Models & Labs · United States

Data Center Energy Trends Force a Rethink of Chip Power Delivery

Credited to SemiWiki · semiwiki.com · 7h ago

Intel Foundry published a two article series on chip power delivery that are worth reading. Here is a quick summary:

Artificial intelligence is changing the requirements for semiconductor innovation. Faster transistors and denser packaging remain essential, but their benefits increasingly depend on another engineering challenge: delivering stable power to the places where computation happens. Intel Foundry’s two-part series on power delivery explains why electricity distribution and voltage regulation are becoming central to the future of data center performance.

The scale of projected demand makes this shift significant. In the first article, Intel Foundry cites Gartner estimates that worldwide data center power demand will rise from 104 gigawatts in 2025 to 132 gigawatts in 2026, reaching 290 gigawatts by 2030. AI-optimized servers are expected to account for 31% of consumption in 2026 and surpass conventional servers by 2027.

These are forecasts, but they highlight a practical constraint: computing expansion depends on available electricity and cooling capacity. Operators must divide limited power budgets among processors, memory, networking, and storage. Improving how electricity travels through compute hardware could help increase useful output within those budgets. That makes power delivery relevant to infrastructure economics as well as chip design.

Conventional frontside power delivery routes electricity and signals through shared metal layers. As designs become more complex, those functions compete for space. Rising current increases voltage losses through resistance, while inductance complicates responses to sudden workload changes. Engineers compensate with wider connections, additional capacitors, and operating margins, consuming resources and potentially sacrificing efficiency to maintain reliable operation.

Intel Foundry’s RibbonFET Gate-all-around (GAA) transistor architecture with Intel 18A, separates these functions by moving power routing to the chip’s backside. Dedicated power layers create shorter paths to transistors and free frontside resources for signals. Intel says this arrangement reduces resistance, inductance, and IR drop, the voltage lost as current passes through resistance, while improving the stability of the supply.

Local energy storage complements that architecture. Intel Foundry describes Omni MIM on-die capacitors, substrate-embedded capacitors with through-silicon vias, and embedded deep trench capacitors. These components provide nearby charge when demand suddenly increases. For bursty AI inference and high-performance computing workloads, that support can help limit brief voltage dips that otherwise threaten timing or require more conservative operating settings.

The second article extends the problem to chiplets and stacked dies. Combining multiple dies enables greater bandwidth and functional integration, but also concentrates current demand and increases the number of voltage domains. Package connections must carry power alongside dense signal traffic. Consequently, advanced packaging can create electrical bottlenecks even as it solves other constraints on performance scaling.

Intel Foundry’s response is to move voltage conversion closer to the load. Distributing power at a higher voltage reduces the current required to carry a given amount of power; conversion near the compute die then supplies its lower operating voltage. This approach can reduce distribution losses, although its overall benefit depends on regulator efficiency and the complete delivery network.

The authors describe several implementations. Fully integrated voltage regulators place the final conversion stage on the processor, supported by CoaxMIL magnetic inductors designed for higher current density. Continuous capacitive voltage regulators, or C2VR, use capacitors instead of traditional inductors. Dedicated regulator chiplets offer placement flexibility, including beneath compute dies or within package structures, potentially shortening demanding power distribution paths.

Why it matters is the connection between electrical stability and usable performance. A processor cannot consistently exploit its capabilities if supply voltage fluctuates beyond acceptable limits. Better routing, localized charge storage, and nearby regulation could reduce the margins required for safe operation. That creates opportunities for more efficient computing, while helping designers manage increasingly complex systems.

Bottom line: Scaling computation now requires coordinated advances across transistors, wiring, capacitors, regulators, and packaging. Power delivery is becoming a decisive factor in whether additional silicon produces worthwhile performance within increasingly constrained energy budgets.

Driving Power Delivery Innovations for the AI Data Center Era

Intel Foundry: AI performance now depends on the whole system

ASML Has High-NA and Chipmakers Can’t Say No

High-NA EUV Moves From Experiment to Manufacturing

CapitalUnited States

AI Model Evaluator Arena Valued at Over $3 Billion in Round

Credited to Bloomberg$ · bloomberg.com8h ago

Useful

Arena Intelligence Inc., the startup behind a popular artificial intelligence model leaderboard, has raised $200 million in funding at a $3.1 billion valuation and plans to expand into measuring AI safety, tapping into growing demand to evaluate the technology’s performance and…

View the article · $Bloomberg

Front Method

Models & Labs · United States

Rein Security raises $25M to rein in insecure AI agents

Credited to SiliconANGLE · siliconangle.com · 11h ago

Photograph · SiliconANGLE

Rein Security Inc. said today it’s trying to help enterprises protect themselves against two distinct security threats relating to agentic artificial intelligence after raising $25 million in early-stage funding from Glilot Capital, Sienna Venture Capital, Corner Ventures, Atlacle and RNP Capital Advisors today.

