By Mike Hodge, AI Solutions Lead, Keysight Technologies

It’s the heart of the AI gold rush, and everyone wants to capitalize on the next big thing. Large language models, multimodal systems, and domain-specific AI workloads are moving from experimentation to production at scale. Across industries, enterprises are building their own proprietary models or integrating pre-trained ones to power applications spanning from video analytics to highly specialized inference services.

This shift has triggered a new wave of infrastructure investment. But while GPUs and accelerators dominate the conversation, scaling AI platforms has produced a less obvious constraint: front-end network performance. In increasingly distributed, multi-tenant AI environments, the ability to move data efficiently into (and across) platforms has become just as critical as raw compute density.

New AI platforms mean new expectations for infrastructure

AI infrastructure is no longer the exclusive domain of a handful of hyperscalers. A growing class of service providers has begun offering end-to-end AI platforms where compute, storage, networking, and orchestration are delivered as a service. Their value proposition is straightforward: customers bring data and models, while the platform handles the complexity of building, operating, and maintaining large-scale data center deployments.

Service models like these, however, place extraordinary demands on networking. Unlike traditional cloud workloads, AI jobs are defined by massive, sustained data movement and tight coupling between data pipelines and compute utilization. GPUs cannot perform at peak efficiency unless data arrives on time, in the right order, and at predictable speeds.

As a result, network performance is now one of the primary determinants of training, inference, and infrastructure efficiency.

The eye of the storm is moving from the fabric to the front end

AI infrastructure discussions often focus on back-end fabrics. Think about things like high-bandwidth, low-latency interconnects between GPUs, for example. However, while these fabrics are indeed essential, they are only part of the picture.

Before training or inference ever begins, data must first traverse the front-end network. This occurs in several ways, but some of the most common paths include:

• From remote object stores or on-premises repositories into the data center
• From ingress points into virtual machines or containers
• From storage into GPU-attached hosts

This is where north-south traffic (external to internal) intersects with east-west traffic (host-to-host and service-to-service). And in AI environments, these flows are not occasional spikes. They are sustained, high-throughput, latency-sensitive streams that run continuously throughout the lifecycle of a job.

When front-end networks underperform, the consequences are costly and immediate: idle accelerators, elongated training windows, unpredictable inference latency, and poor multi-tenant isolation.

Why traditional network validation falls short

Most cloud networks were designed around general-purpose workloads. Think about things like web services, databases, and transactional systems with relatively modest bandwidth demands and fluctuating traffic patterns punctuated by the occasional spike.

AI workloads, on the other hand, break these assumptions. On the front end, AI traffic is characterized by:

• Extremely large data transfers, often using jumbo frames
• Long-lived connections, sustained over hours or days
• Millions of concurrent sessions in multi-tenant environments
• Tight latency and jitter tolerances to avoid starving accelerators

Conventional network testing approaches — such as synthetic benchmarks, isolated link tests, or small-scale simulations — are unable to replicate this behavior. As a result, many issues only surface once customer workloads are already running, which also happens to be when the cost of remediation is highest.

The need for realistic workload emulation

Optimizing front-end AI networks requires the ability to reproduce real workload behavior at scale. That means emulating both north-south and east-west traffic patterns simultaneously, across distributed environments and under sustained load.

For north-south paths, this includes verifying that large datasets can be reliably pulled from diverse external sources into local storage. Moreover, the network must also be able to do so with consistent throughput, predictable latency, and no silent data loss. Transfers like these are essential, as any inefficiency propagates directly into longer training times and underutilized GPUs.

For east-west paths, the challenge shifts to connection density, latency, and scalability. Once workloads are running, virtual machines and services exchange data continuously. Sometimes within the same host, sometimes across racks, and sometimes across geographically separated data centers. Modern AI platforms increasingly rely on SmartNICs and offload technologies to make this feasible, so these components must also be validated under realistic connection rates and protocol behavior.

