CPower, Bentaus and Supermicro Successfully Flex AI Compute Load for Demand Response

CPower, Bentaus and Supermicro (NASDAQ: SMCI) completed a joint demonstration showing GPU-based AI compute can provide fast, reliable demand response to wholesale power markets.

Rhea-AI Summary

CPower, Bentaus and Supermicro (NASDAQ: SMCI) completed a joint demonstration showing GPU-based AI compute can provide fast, reliable demand response to wholesale power markets. The test dispatched CAISO signals through Bentaus to Supermicro-managed GPU clusters, responding in under 20 milliseconds and cutting electricity use by up to 75% while maintaining AI workloads and SLAs.

The partners plan further collaboration with data center operators and markets including PJM, ERCOT and SPP to scale grid-integrated AI compute.

Positive

• Sub-20 ms response from GPU clusters to wholesale market signals
• Up to 75% electricity reduction for AI workloads during grid response events
• Maintained AI workloads and SLAs while performing dynamic curtailment

Negative

• None.

Argus Feb 25 session

+7.93% close to close Open Argus

Details

News Market Reaction – SMCI

In the Feb 25 session, SMCI gained 7.93%, reflecting a notable positive market reaction.

Data tracked by StockTitan Argus on the day of publication.

Market Context

The stock moved +7.9% in the session following this news. A strong positive reaction aligns with rec...

Analysis

The stock moved +7.9% in the session following this news. A strong positive reaction aligns with recurring AI-focused catalysts for SMCI, which has recently highlighted new GPU platforms, cluster solutions, and manufacturing capacity. Historically, AI headlines produced an average 1-day move of -0.57%, with both rallies and selloffs. The latest grid-interactive AI test adds a new angle—energy flexibility—while shares still trade below the 200-day MA, a backdrop that could influence how durable any upside proved.

Key Figures

Current AI power: 5 GW Potential AI power: 50 GW Dispatch response time: 20 milliseconds +1 more

Current AI power

5 GW

Stated current U.S. AI power capacity

Potential AI power

50 GW

Potential U.S. AI power capacity by 2030 mentioned in article

Dispatch response time

20 milliseconds

GPU infrastructure response to full dispatch signal cycle

Power reduction

75% less electricity

Electricity reduction by Supermicro servers when responding to grid conditions

Previous AI Reports

5 past events · Latest: Jan 06

Same Type 5 events

1. Jan 06

   AI systems launch

   24h Move

   +1.6%

   Unveiled new desktop, edge, and client AI systems including Super AI Station.

2. Jan 05

   AI capacity expansion

   24h Move

   +1.6%

   Expanded US manufacturing and liquid-cooling for NVIDIA Rubin-based AI platforms.

3. Dec 11

   Sovereign AI engagement

   24h Move

   -5.0%

   Early-stage role in GPU-based sovereign AI data center planning in the Philippines.

4. Nov 19

   AI server launch

   24h Move

   -3.4%

   Launched 10U air-cooled AI server using AMD Instinct MI355X GPUs with higher TDP.

5. Nov 18

   AI factory clusters

   24h Move

   +2.4%

   Announced turnkey AI Factory cluster solutions based on NVIDIA reference architectures.

24h Move is the share-price change in the day after each event; other market factors may also have contributed.

Key Terms

virtual power plant, vpp, demand response, wholesale electricity market, +4 more

8 terms

virtual power plant technical

"CPower Energy ("CPower"), a leading Virtual Power Plant (VPP) platform, and Bentaus"

A virtual power plant is a software-driven system that links many small energy sources — such as rooftop solar panels, home batteries and flexible electricity use — and operates them together as if they were one larger power station. For investors, it matters because it can turn scattered assets into a reliable revenue stream and reduce costs for energy providers, much like organizing many individual taxis into a single fleet that can be dispatched efficiently to meet demand and earn steady fees.

vpp technical

"CPower Energy ("CPower"), a leading Virtual Power Plant (VPP) platform, and Bentaus"

A virtual power plant (VPP) is a network of distributed energy resources — such as rooftop solar, home batteries, electric vehicle chargers, and small generators — that are coordinated by software to act like a single power station. Investors care because VPPs can create new revenue streams and cost savings by selling capacity, balancing the grid, and avoiding expensive infrastructure upgrades, much like pooling many spare rooms into a rentable hotel can unlock steady income and efficiencies.

