iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
DCD’s October 15, 2026, DCD Studio interview with Matthew Baynes of Schneider Electric examines how data-center developers are responding to rapid AI-driven GPU growth. The discussion centers on modular construction, liquid cooling, future-proof facility design, and grid innovations intended to make new capacity faster to deploy, capable of supporting denser racks, and more resilient as power constraints tighten.
Why AI GPUs are changing the development brief
Traditional data-center planning could treat compute growth as a relatively predictable increase in servers and floor space. AI infrastructure makes that assumption less reliable. GPU clusters can require much greater power and cooling capacity per rack than general-purpose workloads, while demand for new capacity can arrive faster than a conventional build cycle.
That shifts development priorities from simply adding rooms and electrical capacity to designing facilities that can adapt as workloads, rack densities, and power requirements change. The interview presents this adaptation as a coordinated engineering and infrastructure problem rather than a single equipment upgrade.
Modular construction as a response to uncertain demand
Faster, repeatable deployment
Modular construction is highlighted as a way to deploy data-center capacity in repeatable blocks. Instead of completing every part of a large facility as one project, developers can use standardized modules for portions of the power, cooling, or IT environment and add capacity in stages.
This approach can help organizations match construction more closely to demand. It also creates a framework for repeating designs across expansion phases, although the interview does not provide a schedule, cost comparison, or measured performance result for modular versus conventional construction.
Managing changing rack requirements
Modularity is particularly relevant when the final mix of AI systems is still evolving. A design that reserves defined expansion points can make it easier to add higher-density halls, additional electrical capacity, or new cooling sections without redesigning the entire site. The practical value depends on how early those interfaces are defined and whether the surrounding utility and building infrastructure can support later phases.
Why liquid cooling matters for dense AI racks
The interview links liquid cooling to the higher rack densities associated with AI systems. Air cooling remains part of many data-center designs, but air’s ability to remove heat is constrained as more power is concentrated in a smaller footprint. Liquid can transfer heat more directly from high-power components, making it an important option for dense GPU deployments.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Rank #2
- Part number 900-53651-2500-000 and model: P3651
- This is the 2 slot version for when there is no empty slots between 2 slot cards. If you have one or more empty slots between the cards or the cards are 3 slot this NVLink will not work. See the attached images showing the card layout.
- NVLink 3.0 for any brand of RTX Ampere model graphics cards: 3090, A30, A40, A100 / H100 (Requires three NVLinks), A800, A4500, A5000, A5500, A6000
- This is the same as PNY part number: NVLAMP-2SLOT-BSP and RTXA6000NVLINK-KIT
- This is the same as Dell part number: 0RWJ7Y
Cooling is a facility decision, not just a rack decision
Introducing liquid cooling affects the wider site. Developers must account for heat-transfer equipment, distribution piping, controls, maintenance access, leak detection, and the relationship between liquid systems and any remaining air-cooled equipment. Those choices need to be made during facility planning because retrofitting a live environment can be more disruptive than incorporating the required distribution paths and plant capacity from the start.
The source does not identify a particular liquid-cooling architecture, operating temperature, rack power rating, or efficiency figure. Its established point is strategic: rising AI rack density is making liquid cooling an increasingly important part of development planning.
Future-proof design under moving workloads
Future-proofing in this context does not mean predicting one fixed hardware roadmap. It means preserving options as GPU generations, server configurations, rack densities, and power envelopes change. Useful design decisions can include reserving physical space, sizing electrical pathways for expansion, providing adaptable cooling distribution, and separating construction phases so later capacity can be added without invalidating earlier work.
Rank #3
- Video/Sound Cards
- Passive Cooling
There is a trade-off. Designing for every possible future requirement can increase initial complexity and capital commitment. A more practical objective is to identify the changes most likely to affect the site—especially power, heat rejection, and rack layout—and build flexibility around those constraints while keeping the initial deployment proportionate to confirmed demand.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallPower availability and the grid challenge
On-site engineering is only one part of capacity
AI data centers can be technically ready yet unable to expand if the surrounding grid cannot deliver the required power. The discussion therefore treats grid innovation as part of data-center development, alongside buildings, switchgear, generators, and cooling systems.
Developers need to consider when utility capacity will be available, how expansion phases align with grid interconnection milestones, and which operating strategies could reduce stress during constrained periods. The interview does not name a specific utility, market, generation technology, or interconnection timetable, so those details remain site-specific rather than universal recommendations.
Rank #4
- CUDA Cores: 4608 / NVIDIA Tensor Cores: 576 / NVIDIA RT Cores: 72
- GPU Memory: 24 GB GDDR6 with ECC / Bandwidth: 624 GB/Sec
- System Interface: PCI Express 3.0 x16
- Four DisplayPort 1.4 Connectors
- 3D Stereo Support with Stereo Connector
Sustainability and resilience together
Sustainable grid innovation is presented as connected to resilience. A facility that can use power more efficiently, coordinate demand, and integrate lower-carbon energy sources may be better positioned for both environmental goals and future capacity limits. Sustainability cannot be evaluated only by the building’s equipment selection; it also depends on the electricity supply, operating profile, and expansion plan.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What developers should take from the interview
- Plan for density: AI growth can change the amount of power and heat concentrated in each rack.
- Build in phases: Modular construction can support staged deployment and repeatable expansion.
- Design cooling early: Liquid-cooling requirements can affect room layouts, distribution, controls, and maintenance.
- Protect future options: Reserve pathways, space, and capacity for changing hardware and workloads.
- Engage the grid early: Utility availability and interconnection timing can determine whether a technically complete project can actually scale.
- Evaluate sustainability at system level: Facility efficiency, grid interaction, and energy sources all shape the outcome.
What the DCD Studio episode establishes—and what it does not
DCD describes Seb Moss’s conversation with Matthew Baynes of Schneider Electric as covering adaptation to AI-driven GPU growth through modular construction, liquid cooling, future-proof designs, and sustainable grid innovations. It is an editorial discussion of development priorities, not a published benchmark or product comparison.
Free tools Windows power users keep installed
One-click scans. No signup required.
The episode summary provides no named project, geography, price, implementation schedule, verbatim Baynes quotation, or measured performance statistic. Those omissions matter when applying the ideas to a specific site: developers still need detailed load studies, utility assessments, cooling design, permitting analysis, and financial modeling before selecting an architecture.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
