Sponsored by Zscaler

Organizations migrating resources to the cloud are often focused on the end state: how the new public cloud environment will allow them to scale capacity in response to growing demands, reduce data center overhead, and deliver new capabilities to the business more quickly.

But of course, success starts with the migration process, and application migration is about far more than merely moving workloads. It requires a comprehensive understanding of application dependencies, consistent visibility into performance, and secure communications between users, applications, and workloads.

Large-scale enterprise cloud migrations commonly last for at least several months, and they sometimes stretch across years. Throughout this period, organizations must ensure users have reliable access to applications across both cloud and on-premises environments, while navigating hidden dependencies and complex cutovers to avoid disrupting the business.

These dependencies often influence the timing and structure of migration efforts, determining which applications must migrate in concert and which can continue operating from different locations. For example, an application may rely on a database that is shared with several other applications. Or software might depend on another service that is not scheduled to move during the same migration wave.

If teams discover these dependencies too late, they may need to delay the migration, reorganize the wave, or maintain temporary connectivity between application components split across on-premises and cloud environments.

Successful migration requires IT and business leaders to plan not only where applications will run, but also how users will reach them, how teams will identify performance problems, and how workloads will communicate securely throughout the transition. Together, Zscaler and AWS provide these capabilities, helping organizations execute cloud migrations more securely, maintain visibility into application performance, and keep migration plans on track.

The Cloud Imperative

In the hype surrounding AI over the past several years, some attention has shifted away from enterprise IT architecture and cloud migrations. However, public cloud migrations remain critical to business success, and Gartner forecasts 21.3% growth in public cloud services in 2026, largely driven by organizations accelerating workload migration and embracing modernization at scale.

Still, around 42% of strategic workloads remain on-premises today. Some organizations are keeping those workloads in-house because of security requirements, latency requirements, or other carefully considered reasons. But many, no doubt, have failed to move resources to the public cloud due to considerable migration challenges, which include security, speed, visibility, and performance.

This hesitation is understandable, but it is causing companies to miss out on business-critical benefits. For example, the AWS Cloud Adoption Framework benchmarks show a 27% decrease in cost per user, a 57% reduction in downtime, and a 34% decrease in security events after a migration to the public cloud.

Organizations that overcome these migration challenges can realize these benefits without the temporary performance issues and security risks that have often accompanied cloud migrations.

Three Major Migration Challenges

Successful cloud migrations deliver greater agility, scalability, and resilience across applications, infrastructure, and IT operations. But to realize these benefits, organizations must overcome three closely connected migration challenges.

Challenge #1: Maintaining Application Access

Organizations rarely move all of their applications to the public cloud all at once. During a phased migration, some applications remain on-premises, while others run in a cloud environment like AWS, creating a temporary (but often prolonged) hybrid environment.

Users still need reliable access to enterprise applications during this period, but providing this access is often more complicated than IT leaders expect. In many environments, application access is still tied to network location through VPNs, IP addresses, and firewall rules. When applications are moved, IT teams must rework access paths and update policies. Even then, these workarounds can add latency, degrade the user experience, and expand the attack surface.

Solution: Zscaler Private Access (ZPA)

ZPA helps organizations discover private applications and connects authorized users directly to these applications, rather than placing users directly on the network. Instead of basing access policies on static IP addresses, network location, or complex access control lists, ZPA grants access based on the user’s identity, device, and the application’s context, providing authorized users with consistent access without expanding network exposure.

Because ZPA decouples user access from network location, organizations can reduce the need to reconfigure user connectivity when applications move between data centers, VPCs, and public cloud environments. As a result, concerns about application access and poor user experience are less likely to delay milestones or undermine confidence in cloud migrations.

Challenge #2: Securing Workloads Across Distributed Cloud Environments

Organizations running application workloads across multicloud environments face growing security complexity. Securing workload communications across these environments adds another layer of challenge.

