This forum focuses on the innovations already reshaping AI infrastructure – and the transition from centralized AI training to distributed AI deployment.
Most AI infrastructure discussions focus on scaling what exists today.
This forum focuses on the technological transitions that will define what comes next.
As AI expands beyond centralized training into distributed deployment, many of the assumptions that shaped today’s infrastructure are being challenged by fundamentally new approaches to power, compute, communications, cooling, and infrastructure economics.
AI Infrastructure Forum 2026 brings together engineers, operators, investors, technology providers, and standards leaders to examine which innovations are ready for deployment now – and which are likely to become tomorrow’s standard practice.
Seats are reviewed to ensure balanced representation across power, compute, fiber, cooling, capital, development, engineering, and standards leadership.
“The changes reshaping AI infrastructure are not incremental improvements to established models. The technological transitions underway today are foundational. They do not optimize the status quo – they render many of its underlying assumptions obsolete.
Those who understand the magnitude, inevitability, and urgency of that change will help define what comes next. Those who don’t will be left trying to keep pace as the industry shifts from maximizing capability to maximizing capability economically at scale.”
– AI Infrastructure Forum 2026
Most conferences explain what is happening.
This forum examines the forces driving what happens next.
As AI expands beyond centralized training into distributed deployment, many of the assumptions that shaped today’s infrastructure are being challenged by fundamentally new approaches to power, compute, communications, cooling, security, and infrastructure economics.
Developed in collaboration with IEEE Power & Energy Society, AI Infrastructure Forum 2026 brings together engineers, operators, utilities, developers, investors, technology providers, and standards leaders to examine which innovations are ready for deployment now – and which are likely to become tomorrow’s standard practice.
The forum is organized around four integrated pillars:
Because AI infrastructure no longer operates in silos. Constraints in one layer increasingly reshape every other layer.
Built for serious conversations, not expo-floor noise.
AI infrastructure is not one market: it is an ecosystem.
The challenges shaping AI deployment span power, compute, cooling, communications, capital, development, engineering, and public policy. No single organization sees the entire picture.
To keep the discussion valuable, in-person attendance is intentionally limited and reviewed by category. The goal is not maximum volume. It is the right mix of investors, utilities, developers, technology providers, engineers, standards leaders, and infrastructure innovators.
The value of the room comes from cross-disciplinary discussion. The people financing, building, powering, cooling, connecting, regulating, and operating AI infrastructure rarely share the same stage – yet their decisions increasingly shape one another’s success.
Built for serious conversations – not conference volume.
A curated mix of decision makers responsible for financing, building, powering, cooling, connecting, regulating, and operating AI infrastructure.
AI Infrastructure Forum is designed to connect organizations working across the infrastructure stack – from utilities and developers to investors, engineers, communications providers, technology leaders, and standards contributors.
Attendees include:
The value of the room comes from the interaction between these groups.
The next generation of AI infrastructure is redefining how compute is configured, powered, delivered, and economically deployed at scale.
AI infrastructure is expanding beyond a single deployment model.
As new approaches to training, inference, private AI, and edge deployment emerge, the assumptions that shaped previous generations of infrastructure are increasingly being challenged.
This forum examines the innovations, technologies, and emerging standards helping define what comes next.
Technologies that reshape industries succeed not because they are new, but because they prove reliable, repeatable, interoperable, and economically viable.
Emerging infrastructure ideas become industry standards only after they are challenged, validated, and refined by the engineering community.
Developed in collaboration with IEEE Power & Energy Society, AI Infrastructure Forum 2026 brings together engineers, operators, utilities, technology providers, and industry leaders to examine the innovations, architectures, and deployment models defining the next phase of AI infrastructure.
This is where innovation meets engineering credibility.
A curated executive forum examining the technologies, constraints, and decisions shaping the future of AI infrastructure.
The organizations building AI infrastructure no longer operate independently. Decisions around power, compute, communications, security, capital, and policy increasingly influence one another.
The Executive Forum brings together developers, utilities, investors, operators, technology providers, and public-sector leaders for a series of discussions focused on the realities shaping infrastructure deployment at scale.
Designed around conversation rather than presentation, the program explores the market forces, emerging technologies, infrastructure constraints, and strategic decisions likely to define the next phase of AI infrastructure development.
Morning check-in with coffee and networking.
