commitment.
From a sustainability and compliance standpoint, you need clear answers to:
- How will fluids be stored, monitored, and eventually removed or recycled?
- How will you track equipment and materials (racks, tanks, piping, CDUs) as they move between sites, partners, and recycling outlets?
- What documentation will you need for ESG reports, audits, and customer commitments?
Analysts expect liquid-based approaches, including direct-to-chip systems, to continue growing rapidly alongside AI workloads over the next decade. MarketsandMarkets
That means whatever you deploy today is likely the first of many waves, not a one-off experiment.
Guardian is already a single-source provider for:
- On-site data destruction
- Enterprise and data center relocations and decommissioning
- Secure IT packing and logistics, including multi-site projects across North America guardiandatadestruction.com
By integrating liquid cooling lifecycle steps into these existing programs, you can:
- Avoid ad-hoc, non-compliant fluid handling
- Maintain clean audit trails for hardware and materials
- Reuse proven logistics and decommissioning playbooks across every site
Step 6: Start Small, Standardize Fast
A good liquid cooling strategy doesn’t try to boil the ocean.
Instead:
- Launch a tightly scoped pilot
- One row, or a defined group of AI racks
- Clear success metrics: uptime, energy use, thermal performance, and operational friction
- Capture lessons learned in your runbook
- What took longer than expected?
- Where did roles and responsibilities get blurry?
- Which tasks really needed specialist on-site support?
- Standardize and scale
- Turn the pilot design into a repeatable “liquid pod” template
- Use Guardian’s national footprint and project management to replicate that pod across colos, enterprise sites, and edge locations. guardiandatadestruction.com
Done right, liquid cooling becomes just another standard project type in your portfolio, managed with the same discipline as migrations, decomms, and data destruction.
Where Guardian Fits in Your Roadmap
Guardian is positioned to support your roadmap at every step:
- Assessment & planning: Ground-level perspective on what’s realistically deployable in your existing rooms.
- Implementation & logistics: Coordinating equipment, fluids, and field teams “from A to Z,” including staging, packing, transport, and onsite installation. guardiandatadestruction.com
- Run-state support: Onsite services aligned with your SOPs and OEM requirements, making sure the right people touch the right components at the right time.
- Decommissioning & ESG: Secure, documented fluid handling, asset disposition, and reporting that support your ESG commitments as well as your customers’ expectations.
Liquid cooling management isn’t just about cooling more watts. It’s about bringing order, safety, and repeatability to a new operational reality.
Guardian can help you get there, from your first liquid-cooled rack to a multi-site, AI-ready portfolio.
When you think about introducing liquid cooling, what’s your biggest concern: facility infrastructure, operational risk, or lifecycle/ESG impact?