demand response technical

"data centers can contribute to demand response and energy flexibility programs at the server"

Demand response is a program or market mechanism where electricity users are paid or incentivized to reduce or shift their power use when the grid is stressed or prices are high, similar to turning down nonessential appliances during a heat wave to ease a traffic jam. It matters to investors because it can lower peak energy costs, affect utility revenues and market prices, and create opportunities for companies that provide the software, equipment, or services that enable those load changes.

wholesale electricity market regulatory

"CPower dispatched real-time wholesale electricity market signals from California"

A wholesale electricity market is where big buyers and sellers trade large blocks of electricity — think of it like a farmers' market for power where generators sell and utilities or large companies buy ahead of delivery. Prices set in this market influence revenue for power producers and operating costs for energy users, so changes in supply, demand, weather, or regulation can directly affect company profits, utility bills, and the value of energy-related investments.

california independent system operator regulatory

"market signals from California Independent System Operator (CAISO) through the Bentaus"

A nonprofit grid operator that manages the flow of electricity across most of California and coordinates the wholesale energy market, like an air traffic controller for power lines. It balances supply and demand in real time, plans transmission needs, and sets market rules that influence when and how much electricity is bought and sold. Investors watch it because its decisions affect energy prices, utility costs, reliability risks, and the economics of power producers and clean‑energy projects.

caiso regulatory

"market signals from California Independent System Operator (CAISO) through the Bentaus"

CAISO is the nonprofit organization that operates and manages the high-voltage electricity grid and wholesale energy markets for most of California and parts of neighboring states. Think of it as the traffic controller for power: its rules, market prices and reliability decisions influence utility income, the economics of renewable energy and batteries, and the cost of electricity—factors that can materially affect earnings and investment risk for energy-related companies.

gpu technical

"Supermicro-managed GPU infrastructure, which responded to a full dispatch signal"

A GPU (graphics processing unit) is a specialized computer chip designed to handle many calculations at once, originally for rendering images and video but now widely used for tasks like artificial intelligence, data analysis and high-performance computing. Investors watch GPU demand and prices because strong sales often signal growth for chip makers and their customers, affect profit margins and capital spending, and can forecast wider trends in gaming, AI adoption and cloud services.

wholesale power markets regulatory

"participants across multiple U.S. wholesale power markets, including PJM, ERCOT and SPP"

Wholesale power markets are where large quantities of electricity are bought and sold between generators, utilities, and big energy buyers rather than to individual households. Think of it like a commodities exchange or farmers’ market for electricity: prices and contracts set here determine what utilities pay to secure supply and what generators earn, so changes in these markets directly affect energy company revenues, utility costs, and investor risk and returns.

AI-generated analysis. How Rhea-AI works. Not financial advice.

See more from StockTitan in Google Search and AI answers. Adds StockTitan as a preferred source · opens Google

Add on Google Not now

Joint test proves GPU-based workloads can be real-time grid resources as power demand for AI surges tenfold

BALTIMORE and DALLAS and SAN JOSE, Calif., Feb. 25, 2026 /PRNewswire/ -- CPower Energy ("CPower"), a leading Virtual Power Plant (VPP) platform, and Bentaus, a leading developer of AI-driven energy orchestration platforms, in collaboration with Supermicro (NASDAQ: SMCI), a global leader in Application-Optimized Total IT Solutions, today announced the successful completion of a joint demonstration of how AI compute infrastructure can provide fast, reliable flexibility to the electric grid. In the test, CPower dispatched real-time wholesale electricity market signals from California Independent System Operator (CAISO) through the Bentaus energy orchestration platform to Supermicro-managed GPU infrastructure, which responded to a full dispatch signal cycle in under 20 milliseconds.

As U.S. AI power capacity rises from 5 GW to potentially over 50 GW by 2030, the test's results demonstrate that data centers can contribute to demand response and energy flexibility programs at the server level, providing critical grid relief when demand peaks or prices soar, without compromising operations. The cluster of Supermicro servers with NVIDIA B200 GPUs included in the test maintained AI workloads and service-level agreements while using up to 75% less electricity when responding to grid conditions. This was successfully demonstrated through testing under IT load.

