Introduction: AI Infrastructure Is Becoming a Power-Engineering Challenge
The Open Compute Project (OCP) Global Summit 2026, held from 12–15 October at the San Jose Convention Center in San Jose, California, highlighted a fundamental shift in data-centre engineering & scaling artificial intelligence is no longer solely a challenge of compute performance. It is increasingly a question of how electrical power is converted, distributed, regulated, backed up and removed as heat across an increasingly dense infrastructure.
Dinesh Kithany, SME and Founder & Chief Analyst at Wired & Wireless Technology (WAWT), attended the summit to follow developments across power architecture, power conversion and the wider technology ecosystem.
The summit’s theme, Scaling Innovation for the AI Era, was reflected in its technical programme, which covered 22 tracks spanning AI infrastructure, servers, networking, storage, rack power, cooling and system validation.
For WAWT, the significance of OCP extends beyond server and networking innovation. It provides an opportunity to understand how changes in compute architecture are reshaping requirements for AC-DC conversion, DC-DC power supplies, rack-level distribution, backup power and emerging power-delivery technologies.
The Most Important Observation: Power Architecture Is Moving Up the System Hierarchy
The most consequential development at OCP 2026 was the growing alignment around high-voltage DC distribution, particularly 800 VDC, alongside rack architectures designed for power levels approaching or exceeding one megawatt.
This is not simply a transition from one bus voltage to another. It changes the engineering trade-offs across the complete power-delivery chain.
At a fixed power level, increasing distribution voltage reduces current. For example, delivering 1 MW at 800 V requires approximately 1,250 A, compared with 2,000 A at 500 V, assuming ideal DC conditions. At 800 V, the current is still substantial, but the reduction can materially influence conductor sizing, resistive losses, connector design and distribution-system packaging.
For a given conductor resistance, resistive losses scale with the square of current. Consequently, reducing current can improve distribution efficiency and reduce the thermal burden on conductors and interconnects.
However, the benefit cannot be evaluated in isolation. An 800 VDC architecture also introduces requirements for high-voltage isolation, fault interruption, arc management, grounding, service safety, insulation coordination and conversion-stage reliability. Protection strategies must address fault energy and interruption behaviour at the system level, rather than relying solely on conventional low-voltage rack practices.
OCP’s 2026 programme explicitly addressed these challenges through sessions on end-to-end 800 VDC performance, system safety, protection coordination, high-voltage grounding and power delivery to next-generation AI infrastructure.
Source: OCP Global Summit 2026 — Technical Tracks
WAWT perspective
The move towards 800 VDC creates an opportunity for suppliers across the power-supply value chain. The relevant market is not limited to high-power rectifiers. It includes intermediate bus converters, isolated DC-DC stages, high-current point-of-load regulators, protection components, capacitive energy storage, battery interfaces, monitoring electronics and power shelves.
It also changes how the merchant power-supply market should be assessed. Power ratings alone are no longer sufficient to compare suppliers. Voltage range, conversion topology, power density, transient response, fault behaviour, redundancy, thermal performance and compatibility with emerging open rack specifications will increasingly determine commercial relevance.
For power-supply manufacturers, the strategic question is whether their products can support the transition from conventional rack power to integrated, high-voltage, high-density AI infrastructure.
Top Highlights from OCP Global Summit 2026
1. The industry is developing an open framework for 800 VDC
OCP’s power-distribution work is bringing hyperscale operators and technology suppliers together around an open approach to high-voltage DC infrastructure. NVIDIA, Google and Microsoft have been associated with the broader effort to establish 800 VDC as a standardised architecture for next-generation AI data centres.
The technical scope includes power distribution, conversion requirements, system protection and delivery of energy to increasingly demanding computing platforms.
Source: OCP — Powering the Next Era of AI: Google, Microsoft and NVIDIA on LVDC
2. Rack-scale engineering is becoming a system-integration discipline
The summit programme covered power and thermal co-design, direct-to-server power, Open Rack Wide validation and rack architectures exceeding 1 MW.
This reflects a shift away from treating the server, power shelf, rack and cooling infrastructure as independent purchasing decisions. Their electrical, mechanical and thermal interfaces increasingly need to be designed and validated together.
Source: Explore 22 Tracks at the 2026 OCP Global Summit
3. Power and cooling are converging
Exhibitors demonstrated architectures combining high-voltage power delivery with liquid cooling and rack-level integration.
For example, LITEON presented 800 VDC power solutions, power shelves, battery backup integration and DC-DC conversion technologies, alongside a 2.6 MW cooling distribution unit from its strategic investment partner, DCX Liquid Cooling Systems.
These announcements illustrate the growing commercial overlap between power electronics, thermal management and infrastructure integration.