Distributed workloads must communicate across ingress, egress, east-west, and private network traffic paths spanning clouds, data centers, regions, and hosts. When organizations rely on inconsistent security controls across these environments, they increase the attack surface and create opportunities for lateral movement.

Traditional VPN and firewall architectures don’t always consistently enforce least-privilege access across the application estate and also add to operational costs and complexity.

Solution: Zscaler Zero Trust Cloud

Zscaler Zero Trust Cloud extends zero trust security across multicloud environments by providing real-time visibility, consistent threat and data protection, and host-based microsegmentation for workloads. It secures all traffic paths, including ingress, egress, east-west, and private network traffic, while enforcing uniform security policies across clouds.

By forwarding all workload traffic to the Zscaler Zero Trust Exchange (ZTE), the solution performs cloud-scale TLS inspection and applies threat protection, inline data protection, and strict least-privilege access controls, eliminating potential lateral movement.

By using ZTE as the unified platform to secure workloads across cloud and data center environments, organizations can eliminate the cost and complexity associated with legacy firewalls and VPNs, while also providing flexible security and policy management, as well as consistent protection against threats and data loss.

Challenge #3: Enabling End-to-End Performance Visibility

As applications move to the cloud and users become more distributed, performance depends on multiple components across the user-to-application path.

In this new architecture, performance can be affected by user devices, Wi-Fi networks, internet connections, network paths, security services, cloud environments, or the application itself. Traditional monitoring tools provide separate views of endpoints, networks, and applications, leading to fragmented visibility.

As a result, IT teams may have difficulty determining which component is responsible when users experience lags and dropped connections. Remote and hybrid work further complicates troubleshooting because it requires IT teams to consider home networks and personal devices that they don’t control. When performance problems pop up, these issues can reduce productivity for both end users and IT support professionals.

Solution: Zscaler Digital Experience (ZDX)

ZDX is an AI-powered digital experience monitoring solution delivered as a service from the Zscaler cloud. It provides end-to-end visibility into the user-to-application path, helping IT teams monitor and troubleshoot digital experience issues across distributed environments.

The solution gives IT teams insights into performance across endpoints, Wi-Fi, ISPs, network paths, and applications, helping them quickly triage problems and then deploy remote troubleshooting capabilities. This results in far fewer IT tickets and a superior user experience.

ZDX also helps IT teams measure digital experience across users, locations, and departments to identify trends and improve performance over time. For example, Liberty Mutual Insurance deployed ZDX to improve performance for remote users experiencing persistent issues. The company now uses ZDX across the organization to eliminate issues with service provider latency, wireless routers, and desktop computer memory leaks, among other problems.

Zscaler and AWS: A Unified Approach to Secure Migration

AWS and Zscaler combine cloud infrastructure as a service with consistent zero trust access across hybrid and multicloud estates.

Migrations succeed when cloud infrastructure, access, security, and visibility operate as one architecture, rather than as a disconnected collection of point solutions. AWS provides the scalable cloud foundation and native controls for hosting and connecting applications, while Zscaler extends secure, identity-based zero trust access across users, applications, and workloads wherever they reside.

By unifying access, security, and visibility across AWS and the broader application estate, organizations can reduce migration risk without creating unnecessary complexity. The result is a secure, scalable foundation that supports both the immediate move to the cloud and continued modernization.

Learn How to Modernize Securely

During a recent webinar hosted by Techstrong, Zscaler explored how organizations can address access, security, and performance throughout the application migration lifecycle. Attendees will learn how to identify private applications for migration planning, preserve reliable user access and end-to-end visibility, and secure communications for workloads distributed across cloud and on-premises environments.

We also discuss how AWS-native services and Zscaler capabilities work together to apply consistent zero trust controls across hybrid and multicloud estates.

By building security and visibility into the migration process from the very beginning, organizations can reduce disruption, accelerate migration milestones, and establish a foundation for continued application modernization.

The post Modernize Cloud Migration on AWS. Secure It with Zscaler. appeared first on Techstrong IT.

