Power & Sites · United States

Why Custom Silicon Matters in AI Data Centers

Credited to EE Times · eetimes.com · 2h ago

Photograph · EE Times

Custom silicon lets AI data centers cut power, latency and data movement while improving economics at scale.

General-purpose processors are not going away. But increasingly, they are no longer being asked to do everything. AI, machine learning, real-time analytics, and hyperscale cloud workloads put very different demands on data center infrastructure than the applications that shaped the traditional server. Performance still matters, but so do power, latency, data movement, and the ability to optimize an entire system around a specific workload.

That is why custom silicon has moved from a specialized option to an increasingly important part of data center architecture. Instead of forcing every workload through the same general-purpose architecture, designers can optimize silicon around the functions that matter most. The point is not simply to build a faster chip. It is about improving how the entire system performs.

Optimize for the workload

Custom silicon starts with a different question: What does this workload actually need the hardware to do? General-purpose CPUs are designed to handle an enormous range of software. That flexibility is essential, but it also means silicon resources are devoted to functions a particular workload may rarely use. With custom silicon, those resources can instead be concentrated on the operations that matter most.

AI illustrates the tradeoff clearly. These workloads rely heavily on large matrix calculations and tensor operations. Purpose-built AI accelerators are designed specifically to accelerate those calculations. The result is significantly faster performance for targeted workloads. Tasks that might require large CPU clusters can often be executed much more efficiently on specialized hardware designed for the job.

Do more within the power budget

In many data centers, the limiting resource is no longer floor space. It is power. As compute density increases, power delivery and cooling increasingly determine how much additional infrastructure can actually be deployed. Custom silicon improves performance per watt by eliminating unnecessary circuitry and focusing only on the operations needed for a particular workload.

Because specialized processors are more efficient, they can perform the same tasks while consuming less energy. This reduces electricity costs and can allow organizations to deploy more computing capacity within the same power budget. For large-scale infrastructure environments, even modest improvements in power efficiency can translate into significant long-term savings.

Remove latency from the critical path

Many modern services depend on rapid processing and near-instant responses. AI inference, real-time analytics, and high-performance networking all require systems that can process data with minimal delay.

Custom silicon helps reduce latency by implementing key functions directly in hardware rather than relying on layers of software running on general-purpose processors. For example, specialized accelerators can execute machine learning models directly on-chip, while networking processors can handle packet processing without involving the main CPU. This reduces processing overhead and enables faster response times for latency-sensitive applications.

Move data more efficiently

One of the biggest challenges in modern computing is moving data. Large workloads can spend enormous amounts of time and energy moving data between memory, processors, and accelerators. At data center scale, that movement can become as important to system performance as the computation itself.

Custom silicon can be designed to optimize data flow within the processor itself. Engineers can incorporate larger on-chip buffers, specialized memory pathways, and streamlined data pipelines tailored to specific workloads. By reducing unnecessary data movement, custom processors improve overall system throughput and allow applications to process information more efficiently.

Make the economics work at scale

Custom silicon does not make economic sense at every scale. Designing a chip requires substantial upfront investment, engineering resources, and time. But once deployments reach sufficient volume, relatively small gains in performance, power, or utilization can be multiplied across thousands, or millions, of systems.

When a custom chip improves performance or reduces power consumption, those benefits multiply across the entire infrastructure. Over time, efficiency gains can offset development costs and reduce the total cost of operating a data center.

Control more of the roadmap

Organizations built entirely around off-the-shelf processors ultimately inherit someone else’s product roadmap, development cadence, and design tradeoffs. Custom silicon allows companies to design hardware that aligns directly with their software platforms and operational requirements.

This level of control enables tighter integration between hardware and software, faster innovation cycles, and the ability to tailor infrastructure for emerging workloads such as AI. For many cloud providers and large enterprises, this flexibility has become an important competitive advantage.

From general-purpose to purpose-built

The rise of custom silicon reflects a broader change in how modern data centers are designed. Instead of relying solely on general-purpose processors, infrastructure is increasingly built around specialized hardware optimized for specific workloads.

A modern server environment may include several types of processors working together: CPUs for general applications, AI accelerators for machine learning, and specialized hardware for networking or storage functions.

For data center operators and enterprise IT leaders, custom silicon represents an opportunity to rethink how infrastructure is built and optimized. While not every organization will design its own chips, the growing availability of specialized processors means that data centers can increasingly deploy hardware tailored to their most demanding workloads.