Source: LITEON — OCP Global Summit 2026
4. OCP is expanding beyond hardware specifications
The summit also featured OCP Nexus, an AI knowledge assistant designed to make OCP’s technical resources easier to explore. Separately, the Open Switch Platform Architecture (OSPA) initiative was announced as an in-progress contribution to the OCP Networking Project.
Together, these developments reflect a broader objective: making open infrastructure easier to develop, integrate and operate across different hardware and software ecosystems.
5. Solid-state transformers enter the Grid-to-Chip discussion
OCP’s Data Center Facility work is examining how solid-state transformers (SSTs) could connect medium-voltage infrastructure with low-voltage DC distribution. Compared with conventional line-frequency transformer paths, SST-based designs can combine power conversion, control and DC interfacing in a compact architecture. Their potential value is not simply a smaller transformer: it is the ability to coordinate the utility interface, DC bus, energy storage and dynamic AI loads. Trade-offs include converter complexity, protection coordination, fault behaviour, maintainability and the need for standardised interfaces and test profiles. Delta presented an SST-enabled microgrid concept integrating on-site generation, storage and intelligent controls; SolarEdge also demonstrated an SST cell concept for medium-voltage-to-800 VDC conversion. These are architecture developments, not evidence that one SST design has become the universal deployment model.
Sources: OCP Global Summit 2026 recorded sessions
Delta Grid-to-Chip announcement: Delta’s energy-saving Grid-to-Chip infrastructure
SolarEdge demonstration: SolarEdge showcases AI data-centre power technology
6. Megawatt-class racks make thermal design a first-order decision
The second major theme was the move from high-density racks measured in tens of kilowatts toward designs in the hundreds of kilowatts and, ultimately, megawatt-class systems. At these levels, cooling cannot be treated as an accessory added after compute and power are selected. Cold plates, coolant distribution units, pump redundancy, quick disconnects, leak detection, coolant chemistry, service procedures and heat rejection all affect availability and deployment speed. OCP sessions from Supermicro, NVIDIA, MEGMEET, Delta and Danfoss emphasised the need to validate the compute tray and power shelf as part of a common thermal envelope. The engineering bottleneck is increasingly repeatable deployment and commissioning at fleet scale, not simply proving that a liquid-cooling loop can remove heat in a controlled demonstration.
Sources: Supermicro OCP solutions
Danfoss scalable liquid-cooling session information OCP Global Summit 2026
7. Open Rack Wide and sidecar designs separate compute from conversion volume
Two rack-level directions are especially relevant. Open Rack Wide (ORW) uses a wider form factor to accommodate larger, denser AI platforms and their service and thermal requirements. A sidecar or power rack, meanwhile, can move bulk conversion and protection out of the IT rack, preserving space for accelerators while providing modular service interfaces. Sidecars can support brownfield upgrades by converting existing AC infrastructure locally to high-voltage DC, without requiring a complete facility rebuild. These are complementary choices rather than one mandatory topology: the right design depends on rack density, floor space, retrofit constraints, fault isolation, serviceability and the maturity of upstream electrical infrastructure.
Sources: Rittal — OCP Global Summit 2026 and Open Rack Wide
Schneider Electric — 800 VDC rack-level architecture principles White Paper 213
Sources:
Company News from OCP Global Summit 2026
1. Delta targets Grid-to-Chip infrastructure integration
Delta’s OCP 2026 announcement covered high-power conversion and cooling infrastructure, including a 3 MW GoCool liquid-to-liquid cooling system, a Project Deschutes open cooling distribution unit and an 800 VDC, 3.6 MW liquid-to-liquid cooling distribution unit.
The significance is the increasing integration of power and cooling design at the infrastructure level.
For AI data centres, power conversion efficiency is only one part of operating economics. The electrical losses generated by conversion stages contribute to heat that must be removed, while the power density of servers and racks places additional demands on cooling distribution and system reliability.
Integrated design can help address these constraints, but it also makes interface validation, maintenance access and redundancy planning more important.
2. Advanced Energy: bridging 50 V and 800 V architectures
Advanced Energy used OCP 2026 to show products spanning established 48/50 V infrastructure and the emerging 800 VDC direction. Its published portfolio included 48-to-12 V DC/DC converters, 48 V-input server power supplies and ORv3 power-supply and shelf products, alongside its stated roadmap for 800 VDC and ±400 VDC systems. This highlights the transition path: operators will not migrate every installed rack simultaneously, so suppliers able to serve both existing 48/50 V designs and newer architectures can address mixed-generation deployments.