The findings come from a new report by economic research firm ECOnorthwest and the University of Virginia, funded by the Portland-based Lemelson Foundation. The study provides a detailed look at Oregon’s data center industry, including its electricity requirements, employment, property taxes and planned development.

The state hosts 111 data centers occupying approximately 22.1 million square feet. Another 32 facilities, representing 6.9 million square feet, are either planned or under construction.

The rapid expansion is a major challenge for Oregon’s electricity grid, which must accommodate AI infrastructure as well as businesses and residential customers.

Researchers forecast that Oregon’s data centers could consume nearly 25 terawatt hours of electricity annually by 2030, equivalent to the electricity required to power approximately 2.5 million homes. That would represent 31% to 32% of the state’s total electricity demand.

Meeting this requirement will involve not just more generating capacity, but also the transmission infrastructure to deliver electricity to large computing facilities. The challenge is complicated by differences in how data centers operate.

Eastern Oregon hosts most of the state’s hyperscale facilities, which support massive computing operations and tend to maintain consistent electricity consumption. Western Oregon, in contrast, is expected to host a larger share of future development. Its smaller facilities often serve multiple customers that rent computing space, creating electricity requirements that fluctuate throughout the day.

This variation makes forecasting demand difficult for utilities responsible for maintaining a reliable power supply.

Employment and Property Tax Issues

While data centers require vast amounts of electricity and real estate, their direct employment contribution is comparatively modest. Oregon’s data centers employ approximately 2,630 workers.

Eastern Oregon, with its concentration of hyperscale facilities, accounts for 2,205 of these jobs. In Morrow County, data centers directly employ more than 10% of the local workforce. (These figures exclude some of the jobs associated with construction and other contractors that maintain data center equipment.)

Oregon’s data center industry, like data centers in states across the country, has benefited greatly from incentives designed to attract large technology investments. Researchers estimated that property tax exemptions awarded the industry ranged from $230 million to $450 million, although incomplete information prevented a more precise calculation.

In 2025, operating data centers paid approximately $60.2 million in property taxes, substantially less than the estimated value of the property tax exemptions.

However, the researchers cautioned that exempted taxes should not automatically be considered lost government revenue. Without the incentives, some facilities might never have been built in Oregon.

The study also identified substantial gaps in available data concerning water consumption and the effects of data centers on surrounding communities. These limitations complicate efforts to evaluate the industry’s full costs and benefits.

The post Oregon Data Centers Consume 23% of State’s Electricity, With Demand Set to Climb appeared first on Techstrong IT.

Announced September 18, the governor’s Data Center Accountability Framework includes new oversight of power generation, water use and noise pollution. It also proposes restrictions on non-disclosure agreements that have limited public access to information about data center development.

The initiative is part of a dramatic policy shift for Virginia, which hosts the world’s largest concentration of data centers. Northern Virginia has become a global hub for cloud computing infrastructure, attracting vast investment from tech companies.

Yet the industry’s expansion has created concerns among Virginia residents about electricity prices and the impact of large facilities on local communities. “Community members are demanding action,” Spanberger said.

The governor’s initiative combines executive actions that can begin immediately with legislative proposals that require approval during Virginia’s 2027 legislative session. Spanberger has also established an AI task force to examine the technology’s development and its impact on the state.

Concern About Energy Infrastructure

A key part of Virginia’s shift is a move toward greater oversight of the energy infrastructure supporting data centers. Spanberger is proposing limits on on-site natural gas generation while offering incentives for developers to invest in renewable energy sources, including solar and wind power.

These measures are designed to secure enough electricity to operate computing facilities without adding excessive costs for other electricity customers. Under the proposed framework, non-disclosure agreements would be prohibited for commercial data center projects, while facilities requiring more than 25 megawatts of electricity would be subject to local approval. The measures would give communities greater access to information about proposed developments.

Virginia will also push forward noise regulations and the designation of cooling-water scarcity areas, bringing additional scrutiny to the water resources needed to keep computing equipment operating.