The question for data center architects is therefore not simply, “Which processor is fastest?” Increasingly, it is, “Which parts of the workload are important enough to optimize in silicon?” General-purpose computing will remain essential. But as infrastructure becomes larger, more power-constrained, and more specialized, purpose-built silicon gives architects another lever to optimize performance, efficiency, and economics at the system level.

AI Power Demands Push GaN into Data Center Design

AI Data Centers Make Power, Cooling Critical to Scaling

The AI Boom Has a Gigawatt Accounting Problem

Why Data Centers Must Become System‑Aware

RELATED TOPICS: AI AND BIG DATA, CHIPLETS, SEMICONDUCTORS, SILICON, SOC

Scott Seal is a product marketing manager at Marvell where he is responsible for driving custom silicon and IP marketing for hyperscalers and AI data centers. With over 20 years in product marketing, he has experience working across startups, fast growth, and Fortune 500 IT companies, focusing on buyer personas, messaging, and go-to-market execution. Seal also holds an MBA degree from the Kenan-Flagler Business School, University of North Carolina at Chapel Hill. Follow Scott on LinkedIn

Power & SitesUnited States

Education, healthcare projects boosted construction planning in September

Credited to Construction Dive · constructiondive.com3h ago

Useful

Strength outside the data center boom helped lift the Dodge Momentum Index 3.7% above its September 2025 level.

Strength outside the data center boom is beginning to appear in nonresidential construction planning, according to Dodge Construction Network.

The Dodge Momentum Index, a measure of nonresidential construction projects entering the planning stages, increased 2.4% month over month in September, according to a Wednesday news release. Institutional planning jumped 9.4% during the month, though commercial planning declined 1.3%, according to the report.

“Institutional planning led this month’s growth, alongside a strong uptick in education, healthcare, and government building planning,” said Sarah Martin, director of economic research at Dodge Construction Network. “Milder growth was seen across several other commercial sectors, as well.”

Martin added as planning grows in other nonresidential projects, construction spending in those sectors should post a marginally stronger late 2027.

Within the commercial segment, planning improved month over month in retail, hotel, office building and parking garage projects, according to Dodge. Weaker data center and warehouse planning activity, however, offset much of that growth.

Nevertheless, data center projects still account for the lion’s share of commercial planning activity, despite the weaker figures over the past couple of months. Without the data center work, commercial planning would have dropped 14.9% year over year in September, instead of a 3% dip, according to the report.

Institutional planning, which includes healthcare and education, increased 17.2% year over year in September. The boost pushed the overall index up 3.7% above its September 2025 level.

A total of 45 projects valued at $100 million or more entered planning in September, according to Dodge. The largest commercial projects included:

On the institutional side, the largest projects included:

Permits, Not Power, Halt Google’s Finland Data Centers

Credited to Data Center Knowledge · datacenterknowledge.com · 4h ago

Photograph · Data Center Knowledge

Google had secured access to Finland’s northern grid, but environmental rules are blocking site work at two facilities in its €13 billion data center expansion.

Finland told Google to halt major site work at two planned data centers after regulators found that substantial land clearing and other site preparation had occurred before the required environmental impact assessments were completed.

For data center operators, Google’s experience shows that securing power and grid capacity does not guarantee that construction can proceed. Google had access to significant grid capacity in northern Finland through the Finnish grid operator Fingrid, but the project encountered another constraint when site preparation crossed into permanent environmental alteration.

“Power was the part Google had already solved. Fingrid said the northern grid had room,” said Stephen Sopko, practice lead for semiconductor and deep tech at HyperFrame Research. “The failure was treating forest clearance as preparatory work when Finnish law treats permanent conversion above 200 hectares as the action regulators are there to judge.”

Related: Google’s Grid-Interactive AI Data Centers: From Backup to Grid Partner

The Finnish Licensing and Supervisory Authority, known as LVV, directed Google ’ s subsidiary Tuike Finland Oy to suspend work that significantly alters the environment at planned sites in Muhos and Kajaani. The agency called for the work to stop immediately and no later than Oct. 23.

The request covers tree removal, topsoil removal, excavation, blasting, crushing, earthmoving, drainage work, and construction of site roads and storage areas. Planning, measurements, soil surveys and other limited or reversible activities can continue.

At Muhos, Tuike reported that about 330 hectares had been felled or prepared. At the Kajaani-Otanmäki site, the reported area was just under 200 hectares.