"The ability to use AI compute loads to help the grid keep up with rapid load growth has been largely unrealized until now. This successful end-to-end test is a breakthrough, accelerating flexible compute's emergence as a controllable, grid-interactive load, and moving it from the theoretical to the actual," said Michael D. Smith, CEO, CPower.

The test established a foundational architecture for integrating AI compute into demand response by validating abilities such as translating energy market signals into compute-level orchestration actions, dynamically curtailing and restoring AI workloads across GPU clusters and coordinating compute control with infrastructure lifecycle management. CPower, Bentaus and Supermicro will collaborate with data center operators and market participants across multiple U.S. wholesale power markets, including PJM, ERCOT and SPP, to accelerate the transition to grid-integrated AI infrastructure.

"Demonstrating that GPU workloads can react to market signals in the blink of an eye is an important step toward aligning the growth of AI with the operational realities of the grid. AI factories can evolve from passive consumers of electricity into active participants in wholesale power markets," said Robert Davidoff, CEO, Bentaus.

For those interested in learning more about AI workloads for demand response, register for CPower's upcoming webinar featuring Bentaus and Supermicro: From AI Load to Grid Asset: How CPower, Bentaus & Supermicro Are Making AI Compute Flexible.

About CPower Energy
CPower is a leading Virtual Power Plant platform, monetizing the value of customer-sited energy to intelligently strengthen the grid. For over a decade, we've made turning flexible energy into revenue simple for partners and large energy users such as businesses, manufacturers, public institutions and healthcare organizations, delivering $1.4 billion from demand response and energy flexibility programs to customers since 2015. Learn more at https://cpowerenergy.com/. 

About Bentaus
Bentaus develops AI-driven energy orchestration platforms that unify power markets, grid telemetry and compute infrastructure. Its technologies—including PAO (Power Asset Orchestrator), Ziani EOS (Energy Orchestration System), Ziani Control Plane, and GPUflex—enable real-time coordination between energy systems and AI workloads. Learn more at https://bentaus.com/.

View original content to download multimedia:https://www.prnewswire.com/news-releases/cpower-bentaus-and-supermicro-successfully-flex-ai-compute-load-for-demand-response-302697010.html

SOURCE CPower Energy

FAQ

AI-generated questions and answers. How Rhea-AI works. Not financial advice.

What did CPower, Bentaus and Supermicro demonstrate about AI compute and demand response on Feb 25, 2026 (SMCI)?

They demonstrated that GPU-based AI compute can act as a fast demand response resource, reacting in under 20 milliseconds. According to the company, the test showed clusters could curtail and restore AI workloads while preserving service-level agreements and reducing electricity use.

How quickly did Supermicro-managed GPU infrastructure respond to CAISO signals in the SMCI demo?

The GPU infrastructure responded to a full dispatch signal cycle in under 20 milliseconds. According to the company, that sub-20 ms latency was achieved while translating wholesale market signals through the Bentaus orchestration platform to server-level actions.

How much energy reduction was achieved by Supermicro GPU clusters during the demand response test (SMCI)?

The tested GPU cluster used up to 75% less electricity when responding to grid conditions. According to the company, this reduction occurred during dynamic curtailment while AI workloads and SLAs were maintained under IT load testing.

Will the SMCI, CPower and Bentaus collaboration expand to other U.S. power markets like PJM and ERCOT?

Yes, the partners plan to collaborate with data center operators across multiple wholesale markets including PJM, ERCOT and SPP. According to the company, the aim is to accelerate adoption of grid-integrated AI compute infrastructure in those markets.

Did the Supermicro (SMCI) test maintain AI performance while providing grid flexibility?

Yes — AI workloads and service-level agreements were maintained during grid-responsive curtailment. According to the company, the test validated that compute-level orchestration can curtail and restore GPU workloads without compromising service commitments.

How does the Feb 25, 2026 SMCI demonstration affect the potential role of data centers in wholesale power markets?

The demonstration indicates data centers can evolve into active participants in wholesale power markets by offering controllable compute loads. According to the company, validated orchestration and sub-second response enable participation without sacrificing AI service levels.