Opening remarks and event introduction.
An executive overview examining the market forces accelerating AI infrastructure investment and the engineering realities beginning to reshape deployment strategies.
A forward-looking discussion exploring the technologies, infrastructure trends, and strategic decisions likely to influence AI deployment over the coming decade.
How utilities, power providers, compute platforms, and infrastructure suppliers are responding to AI-scale demand, transmission constraints, and the growing need for coordinated infrastructure planning.
Coffee, refreshments, and sponsor showcase access.
How organizations are planning, deploying, and operating AI infrastructure today – including enterprise adoption, operational priorities, governance, and organizational readiness.
Catered lunch with sponsor showcase and networking opportunities.
An executive perspective on one of the major strategic forces reshaping AI infrastructure, deployment, or investment.
How AI architectures, accelerated compute, power delivery, cooling technologies, and communications infrastructure are evolving as AI expands from centralized training campuses toward increasingly distributed inference deployments.
Refreshments and sponsor showcase access.
How investment priorities, infrastructure execution, deployment strategy, and operational resilience are shaping the next phase of AI infrastructure expansion.
A moderated wrap-up highlighting recurring themes from the day's discussions, followed by audience questions and open exchange.
Closing reception with sponsors, speakers, and attendees.
Four technical tracks examining the technologies, architectures, and engineering challenges defining next-generation AI infrastructure.
Developed in collaboration with IEEE Power & Energy Society, the Technical Program explores the practical realities behind next-generation AI infrastructure.
Across four focused tracks, engineers, operators, technology providers, and industry specialists examine the innovations, deployment models, and emerging standards influencing power delivery, compute infrastructure, communications networks, and security.
Designed for licensed engineers and technical professionals, the program combines practical engineering discussion with continuing education opportunities. Attendees may earn Professional Development Hours (PDHs) while gaining insight into the technologies and infrastructure models helping define the next phase of AI deployment.
Four sequential sessions for engineers and technical professionals. Attendees may register for up to 4 sessions. PDH-eligible.
Battery technologies, storage integration, grid interaction, and operational considerations for AI-scale facilities.
Power conversion architectures, high-frequency switching, DC bus topology, and conversion efficiency for AI infrastructure.
Break for lunch and networking between sessions.
DC mesh topology, bus bar design, current sharing, thermal performance, and real-time distribution monitoring for high-density deployments.
MEMS-based sensing and actuation, energy harvesting, environmental monitoring, and smart grid applications.
Four sessions. Four engineers working on the infrastructure problems that matter. Each session is a focused, PDH-eligible technical program built around engineering substance — not case studies or vendor pitches.
Dr. Petrovic brings over 30 years of industry and research experience spanning photovoltaics, batteries, fuel cells, hydrogen, and energy storage integration. He has authored 8 books, holds 50+ patents, and published more than 120 papers in the field.
His industry career includes roles as CTO of a battery company, VP of Engineering for a stationary fuel cell company, and Director of Systems Integration for a portable fuel cell company. He is also Founder and President of Solar Hope, which has completed 120+ solar PV installations across Africa, reaching over one million people.
Selecting the right energy storage architecture requires more than comparing technologies — it requires understanding how each performs under AI-scale operating conditions.
AI infrastructure is placing unprecedented demands on electrical systems, making energy storage a critical design consideration rather than simply a backup resource. Selecting the right solution requires balancing performance, response time, safety, lifecycle cost, operational complexity, and long-term reliability.
Rather than promoting a single approach, this session provides an engineering-focused comparison of today's major energy storage technologies — including lithium-ion batteries, compressed air, hydrogen, fuel cells, supercapacitors, flywheels, thermal storage, pumped hydro, SMES, and emerging hybrid systems — examining where each performs well, where its limitations emerge, and how different architectures align with evolving AI infrastructure requirements.
Whether designing new facilities or evaluating next-generation architectures, participants will gain a practical framework for comparing technologies, understanding their engineering tradeoffs, and identifying where the industry is likely headed over the next decade.
Joseph Jacob founded Blueflux Power to address one of the most pressing bottlenecks in data center development: power infrastructure that takes years to procure and months to commission. With over 10 years leading the company, he has built a practice centered on delivering complete, engineered power systems — transformers, medium voltage switchgear, low voltage distribution, and integration services — for high-density, mission-critical facilities on schedules the grid can't match.