Without large-scale, workload-accurate testing, subtle bottlenecks — such as rule-processing limits, connection-tracking inefficiencies, or unexpected latency spikes — can remain hidden until production traffic exposes them.

Front-end optimization is a competitive differentiator

In response, the most advanced AI platform operators are shifting left: validating their front-end networks before customers ever deploy workloads. Along the way, their proactive approach is changing the economics of AI infrastructure.

Stress-testing networks under real-world conditions offers a range of benefits for network operators:

• Identifying performance cliffs at high line rates
• Understanding how different layers of the stack interact under load
• Resolving scaling limitations in NICs, virtual networking, or storage paths
• Delivering predictable performance across tenants and geographies

It’s not just about improving peak throughput. It’s about building confidence that platforms perform as expected under peak pressure. And in a market where AI workloads are expensive, time-sensitive, and strategically important, this confidence becomes a differentiator. Customers may never see the network directly, but they feel its impact in faster training cycles, lower inference latency, and fewer production surprises.

Looking ahead: front-end networks and the next generation of AI

AI workloads continue to evolve. Microservices-based architectures, distributed inference pipelines, and increasingly stateful services are placing even more emphasis on low-latency, high-availability front-end connectivity. At the same time, data is becoming more geographically distributed, pushing platforms to span multiple regions and network domains.

In this environment, front-end networks are no longer a supporting actor. They are a core component of AI system design. That means they must be engineered, validated, and optimized with the same rigor applied to compute and accelerators.

The lesson is clear: operators cannot optimize AI infrastructure by focusing on GPUs alone. The performance, efficiency, and reliability of tomorrow’s AI platforms will be defined just as much by how well they move data as by how fast they process it.

The post AI’s Overlooked Bottleneck: Why Front-End Networks Are Crucial to AI Data Center Performance appeared first on Data Center POST.

AI workloads are pushing data centre infrastructure towards higher rack densities, new cooling strategies and greater power demand. Jamie Darragh, Data Centre Director, Europe, at global data centre engineering design consultancy Black & White Engineering, examines the design implications for the next generation of facilities.

AI and high-performance computing are placing new demands on data centre infrastructure. Rack densities are increasing; facilities are being delivered at larger scale and operators are under pressure to support workloads that consume far greater levels of power and generate far higher heat loads than conventional cloud environments.

Independent forecasts underline the pace of expansion. Gartner estimates global data centre electricity consumption will rise from around 448TWh in 2025 to roughly 980TWh by 2030, driven largely by AI-optimised computing infrastructure. Within that growth, AI servers alone are expected to account for close to 44% of data centre power consumption by the end of the decade.

For our engineering teams, these workloads are altering the practical limits of traditional infrastructure design. Rack densities exceeding 100–200kW are now appearing in project specifications, particularly where large AI training clusters are planned. These loads influence every part of the building environment, from electrical distribution and cooling capacity to structural loading and cable management.

Designing for extreme density

Under these conditions, air cooling alone becomes difficult to sustain across entire facilities. Liquid cooling is therefore increasingly included in the baseline design of new data centres rather than introduced later as a specialist solution. This cooling method is becoming increasingly favourable due to its higher specific thermal capacity compared with air, which enables more efficient heat transfer and removal. Direct-to-chip and rack-level systems are being designed alongside air cooling so facilities can accommodate different densities and equipment types across the same site.

The introduction of liquid systems requires careful coordination between disciplines. Facilities must manage environments where air and liquid cooling operate together, supported by monitoring platforms, safety controls and operational procedures capable of supporting both approaches.

Some IT chips require different liquid cooling temperatures than those used in air-cooling systems, creating technical hurdles for the overall heat rejection system and requiring precise control of the cooling circuit temperature. Another engineering challenge lies in integrating these systems with power distribution, control platforms and maintenance strategies rather than selecting one cooling method over another.