LVV said environmental impact assessment reports for the two projects are expected to be completed in 2026. The agency said allowing major site work to continue could undermine the assessment process by limiting the ability to identify environmental impacts, examine alternatives and prevent harmful effects.

Grid Capacity Wasn’t the Constraint

LVV’s intervention followed a Sept. 18 request for information from Tuike Finland, the Municipality of Muhos, the City of Kajaani and local building-control authorities. Based on the information it received, the agency determined that the preparatory activities warranted intervention under Finland’s EIA Act.

The issue is the sequencing of the work. An environmental impact assessment is intended to gather information about a project’s environmental effects before that information is used in decision-making, including identifying impacts, considering alternatives and determining measures to prevent or reduce environmental harm.

Related: Google Says AI Growth Is Outrunning Grid Decarbonization

With major site alteration already underway, the assessment has less ability to serve that function, according to Tiina Paloniitty, associate professor of environmental and sustainability law at the University of Helsinki, in an email.

“You can often do surveys and reversible site work ahead of the permit.... What you can’t do is clear 300 hectares and call it anticipation.” – Stephen Sopko, practice lead, HyperFrame Research

Google announced the €13 billion investment in September, including new data centers in Muhos, Kajaani and Vaala and an expansion of its Hamina site.

Google selected the northern Finland sites in part because of the available grid capacity near carbon-free electricity sources. Fingrid has said Muhos, Kajaani and Vaala have significant grid capacity in an area where electricity production exceeds consumption.

For operators planning large data centers, the Finnish case shows that environmental approvals can impose separate sequencing requirements even when a site has the electricity and transmission capacity needed for a large facility.

“This is not just a data center rule. It is the rule for any major project,” Sopko said. “You can often do surveys and reversible site work ahead of the permit. You certainly can in Finland. What you can’t do is clear 300 hectares and call it anticipation.”

Related: The Ripple Effect of Data Center Project Cancellations and Delays

Google Says It Followed Forestry Rules

Google said it acted in good faith under Finland’s Forest Act, conducted nature surveys and took measures to protect environmentally valuable areas.

“We understand the concern caused by the situation. In this case, we did not meet the high standards we set ourselves,” said Sondre Ronander, communications manager for Finland and Norway at Google, in a public statement.

Google said it is advancing landscaping and biodiversity plans, including tree planting across 130 hectares at Muhos.

Paloniitty said compliance with forestry rules does not necessarily satisfy separate environmental impact assessment requirements. The two laws serve different purposes, and complying with one does not remove obligations under the other.

“Compliance with one applicable law does not equal … compliance with other applicable laws,” Paloniitty said.

Based on publicly available information from LVV and Finnish media, Paloniitty said it appears that Google failed to comply with the Finnish EIA Act, even after the authority indicated that the projects required an environmental impact assessment.

“Good faith is not a legally relevant argument here,” she said.

Paloniitty said the environmental review cannot fully serve its intended purpose once substantial site work has already altered the area being assessed. That also limits the ability to evaluate alternatives and environmental impacts before decisions are made, she said.

Google’s plans to plant new trees do not resolve that issue, she said.

“New trees do not replace a full-grown forest,” Paloniitty said.

Intervention Is ‘Highly Unusual’

Paloniitty described LVV’s intervention as “highly unusual.” She said she could not recall a comparable intervention at this scale, although such action is not entirely without precedent.

The intervention does not cancel either data center project. It seeks to stop major site work while the environmental assessment procedures are completed.

The projects can continue through limited activities such as planning, measurements and soil surveys.

Tuike Finland has until Oct. 14 to provide LVV with a written explanation and state whether it will comply. Separately, the agency set Oct. 23 as the latest date by which the suspension of major site work must be in effect.

LVV is initially seeking voluntary compliance. If Tuike does not comply, the agency said it will consider initiating a formal enforcement procedure that could be backed by a conditional fine.

This article was originally published on TechTarget.com.

About the Author

Shane Snider is a veteran journalist with more than 20 years of industry experience. He started his career as a general assignment reporter and has covered government, business, education, technology and much more. He was a reporter for the Triangle Business Journal, The Raleigh News & Observer and most recently a tech reporter for CRN. He was also a top wedding photographer for many years, traveling across the country and around the world. He lives in Raleigh, N.C.

Want more Data Center Knowledge stories in your Google search results?

Editor's Choice

Industry Voices

Join the entire data center ecosystem September 21 – 23, 2026, at the Gaylord Texan Resort & Convention Center in Dallas, TX.