Data Center Tax Breaks Remain Contentious

Earlier this year, lawmakers from both parties sought to eliminate tax incentives that have helped attract technology companies to the state. Spanberger opposed repealing those benefits, arguing that withdrawing existing commitments could damage Virginia’s reputation among businesses considering investments.

The resulting budget agreement established a temporary electricity consumption tax on data centers, with general-fund revenue from the tax capped at $600 million per fiscal year. A legislative panel will also examine the industry’s subsidies and make recommendations for 2027.

Nicole Riley, director of Virginia government affairs for the Data Center Coalition, warned that major policy changes introduced without sufficient industry consultation could undermine investment and economic development. The Natural Gas Coalition of Virginia also raised concerns about proposed restrictions on gas-powered electricity generation.

Virginia’s shift is similar to those happening across the U.S. Governors in Pennsylvania, Texas and New York have pursued measures addressing data center expansion. Nevada Governor Joe Lombardo also announced restrictions on data centers seeking state tax incentives on September 18.

The post Virginia Unveils Sweeping Data Center Restrictions appeared first on Techstrong IT.

Sen. Jon Husted, R-Ohio, attempted Thursday to win unanimous Senate approval for the Ratepayer Protection Act, one day after the House passed the legislation by a 417-3 vote. Sen. Martin Heinrich, D-New Mexico, blocked the request, arguing that the bill lacks the enforcement needed to protect consumers.

The dispute highlights an urgent challenge created by the AI boom: Data centers require vast amounts of electricity, and connecting new facilities often requires expensive upgrades to power generation equipment and related infrastructure. The pressing question for lawmakers and voters is who pays for those upgrades.

The Ratepayer Protection Act would require states to consider standards under which large electricity users cover the full, incremental cost of infrastructure needed to serve them. The legislation applies to large customers consuming 100 megawatts or more.

But the legislation passed by the House does not require states to adopt those standards. Instead, state utility regulators would be required to consider them through a regulatory proceeding. That distinction became the central point of disagreement on the Senate floor.

Heinrich, the ranking Democrat on the Senate Energy and Natural Resources Committee, contends that federal policy should directly require large power users to pay the grid costs they create.

“It’s not enough for us to tell states to consider making data centers pay for grid updates,” Heinrich said. “Rather than voluntary pledges or suggestions to states, Congress needs to pass real legislation with real teeth.”

A Detailed Review

Heinrich sought unanimous consent for his own legislation, the GRID Savings Act. That bill would give the Federal Energy Regulatory Commission a larger role in establishing rules for connecting major electricity users to the transmission system.

Under Heinrich’s proposal, large customers would undergo a detailed review of the infrastructure required for a grid connection and would be responsible for paying for upgrades. The bill also calls for financial commitments from companies before those upgrades move forward.

But Heinrich’s legislation immediately ran into the same Senate obstacle. Sen. Bernie Moreno, R-Ohio, objected to its passage by unanimous consent. The result was that neither proposal advanced.

Husted argued that Congress should move ahead with legislation that has already demonstrated overwhelming bipartisan support. The House version was led by Reps. Gabe Evans, R-Colo., and Kathy Castor, D-Fla.

“The Ratepayer Protection Act represents the most meaningful, bipartisan step Congress could take to protect the American people from higher prices for electricity,” Husted said.

The post Senate Clash Leaves Data Center Energy Cost Rules Unresolved appeared first on Techstrong IT.

Crusoe

Round: Series F, initial closing | Sector: Cloud

Crusoe announced the initial closing of its anticipated $3.9 billion Series F, an oversubscribed round co-led by Atreides Management, Mubadala Capital, and Valor Equity Partners. The AI infrastructure provider plans to use the capital to build AI factories, expand modular data-center capacity, and grow Crusoe Cloud, extending its control from power and facilities into the services enterprise customers consume. The size of the financing underlines how quickly AI capacity is becoming an infrastructure-scale capital commitment.