Higher density also narrows operational tolerance. Commissioning becomes more demanding and redundancy strategies require more detailed modelling. Infrastructure must be capable of supporting peak compute demand while maintaining efficiency when loads are lower, placing greater emphasis on flexible electrical and mechanical systems.

The scale of development is also increasing. Buildings that once delivered a few megawatts of capacity are now part of campus-scale developments where multiple data halls contribute to facilities delivering hundreds of megawatts. data centres are increasingly planned and delivered as long-term infrastructure assets rather than individual projects.

This environment encourages repeatable design and industrialised delivery methods. Developers and investors expect predictable construction schedules and consistent performance across multiple sites. As a result, engineering teams are placing greater emphasis on modular infrastructure systems and digital design methods that allow mechanical and electrical systems to be configured and deployed repeatedly.

Power, control and operational intelligence

Power availability is also becoming a determining factor in project planning. In many regions, grid connection capacity is now one of the main constraints on new development. Gartner has warned that by 2027 as many as 40% of AI data centres could face operational limits because of power availability.

Developers are therefore engaging more closely with utilities during early feasibility stages and exploring complementary infrastructure such as on-site generation and energy storage. In some cases, data centres are also being designed to contribute to wider grid stability through demand response and energy management capability.

Artificial intelligence is also beginning to influence how facilities themselves are operated. Machine-learning systems are already being used in some environments to optimise airflow patterns, cooling plant performance and power distribution using live operational data.

The next stage will see more widespread use of integrated control platforms and digital twins capable of modelling facility behaviour in real time. These systems allow operators to simulate infrastructure performance under different load conditions, test operational changes and identify maintenance requirements before faults occur.

Environmental performance remains another constraint as compute density increases. Higher workloads place additional pressure on energy supply while raising questions around water consumption, construction materials and waste heat recovery. Planning authorities and investors are increasingly looking for measurable improvements in efficiency and carbon reporting before approving new developments. Sustainability therefore sits alongside power and cooling as a central engineering consideration rather than a secondary design feature.

Taken together, these conditions create a more complex design environment for data centre infrastructure. Higher compute densities, power constraints and new operational technologies require mechanical, electrical and digital systems to be considered together from the earliest design stages.

Facilities intended to support AI workloads must accommodate far greater performance requirements than earlier generations of data centres while remaining adaptable as infrastructure technologies and operating practices continue to develop.

# # #

About the Author

Jamie Darragh is Data Centre Director, Europe at Black & White Engineering. He leads the delivery of complex, mission-critical projects across the region, with a focus on technical quality, design coordination and strong client relationships. A Chartered Engineer and member of CIBSE and the IET, Jamie has worked across Europe, the Middle East and the UK since 2005. He brings a clear, practical approach to engineering challenges, combining technical expertise with commercial awareness. He is committed to developing teams that work collaboratively and perform at a high level. Jamie has received several industry awards, recognising both his technical capability and his impact on the built environment including ‘Engineer of the Year’ at leading Middle East industry awards.

The post AI Workloads and the Implications for High-Density Data Centre Design appeared first on Data Center POST.

Datalec Precision Installations (DPI) has introduced its next-generation Data Centre Modularisation Solution, targeting operators that need to add capacity quickly without sacrificing control, reliability or lifecycle value.

Developed in response to surging demand for rapid capacity expansion, the new solution is designed to compress delivery timelines while maintaining full flexibility over configuration, performance and long-term scalability. Each system is precision engineered and manufactured by Datalec to ensure compatibility across structural, mechanical and electrical systems, helping to reduce onsite risk and integration challenges.

Datalec’s modular approach combines pre-engineered design principles with tailored manufacturing, enabling customers to adapt deployments to specific site conditions, operational requirements and growth strategies, including AI-intensive workloads. By shifting more work offsite into a controlled manufacturing environment, the solution minimises disruption associated with traditional construction-led projects and supports safer, more succinct installations and a faster speed to market.