The Series A round, which brings Rein’s total amount raised so far to $35 million, comes at a time when the risks of AI agents have become more acute than ever. In fact, the startup argues that autonomous agents bring two entirely new types of risks that enterprises have never had to deal with before, which means there’s no hope that their existing cybersecurity platforms are going to be able to protect them.

For enterprises that are adopting AI agents themselves, the big challenge is to secure those digital workers and make sure they don’t unwittingly cause problems or leak crucial secrets. And for everyone else, regardless of whether they’re using agents, they’re going to need to learn how to protect their information technology infrastructure from sophisticated, hacker-controlled agents, which are capable of continuously probing organizations’ defenses to search for vulnerabilities they can exploit.

Rein co-founder and Chief Executive Matan Bar-Efrat (pictured, left, alongside co-founder and Chief Technology Officer Netanel Rubin) told SiliconANGLE the main issue with autonomous AI agents is that the risks are completely new and therefore hidden from most security teams. For companies working with AI agents, the main danger is the downstream actions taken by those autonomous systems, which are usually based on instructions received from other systems and agents, rather than their human operators.

When AI agents are given a task to complete, they’ll strive to finish it no matter what, without thinking of the negative consequences their actions may have. To secure these agents, Bar-Efrat argues that companies need a way to understand every action they take so they can prevent the harmful ones from happening.

“Enterprise AI agents are becoming essential to how businesses operate, but security hasn’t kept pace with the autonomy these systems now have,” Bar-Efrat explained. “This funding enables us to keep building the infrastructure enterprises need to deploy AI agents with visibility, control and confidence.”

Rein Security tries to secure AI agents with a proactive approach that prioritizes prevention and protection at the execution layer with a platform that operates directly at runtime. Its platform is deployed as a lightweight sidecar that analyzes everything from the user’s prompts to the application programming interface calls and the database access.

It works by creating a baseline of each agent’s behavior, so it can understand what’s normal and what isn’t. Then it applies contextual, action-level controls that prevent deviations in real time, without shutting down the agent completely.

“Our approach is to micro-segment the resources that an agent can interact with and enforce controls at the point of execution,” Bar-Efrat explained. “We aren’t terminating the agent process, thread or the request, we only prevent the specific malicious action from taking place while allowing the rest of the workflow to continue uninterrupted.”

Rein’s understanding of agents’ baseline behavior is key, because there are so many ways that AI agents can be compromised that it’s basically impossible to prevent them all from occurring, Bar-Efrat said. Most agentic security tools are focused on identifying specific techniques and payloads used to compromise AI agents, such as indirect prompt injections, zero-day vulnerabilities in their code and misconfigurations that can allow unauthorized access or actions.

As far as Bar-Efrat is concerned, the method used by the attacker is irrelevant. What matters more is being able to understand an agent’s legitimate or normal behavior, so that any deviations from this can immediately be flagged.

“Attackers will continue to find new ways to manipulate AI systems, so rather than trying to predict every possible attack technique, Rein focuses on the agent’s behavior at runtime and the impact that an action could have on the enterprise,” he said.

Rein’s platform has already been deployed by enterprises such as Flex Ltd., Swimlane Inc., Lemonade Inc. and Dun & Bradstreet Holdings Inc. to secure thousands of new AI agents that are executing millions of actions each week. The company is also playing an important role in helping enterprises to understand the new risks AI presents. At Black Hat USA 2026 in August, its research team Agent Breakers showed how it was able to compromise the AI shopping agents of a top-five U.S. retailer, before detailing how that vulnerability can be prevented.

Rein has already grown its customer base by more than five-times this year, and the money from today’s round will help it to keep up that momentum by supporting a global hiring push and funding its product research.

Sienna Venture Capital’s Thomas Visan said the direct impact AI agents have in production environments means that companies cannot compromise on security a moment longer. “The frameworks that exist today are simply not equipped to handle this transition,” he said. “Rein’s differentiated platform explicitly addresses current, as well as new agentic AI security threats, giving the company a distinct position in this rapidly evolving market.”

Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.

Are you an AWS customer? Support SiliconANGLE financially by buying your AWS services from our Marketplace portal page and links: https://siliconangle.com/aws-marketplace/

Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.

Automation Anywhere acquires Boost.ai to expand customer-facing voice AI

Liquid AI builds personal AI around device-level context

Willow taps CoreWeave to simplify AI model training

Oratomic raises $475M for neutral-atom quantum computing, Universal Quantum nets $100M+

Docusign adds a desktop PDF editor to its Intelligent Agreement Management platform

Harness acquires Augment Code to bridge software engineering from idea to deployment

EMERGING TECH - BY KYT DOTSON. 4 HOURS AGO

Join our community