His engineering background spans power system design, protection coordination, microgrid architecture, and factory acceptance testing. Blueflux Power is a WBE/MBE-certified company serving hyperscale and colocation data centers, microgrid developers, utilities, and EPC contractors nationwide.
The move to high-frequency DC conversion isn't just an efficiency story — it's a fundamental rethinking of how power is conditioned, distributed, and controlled inside modern data center infrastructure.
Conventional AC distribution architectures were designed for a different era of compute density. As AI workloads push power demand per rack into ranges that legacy systems weren't built to handle, high-frequency DC power conversion is emerging as a critical enabling technology — reducing conversion stages, shrinking passive component sizes, and unlocking distribution architectures that simply aren't possible at line frequency.
This session provides a practical engineering overview of high-frequency DC power systems: how switching frequency affects transformer size, filter design, and conversion efficiency; the tradeoffs between topologies at different power levels; and how these systems integrate with DC bus architectures, on-site generation, and storage. The discussion is grounded in real infrastructure requirements, with the goal of giving attendees a working framework for evaluating high-frequency conversion as a design choice rather than a vendor specification.
Joseph's engineering work spans power distribution design, microgrid architecture, hybrid energy systems, and control-system implementation for mission-critical infrastructure. He works across the full energy stack, connecting generation, storage, distribution, and critical loads through practical system design, real-time monitoring, protection coordination, commissioning, and autonomous power management.
His control-system work bridges electrical infrastructure and intelligent operation, using sensors, metering, communication protocols, and automation logic to monitor system behavior and coordinate power assets in real time. This includes Modbus/RS485 instrumentation, IPMI-based power metering, telemetry-driven decision logic, generator and battery coordination, DC distribution monitoring, and load-side control. His focus is on designing resilient power systems that improve efficiency, uptime, safety, and operational visibility across modern energy and data-center environments.
Most engineers treat the bus bar as a solved problem — until they are responsible for one at the scale where the physics actually bites back.
The shift to DC-native data center architectures has changed what power distribution infrastructure needs to deliver. Higher power densities, tighter reliability requirements, and the operational demands of AI-scale facilities have elevated the bus bar from a background component to a major design decision with real consequences for efficiency, uptime, thermal performance, serviceability, and total cost of ownership.
This session examines DC bus bar mesh architecture as a power infrastructure choice — how different distribution topologies perform under the load profiles of modern data centers, where conventional approaches create operational and cost constraints, and what a well-designed DC mesh can deliver in terms of flexibility, resilience, current sharing, maintainability, and long-term scalability.
The session also explores how real-time monitoring and closed-loop control can improve the value of DC distribution systems. At high-current densities, efficiency and uptime are not determined by copper alone; they depend on visibility into load behavior, voltage drop, thermal rise, imbalance, fault conditions, and asset utilization. By combining DC mesh architecture with sensor telemetry, power metering, control logic, and automated decision-making, operators can move from passive distribution to intelligent power management.
The focus throughout is on making better infrastructure decisions — not just better engineering calculations — by connecting electrical design, monitoring strategy, controls, and operational reliability.
Dr. Petrovic's research spans MEMS sensing, photovoltaics, hydrogen production, batteries, and low-cost energy technologies. He holds 50+ patents and has authored 8 books across his 30-year career in science and engineering.
Prior to academia, he served as Director at a semiconductor company and in senior technical leadership at multiple energy technology companies — giving him a rare combination of deep research expertise and real-world deployment experience.
Most data center engineers know sensors matter — few have a working model of what MEMS can actually do at the component level, and where that capability is headed.
As AI infrastructure scales, the demands on sensing, monitoring, and physical control systems are growing faster than conventional approaches can keep pace. MEMS technology — already embedded in consumer electronics, aerospace, and medical devices — is now finding serious traction in high-density compute environments, where precision, response time, and miniaturization create real operational advantages.
This session provides a practical engineering overview of MEMS technology across its major domains: optical, mechanical, RF, electrical, magnetic, and bio MEMS. Coverage includes design principles, structural analysis, packaging, reliability, and the path from lab to commercial deployment. The discussion then turns to data center applications — where MEMS sensors and actuators outperform conventional approaches, what integration looks like in practice, and how the technology roadmap aligns with the escalating demands of AI-scale facilities.