“With organisations under pressure to scale quickly while managing capital expenditure and quality, this launch marks a pivotal shift in how data centre capacity can be delivered,” said John Lever, Director of Modular Solutions at Datalec. “Our modular solution brings these priorities together, giving customers the confidence and agility to develop at the pace their business requires.”

By emphasising reliability, engineering excellence and lifecycle value, Datalec’s new Modularisation Solution reinforces the company’s role in delivering robust, scalable infrastructure for today’s data-driven enterprises and AI-led digital transformation. More information on Datalec’s modular critical infrastructure solutions is available at www.datalecltd.com/critical-infrastructure/modular.

The post Datalec Unveils Next-Generation Modular Data Centre Solution to Accelerate Deployment appeared first on Data Center POST.

Originally posted on Compu Dynamics.

The data center industry is entering a new phase — one defined less by generic flexibility and more by purpose-built design. For years, operators relied on large, adaptable white-space shells to support a wide range of workloads. That model served the cloud era well. But the rise of AI and high-density computing is reshaping infrastructure requirements, pushing the industry toward more integrated, modular, and performance-driven environments.

In a recent QTS podcast with David McCall, Steve Altizer, CEO of Compu Dynamics, shares his perspective on how prefabrication and modular white-space design are becoming foundational to building data centers ready for the AI era.

Why the White Space Is the New Frontier for Modular Innovation

As AI workloads push power density to new extremes, long-standing assumptions about how data centers are designed and built are being challenged. White space, once treated as a static and custom-built environment, is rapidly becoming the next frontier for modular innovation.

Why Density Changes Everything

AI workloads aren’t just hotter, they’re architecturally different. When you’re deploying GPU arrays that demand 100kW per rack today and 600kW tomorrow, you’re not simply installing servers; you’re building a machine. The sheer volume of structural steel, high-pressure liquid cooling pipes, power distribution, and network infrastructure required to support these dense deployments creates an entirely new opportunity: factory assembly.

Traditional cloud data centers were too light and airy to justify prefabrication – components would literally fall apart in transit. But modern AI infrastructure is robust, dense, and highly engineered. It’s perfect for modular construction. Think of it as building a motherboard rather than a room. Every element – power, cooling, network – works in precise coordination to support the chips doing the computational work.

To continue reading, please click here.

The post Rethinking Data Center Construction In The AI Era – The QTS Experience Podcast appeared first on Data Center POST.

Industry Analysts Examine Revenue Models, Investment Rationale, and Growth Projections Shaping the Future of Cloud Infrastructure

At the infra/STRUCTURE Summit 2025, held October 15-16 at The Wynn Las Vegas, a distinguished panel of industry analysts gathered to tackle one of the most pressing questions facing the digital infrastructure sector: Is the explosive growth in AI-driven cloud infrastructure sustainable, or are we witnessing an investment bubble destined to burst?

The analyst team conclusion session brought together the Structure Research team to examine the economic fundamentals underpinning today’s unprecedented capital expenditure in AI infrastructure. Moderated by Philbert Shih, Managing Director of Structure Research, the panel featured Jabez Tan, Head of Research; Sacha Kavanagh, Research Director, EMEA; Swapna Subramani, Research Director, IMEA; and Ainsley Woods, Research Director. Together, these experts have been tracking hyperscale investments, cloud infrastructure evolution, and AI revenue generation patterns across global markets.

The Double-Counting Concern: Are AI Revenues Real?

One of the session’s most critical discussions centered on a concern that has been circulating throughout the industry: whether hyperscalers are truly generating new revenue from AI, or simply recycling existing workloads through new pricing models.

“There’s this concern that if off-takers are recycling it or reselling it, and not making money off of that compute, then that’s when I started looking around at the structures,” one analyst explained. “For now, it seems like there is a lot of healthy revenue from very basic off-takers.”

The panel pointed to concrete examples of AI-driven revenue growth. They highlighted Cursor, a development tool that now derives roughly 30% of its revenue through AI capabilities. “If LLMs get to the level of a senior software engineer at Google, just by using all these papers and other resources, there’s real value being created,” the analyst noted.