Temporal

Round: Late-stage funding round | Sector: Cloud

Temporal has raised $550 million in a round led by Lightspeed Venture Partners, with Wellington Management, Goldman Sachs Alternatives’ Growth Equity business, and Tiger Global joining as co-leads. The company’s open-source workflow engine and commercial Temporal Cloud help applications, including AI agents, recover from failure, and the new capital will support global operations and platform research. That puts reliability tooling alongside compute as a core part of the enterprise AI stack.

Exein

Round: Equity funding round | Sector: Cybersecurity

Exein closed a significantly oversubscribed $270 million equity round led by Headline at a $1.7 billion valuation, with participation from Sofina, Goldman Sachs, EIB Group, KfW Capital, and T.Capital, among others. The Rome-based company builds security for connected machines, including robots, vehicles, and industrial systems, and plans to use the funding for product development, acquisitions, and expansion in the United States and Asia Pacific. Its focus extends cybersecurity budgets into the devices that make decisions at the edge.

EUCLYD

Round: Series A | Sector: Semiconductors

Eindhoven-based EUCLYD has raised more than €200 million in a Series A led by Samsung, Somerset Capital Partners, EQT’s Scaleup Europe Fund, and Innovation Industries, with additional backing from EIFO, imec.xpand, BOM, and Quadri. The startup is combining proprietary chips, memory architecture, and data-center systems to improve the energy efficiency of AI infrastructure, with the funding directed toward commercial deployment and technical hiring. The round reflects how much capital is moving toward system design beyond the processor itself.

Cornelis

Round: Funding round | Sector: Networking

Cornelis has raised $205 million in a funding round led by IAG Capital Partners to develop its Active Compute Fabric, a networking architecture intended to reduce the time AI chips spend waiting for data. The company has already begun shipping its product and is building a new generation around an open architecture that can work with different GPUs and accelerators. For enterprise infrastructure teams, the practical question is increasingly how efficiently compute can communicate, not only how much compute is installed.

The Weekly Funding Pulse tracks the most significant IT infrastructure, cloud, semiconductor, networking, and ITSM funding rounds each week. Coverage is curated for enterprise IT professionals and decision-makers.

The post 5 IT Funding Deals to Watch appeared first on Techstrong IT.

Azul’s recent CFO Cloud Cost Optimization Report found that nearly nine in 10 CFOs say cloud spending is growing, and two-thirds call it a board-level issue. The default response is to demand more visibility. Over the past few years, organizations have invested heavily in tools designed to surface waste and optimize spend. Plenty of tools now provide accurate, highly granular visibility. But visibility without incentive doesn’t change anything, because knowing where waste lives is not the same as eliminating it.

Growth Always Wins

CFOs say they care deeply about controlling cloud costs. In practice, growth wins every time.

Whether a business is actively expanding or under pressure to expand, the instinct is to provision first and optimize later. In competitive markets, the fear of losing ground outweighs any efficiency argument. Cloud cost discipline becomes a last resort. It’s something organizations get serious about only when they have no other choice.

A CTO presents a compelling case for infrastructure optimization. The board nods approvingly, then approves a new AI initiative that doubles the compute footprint. The board drives the AI investment, while finance owns cost discipline. The money can’t flow in two directions at once.

This pattern is accelerating. Gartner forecasts worldwide AI spending will hit $2.59 trillion in 2026, a 47% year-over-year increase. BCG’s 2026 AI Radar found that 94% of companies plan to keep investing in AI even without immediate returns, and most are doubling their budgets. Meanwhile, Flexera’s 2026 State of the Cloud Report shows that wasted cloud spend ticked up to 29% for the first time in five years, reversing a half-decade of improvement, as AI workloads are surging faster than organizations can manage.

Work at a Different Layer

If growth will always take priority, the only realistic path is to reduce the cost of growth itself. Not to slow it down, but to make it structurally cheaper to deliver.

Most organizations are looking in the wrong place. The visibility layer tells you what was spent. It doesn’t change what gets spent. A significant leverage point is at the application layer, where compute consumption is created and where the number of servers required, how much memory each one needs, and how hard each one works are determined.