Visibility matters. Relevance matters more.
AI Infrastructure Forum 2026 is intentionally designed around quality of interaction rather than audience volume.
In-person attendance is curated and balanced across utilities, infrastructure developers, investors, operators, technology providers, engineering leadership, and standards organizations.
Sponsors are not simply promoting products or services. They are positioning their organizations alongside the technologies, ideas, and industry discussions helping shape future infrastructure decisions.
For organizations seeking meaningful engagement with decision makers, technical leaders, and strategic stakeholders, fewer conversations often matter more than larger crowds.
Limited to one sponsor per experience.
The event's largest shared networking moment – providing direct visibility across the full in-person audience.
Limited to one sponsor per experience.
Where the day's highest-value conversations continue after formal sessions conclude.
Sponsors receive early category visibility, curated engagement opportunities, and direct access to a concentrated audience of infrastructure decision makers.
Early sponsors receive category priority, expanded visibility, speaker consideration, and custom activation opportunities.
The full prospectus includes tier benefits, speaking opportunities, pass allocations, streaming visibility, sponsor showcase details, and category-priority options.
Developed by Blueflux Power in technical collaboration with IEEE Power & Energy Society – Atlanta Chapter.
Blueflux Power develops and deploys modular, high-density power systems designed to meet the evolving demands of AI infrastructure. The company focuses on integrated generation, storage, distribution, and control architectures that enable faster deployment, improved efficiency, and greater operational flexibility across both grid-connected and off-grid environments.
With a multidisciplinary engineering team and a focus on real-world deployment, Blueflux is advancing new approaches to power delivery that address the increasing scale, volatility, and performance requirements of AI-driven compute environments. Its work spans next-generation DC architectures, hybrid energy systems, and intelligent control platforms designed to optimize power at both the system and rack level.
Blueflux is actively engaged in industry collaboration and standards development, including contributions within the IEEE ecosystem, helping shape how future AI infrastructure is designed, powered, and operated.
AI Infrastructure Forum 2026 is presented in technical cooperation with the IEEE Power & Energy Society (PES) Atlanta Chapter, one of the leading professional organizations advancing the science and practice of electric power engineering.
IEEE PES connects a global community of engineers, researchers, utilities, and industry leaders working across generation, transmission, distribution, and emerging energy technologies. Through its standards, publications, and technical programs, PES plays a central role in defining how modern power systems are designed, operated, and integrated with new demands such as large-scale data center infrastructure.
The Atlanta Chapter brings regional expertise and access to a strong network of technical professionals across the Southeast, ensuring the forum is grounded in real engineering challenges and aligned with the latest developments in power systems and infrastructure design.
This is what makes the room structurally different from other events: executives, investors, utilities, and developers sit alongside engineers, architects, and IEEE standards contributors — in the same building, on the same day, working through the same infrastructure realities.
Executive discussions and technical sessions designed for the people building, operating, and enabling AI infrastructure.
Whether your focus is infrastructure strategy, technical implementation, or both, AI Infrastructure Forum 2026 offers multiple ways to engage with the conversations shaping the next phase of AI deployment.
Attendance is intentionally shaped to create a balanced room across power, compute, fiber, cooling, development, EPC, finance, government, standards, and innovation leadership.
Choose from the Executive Forum, the IEEE Technical Program, or a combined experience that provides access to both perspectives. Registration options and participation details are outlined below.
Curated in-person attendance
IEEE members receive preferred pricing and access to accredited breakout sessions (PDH eligible).
Four PDH-eligible technical sessions
Developed in collaboration with IEEE Power & Energy Society. PDH eligibility subject to IEEE PES confirmation.
Hosted at Georgia Tech’s Global Learning Center in the heart of Atlanta’s innovation district.
AI Infrastructure Forum 2026 is hosted at Georgia Tech’s Global Learning Center, a purpose-built executive education and conference facility located in Midtown Atlanta.
Surrounded by one of the nation’s leading engineering and technology ecosystems, the venue provides an environment designed for learning, discussion, and meaningful professional engagement.
Attendees are within walking distance of hotels, restaurants, and Atlanta’s growing innovation corridor.
Georgia Tech Hotel & Conference Center · Renaissance Atlanta Midtown Hotel
Hartsfield–Jackson Atlanta International (ATL)