The consensus: As long as overall net new revenue tied to AI can be measured and we’re still in the very initial stages, the growth is legitimate. “It’s all about recognizing that new revenue,” the analyst emphasized.

Existential Investment: Why Hyperscalers Are Spending “Stupid Amounts of Capital”

The discussion took a fascinating turn when panelists examined the psychology driving hyperscale investment decisions. One analyst posed a revealing question to frame the issue: “If I asked the audience today, how many of you can say with 100% certainty that your jobs will not be displaced by AI, most of us would not be able to say that with 100% certainty.”

This uncertainty, the panel argued, is exactly what’s driving hyperscaler behavior.

“That’s exactly how the hyperscalers feel as well, and that’s why they’re investing stupid amounts of capital because that’s an existential threat to their leadership,” the analyst explained. “They’re not really investing their capex in a purely rational economic framework. They’re investing in it because they don’t want to be the next Cisco.”

This perspective reframes the massive capital expenditure not as irrational exuberance, but as strategic survival. The hyperscalers remember the cautionary tales of technology giants that failed to adapt to paradigm shifts, and they’re determined not to repeat those mistakes.

Cloud as a Delivery Vehicle: Learning from Historical Precedent

To assess the sustainability of current AI infrastructure growth, the panel drew parallels to previous technology transitions. Specifically, Microsoft’s shift from licensing Windows Server to delivering it as a cloud infrastructure service on Azure.

“I view the real clouds that are taking GPUs from Nvidia as a delivery vehicle, a service provider for cloud infrastructure not unlike what Microsoft did with Windows Server,” one analyst explained. “Instead of selling licenses and having customers install software in back offices, they simply delivered it as a cloud infrastructure service off the Azure platform.”

This historical comparison provided the panel with confidence in the current trajectory. However, they acknowledged a critical difference: velocity.

“The Windows Server transition happened over the course of five to ten years and it was slow-moving,” the analyst noted. “What’s happening now moves so fast. That basic velocity at which it happens gives us optimism, but it also makes it harder to predict.”

Projecting Five Years Out: Methodology and Data Points

When asked about revenue projections five years into the future and the data points supporting such tremendous growth, the panel outlined their analytical approach.

“First of all, I’ll say it once again: it’s very difficult to see what’s going to happen,” one analyst acknowledged candidly. “But the methodology is grounded in how we view cloud infrastructure growth historically.”

The team’s approach involves:

1. Historical Evidence Analysis: Examining how first-generation cloud and free hyperscale infrastructure evolved, then comparing that to current hyperscale growth patterns.
2. Phase-Based Growth Modeling: Dividing growth into distinct phases to understand acceleration patterns and inflection points.
3. Fundamental Technology Comparison: Recognizing that “GPU clouds are the same thing, right? Servers with chips and storage.” Building projections on these technological fundamentals.

“When something has no historical precedent, the best way to understand it is to look at the closest analog,” the analyst explained. “That’s how we did it, we built on historical patterns and then tried to say, ‘Okay, this is going to be bigger and faster,’ but it’s based on actual precedent.”

Key Takeaways: Why This Matters for the Industry

The analyst panel’s conclusions carry significant implications for stakeholders across the digital infrastructure ecosystem:

1. AI Revenue is Real: Despite concerns about double-counting, evidence suggests genuine net new revenue generation from AI workloads, with companies like Cursor demonstrating meaningful AI-driven revenue streams.
2. Investment is Strategic, Not Irrational: Hyperscaler capital expenditure, while massive, reflects existential competitive dynamics rather than speculative excess. Companies are investing to avoid obsolescence.
3. Historical Models Provide Guidance: While the current AI infrastructure buildout is unprecedented in scale and speed, previous cloud transitions offer methodological frameworks for understanding and projecting growth.
4. Velocity Creates Uncertainty: The rapid pace of change makes prediction challenging, but it also creates opportunities for those who can move quickly and adapt.
5. Fundamentals Still Matter: Despite the transformative nature of AI, the underlying infrastructure still consists of servers, chips, and storage—grounding analysis in tangible technological realities.