A good example comes from the early days of online travel booking. When consumers were the ones searching for flights, the ratio of searches to actual bookings was manageable, maybe three or four to one. Then automated search engines arrived, and that ratio exploded to 1,000-to-1. It completely transformed the economics of running a travel platform. The same pressure is building now as AI agents and automated workflows multiply the demands on backend systems. The infrastructure needs to be handled more efficiently, or the economics break.

For companies running Java-based workloads—which still represent the majority of production applications in large enterprises—techniques like runtime optimization and JVM tuning can directly reduce compute usage per transaction, not just report on what’s been consumed. That translates to fewer servers, smaller footprints, and lower bills in a way that’s structural rather than temporary.

Making Efficiency Fund What Comes Next

For CFOs navigating the tension between cost control and AI ambition, the most important shift is to stop treating them as competing priorities. In the right sequence, efficiency funds innovation.

Treat efficiency as an innovation enabler, not a constraint: The organizations getting this right aren’t choosing between cost discipline and AI ambition. They leverage cloud efficiency to create budget headroom, enabling sustained AI investment.

Move FinOps from reactive to integrated: For many organizations, FinOps analysis begins after the bill arrives. Instead, cost awareness needs to be embedded in architecture decisions, development workflows, and infrastructure choices from the start.

Make cost an engineering KPI: Cost can’t remain a finance-only metric reviewed at the end of the month or quarter. It needs to become an informed, shared KPI that engineering and finance teams optimize together as part of everyday decision-making.

Cloud economics erode workload by workload, agent by agent, until the bill is everyone’s problem and no one’s responsibility. The reactive dashboard was never the answer. It’s all about integrating cost visibility and discipline into every phase of the SDLC.

The post Why Knowing Your Cloud Waste Doesn’t Reduce It appeared first on Techstrong IT.

The competition is open to for-profit domestic companies that can deliver a complete quantum computing stack. DOE anticipates selecting between three and 10 applicants to participate. Its request for applications calls for a “scientifically relevant quantum computer,” or SRQC, that can operate under quantum error correction on a DOE-relevant scientific workload.

DOE’s benchmark for a first-generation system includes 100 logical qubits and roughly 100,000 hard operations, meaning operations that are especially costly under a system’s error correction scheme. The agency says the specific metrics may vary by architecture. Awardees must also show that the machine can complete a scientific workflow, potentially in conjunction with a classical supercomputer. The RFA notes that the workload does not necessarily have to operate at a scale already beyond classical computing.

Making scientific utility part of the milestone indicates DOE will not judge progress on hardware performance alone. An Office of Science advisory committee report released Thursday takes a similar stance, recommending that quantum systems be evaluated against scientific problems and integrated with high performance computing and other existing research infrastructure.

Funding is also tied to milestones. Selected companies can receive $250,000 after DOE approves a verification and validation plan, followed by $1.25 million for a validated prototype. DOE also plans a $100 million incentive pool to be divided evenly among participants that reach the first-generation SRQC target, plus separate $50 million bonus pools for awardees that meet those requirements while reaching at least 150 and 200 logical qubits.

The competition’s structure gives DOE a direct role in validating the systems. National laboratory personnel are expected to receive physical and virtual access to participating machines and operate them to verify performance. The agency separately opened a lab call with $45 million in planned funding to establish quantum and high performance computing validation and verification capabilities to support the competition.

Darío Gil, under secretary for science, said the competition will help build a new era of compute power through this public-private partnership model.

“This competition will enable new quantum computing capabilities and unleash new computational frontiers, driving new scientific and technological discoveries that have previously only been theorized,” Gil said.

Applications are due Oct. 19, and DOE anticipates announcing initial participant selections no earlier than Nov. 13, 2026. The agency plans to evaluate first-generation systems in September 2028, although it will consider proposals that reach the target as late as 2030.

Most of the planned $215 million depends on future appropriations. DOE says $2.5 million is available in fiscal 2026, while funding for later years is contingent on congressional appropriations.

The post DOE Opens $215M Quantum Genesis Competition for Fault-Tolerant Systems appeared first on Techstrong IT.

The work is being presented this week at IEEE Quantum Week in Toronto, where it received a third-place Best Paper award in the Quantum–GenAI Co-Design & Co-Discovery track.

Training AI To Generate Quantum Circuits

The framework, called DQAOA-GPT, combines the distributed quantum approximate optimization algorithm, or DQAOA, with a generative AI model trained on high-quality quantum circuits. It then uses that model to generate new circuits directly instead of repeatedly tuning each one.

DQAOA works by breaking a large problem into smaller subproblems that can be solved separately and folded back into the overall solution. As those subproblems grow, however, standard DQAOA requires more circuit evaluations and classical optimization steps, driving up the computational cost. To build DQAOA-GPT, the researchers used ADAPT-QAOA, an adaptive method that builds circuits iteratively, to generate high-quality reference circuits that taught the model to map each subproblem to a suitable circuit.

In a benchmark problem with 100 variables, the researchers tested what happened as the smaller pieces handled by DQAOA grew in size. IonQ said the generative approach held runtime nearly flat at about 28 seconds, while the conventional method rose from about 34 seconds to more than 11 minutes. The company said the quality of the AI-generated solutions also roughly doubled as the subproblems grew. Here, solution quality refers to relative accuracy against the best-known results from earlier work, rather than a direct comparison with a classical solver.

GenAI Changes the Cost Curve

The significance of this research is in that very relationship between size, quality and cost. Larger subproblems can capture more of the interactions in the original problem and produce better solutions, but they also make the conventional tuning process much more expensive. The results suggest generative AI could allow researchers to work with larger pieces of a problem without the same increase in runtime. If that result holds at larger scales, it could make hybrid quantum and high performance computing approaches more practical for complex optimization problems in areas such as materials design, logistics and networks.

Nvidia was represented among the research team, and the experiments relied on its hardware and software. The experiments ran on ORNL’s Defiant2 system using a single Nvidia H200 GPU, with Nvidia’s CUDA-Q platform and cuQuantum software used to simulate the quantum circuits. That gave the researchers the same computing environment for testing both the conventional and generative approaches, making the runtime comparison more controlled.

Another important limitation: no quantum processor was used. Because every circuit was simulated, the work does not demonstrate quantum advantage or establish how DQAOA-GPT performs on quantum hardware. The paper describes the study as benchmark-scale validation. The researchers say the framework could eventually be distributed across multiple GPUs and computing nodes because its subproblems can be handled independently. ORNL researchers said they are extending the framework to real-world applications and larger HPC systems. Access the research paper at this link.

The post IonQ, ORNL Use Generative AI To Cut Quantum Optimization Costs appeared first on Techstrong IT.

The group’s core proposal is that data centers would temporarily reduce or shift their electricity consumption when the grid is under heavy strain. In return, utilities and regulators could allow qualifying facilities to connect to the power system more quickly.

The approach addresses one of the largest challenges for AI expansion. Building new computing infrastructure requires enormous amounts of electricity, while connecting large new power users involves lengthy wait periods and sometimes costly grid upgrades.

AEMA argues that data centers do not necessarily need to operate as fixed loads that consume the same amount of grid power regardless of conditions. AI computing workloads can be shifted, storage can provide electricity during high-demand periods, and facilities can use on-site generation. The alliance supports multiple technologies rather than a single method.

Beyond its three founders, 18 launch partners include AI developer Anthropic, semiconductor vendor Analog Devices and energy companies such as Constellation, National Grid, AES, NRG and RWE. Frank Lacey, an energy industry executive, will serve as AEMA’s executive director.