For infrastructure operators, investors, and technology providers, these insights suggest that while caution is warranted given the pace of change, the fundamental economics of AI infrastructure appear sound. The key will be distinguishing between companies delivering genuine value and those merely riding the hype cycle.

Infra/STRUCTURE 2026: Save the Date

Want to tune in live, receive all presentations, gain access to C-level executives, investors and industry leading research? Then save the date for infra/STRUCTURE 2026 set for October 7-8, 2026 at The Wynn Las Vegas. Pre-Registration for the 2026 event is now open, and you can visit www.infrastructuresummit.io to learn more.

The post The Economics of AI Infrastructure: Separating Hype from Reality appeared first on Data Center POST.

Submarine Networks World 2025, held September 24–25 in Singapore, once again cemented its position as the premier global gathering for the subsea communications community. Bringing together leaders across undersea infrastructure, cable technology, and digital connectivity, this year’s event delivered fresh insights on innovation, collaboration, and the future of resilient global networks.

Event Overview

Hosted at the Sands Expo & Convention Centre, Submarine Networks World 2025 welcomed more than 1,000 attendees from across the industry, including cable operators, technology vendors, regulators, investors, and infrastructure developers. The program featured over 130 speakers and more than 70 sponsors and partners, including recognized industry leaders such as Nokia, Ciena, and Digital Realty. Keynotes, debates, technical theatre presentations, and high-value networking sessions created a dynamic environment for exchanging ideas and forecasting trends shaping the subsea ecosystem.

Key Themes and Highlights

Cable Resilience and Security

A central theme throughout the event was the industry’s increasing focus on resilience. Panels explored strategies for diversifying routes, improving fault detection, strengthening data openness, and protecting subsea assets from risks ranging from climate events to geopolitical tensions.

Technological Innovation

Speakers highlighted major advancements transforming the subsea landscape, including pluggable technologies for submarine networks, fiber sensing for predictive maintenance, and the evolution toward petabit-scale cable systems. These innovations mark an important shift as operators aim to deliver higher capacity with greater efficiency.

Scaling to Meet Demand

With global bandwidth needs accelerating due to cloud growth, AI workloads, and digital expansion, the conference underscored the pressing need for large-scale infrastructure development. Experts noted that traffic requirements could double by 2030, emphasizing the urgency for new systems, expanded routes, and increased investment.

Sustainability and Transparency

Sustainability also took center stage, with leaders calling for enhanced mapping practices, standardized open-data models, and more environmentally responsible construction. The conversation pointed toward building not only faster and stronger networks, but smarter and cleaner ones as well.

Regional Collaboration

Sessions highlighted the rising influence of emerging markets, particularly in the Asia-Pacific region. Indonesia stood out for showcasing its connectivity initiatives, unique subsea challenges, and growing leadership role in regional digital infrastructure.

Community Impact and Takeaways

Attendees praised the depth and relevance of the discussions, as well as the diversity of perspectives from C-suite executives to highly specialized engineers. The event reinforced a collective commitment to innovation, security, and global cooperation as the subsea community navigates rising demand and an increasingly complex operating environment.

Looking Ahead

Submarine Networks World 2025 reaffirmed its status as the definitive annual forum for subsea connectivity. By bringing together the industry’s brightest minds and boldest strategies, the event set the tone for continued progress heading into 2026 and beyond. With momentum building across technology, sustainability, and international partnership, the global subsea communications community is well positioned to meet the challenges and opportunities of the next decade.

To learn about the upcoming Submarine Networks World 2026 and to register for the event, visit www.terrapinn.com/conference/submarine-networks-world/index.stm.

The post Submarine Networks World 2025: Advancing Global Connectivity Beneath the Waves appeared first on Data Center POST.