Faster Access

For tech giants, the payoff is faster access to electricity for new AI capacity. Google already manages one gigawatt of flexible load. NVIDIA and Emerald AI have conducted six demonstrations, and a 100-megawatt data center designed for large-scale flexibility is scheduled to begin operating in Manassas, Virginia, before the end of 2026.

Research published in 2025 by Tyler Norris, Google’s head of energy market innovation for AI and infrastructure, found that modest levels of data center flexibility could make room for an additional 100 gigawatts of load on the U.S. power grid.

AEMA will attempt to translate this technical potential into changes in energy regulation. One area of focus is the Federal Energy Regulatory Commission’s June orders directing the six regional grid operators under its jurisdiction to justify or revise their rules for connecting large loads like data centers.

Similar efforts are taking shape across the country. Texas is developing a framework that addresses large flexible loads, while Silicon Valley Power has established a flexible load interconnection program in partnership with Emerald AI. AEMA plans to advocate for technology-neutral rules that evaluate data centers on measurable grid performance rather than requiring a particular energy technology.

Data Center Energy Costs Draw Congressional Attention

The launch of AEMA comes as Washington is also moving to address the cost of supplying power to AI data centers. On Sept. 16, the House passed the bipartisan Ratepayer Protection Act by a 417-3 vote. The bill would require state utility regulators to consider standards under which large data centers pay the added costs of power generation and related infrastructure. States would have authority over how those standards are applied.

The legislation was prompted by growing concern that the AI infrastructure buildout could boost electricity bills for other customers. Nearly two-thirds of Americans are extremely or very concerned about the effect of data centers on energy prices, while 57% express the same level of concern about their impact on water supplies, according to an AP-NORC Center for Public Affairs Research and University of Chicago Energy Policy Institute poll.

Some lawmakers have called for broader restrictions on AI data center development, while others argue that expanding domestic computing and power capacity is important to U.S. technology development.

The post Google, NVIDIA Push Flexible AI Data Centers as Congress Targets Energy Costs appeared first on Techstrong IT.

The funding will support research and development at Anderon’s 300-millimeter quantum wafer foundry in Albany, New York. IBM said it is also investing another $1 billion in the foundry, which will manufacture quantum wafers for IBM and other quantum hardware developers, increasing domestic capacity to produce specialized quantum hardware.

When IBM and the Commerce Department disclosed the planned award, the $1 billion commitment was still subject to negotiation of a final agreement. The Commerce Department now describes the deal as a final R&D award of up to $1 billion under the CHIPS and Science Act.

Anderon also announced that the first quantum wafers are now moving through its manufacturing facility. The company said it offers specialized wafers for superconducting qubit arrays, quantum input/output signaling and components used in quantum readout systems. The firm plans to expand to other quantum modalities over time.

IBM describes Anderon as a pure-play quantum foundry, borrowing a model familiar from the conventional semiconductor industry in which a specialized manufacturer fabricates chips for outside customers. In Anderon’s case, the facility is dedicated to quantum wafers and offers established fabrication processes, process design kits, wafer testing and advanced packaging.

“This investment helps establish the manufacturing foundation needed to support the next generation of quantum computers, enabling the industry to move from scientific progress to production-scale innovation,” Jay Gambetta, director of IBM Research and an IBM Fellow, wrote in a social media post. He noted that scalable wafer fabrication will become a more important part of the technology stack as the industry works toward fault-tolerant quantum computing.

The Anderon award is the largest share of the roughly $2 billion federal quantum funding package first outlined by the Commerce Department in May. That package includes awards for other quantum companies including GlobalFoundries, D-Wave, Rigetti Computing, PsiQuantum and Quantinuum. Commerce finalized those five awards on Sept. 8.

With its first wafers now moving through the Albany facility, Anderon’s next challenge will be turning that initial production into repeatable manufacturing for outside quantum hardware companies. The company plans to expand beyond superconducting systems over time, which could make the facility useful to developers working with other qubit technologies.

The post IBM’s Anderon Begins Quantum Wafer Runs as $1B CHIPS Award Is Finalized appeared first on Techstrong IT.