3. OmniOn Power: megawatt sidecar and high-density power shelves
OmniOn Power announced an AI data-centre portfolio spanning LVDC and HVDC, including a 1.2 MW, 800 V power-sidecar rack with battery backup units, a 108 kW 800 V DC/DC power shelf, a 72 kW AC/DC shelf, bus converters and rack controllers. Its OCP presentation also focused on weighted efficiency across the operating range: AI loads fluctuate, so a power chain should not be judged only at its peak-efficiency point. This matters because conversion losses at common operating states affect both electricity consumption and the heat cooling systems must remove.
4. AcBel: 100 kW and 72 kW shelf platforms
AcBel announced high-density power solutions supporting ±400 V/800 V DC and 50 V DC architectures. Its OCP showcase included 100 kW and 72 kW power-shelf platforms, reflecting the need for modular conversion capacity that can scale with rack-level demand. Shelf rating alone is not enough: input architecture, efficiency across the load curve, redundancy, fault containment, telemetry and compatibility with the selected rack standard determine whether a product can be integrated into a deployable system.
5. Advanced Energy: bridging 50 V and 800 V architectures
Advanced Energy used OCP 2026 to show products spanning established 48/50 V infrastructure and the emerging 800 VDC direction. Its published portfolio included 48-to-12 V DC/DC converters, 48 V-input server power supplies and ORv3 power-supply and shelf products, alongside its stated roadmap for 800 VDC and ±400 VDC systems. This highlights the transition path: operators will not migrate every installed rack simultaneously, so suppliers able to serve both existing 48/50 V designs and newer architectures can address mixed-generation deployments.
6. Flex: integrated infrastructure and deployment predictability
Flex positioned its OCP participation around treating AI infrastructure as an integrated product rather than a collection of separately sourced components. Its event messaging emphasised scalable, resilient and predictable infrastructure delivery as power, cooling and capacity constraints become linked. For operators, the value proposition is coordinating design, manufacturing and deployment across a complex bill of materials. The practical test will be whether integration reduces commissioning time and interface risk while preserving the flexibility that open standards are intended to enable.
7. Vertiv: grid-to-rack design for 300 kW, 600 kW and megawatt systems
Vertiv’s OCP programme examined power paths for AI factories as rack densities move through 300 kW and 600 kW toward megawatt-class systems. Its session description highlighted 800 VDC fault behaviour, system-level protection and the need for standardised interfaces. A related DC Fault Manager concept separates rapid local solid-state breaker action from central fault analysis, safety interlocks, telemetry correlation and controlled bus re-energisation. High-voltage DC protection therefore requires coordinated sensing and recovery behaviour across the distribution system, not simply a higher-rated breaker.
8. Schneider Electric: sidecar architecture and operational readiness
Schneider Electric’s published 800 VDC guidance identifies rack-level power racks, often called sidecars, as a near-term route to higher-voltage distribution. Moving conversion away from the IT rack can release space for compute and reduce congestion around power feeds and connectors. The design must still coordinate grounding, protection, energy storage, service procedures and ecosystem readiness. Schneider’s OCP-related discussion of fault ride-through and load smoothing also underlines that power architecture has to accommodate dynamic AI loads and disturbances at the grid interface.
9. EATON: safety and serviceability for 800 VDC
Eaton’s OCP session, “Beyond Power Density: Building a Safe 800V DC Ecosystem,” focused on grounding and bonding, touch-safe design, maintenance procedures, arc-flash and energy-isolation considerations, and circuit-protection approaches for DC faults and stored energy. These topics are central to the move into higher-voltage distribution within the data hall. An efficient architecture is not deployment-ready unless technicians can isolate, inspect, replace and recommission equipment under defined procedures, with fault energy and residual stored energy appropriately controlled.
10. LITEON: power shelves, backup and cooling as one design and 800 VDC-to-GPU power-delivery
LITEON’s portfolio spans 800 VDC distribution, battery or capacitor backup, DC/DC conversion and rack cooling. Its announced examples included a 95 kW DC/DC power shelf paired with a 280 kW in-rack liquid-to-liquid CDU for an NVIDIA Vera Rubin NVL72-oriented solution, and a 2.6 MW enterprise CDU from its cooling partner. These are vendor-reported configurations rather than independent comparative test results, but they illustrate the direction of system integration: power delivery, ride-through and thermal management must be sized and validated together.
LITEON’s OCP 2026 announcement is particularly relevant to the merchant power-supply industry. The company presented an 800 VDC power shelf with battery backup integration, solutions aligned with OCP ORv3 architecture, and a conversion portfolio spanning 800 V power bricks, intermediate bus converter modules, 50 V power bricks and voltage regulator modules.
This portfolio illustrates the number of conversion stages and interfaces that can sit between the incoming electrical supply and the processor.
As distribution voltage increases, the design challenge is not eliminated; it shifts towards efficient conversion, insulation, transient control and delivery of high current at lower voltages near the computing load.
The 800 VDC distribution stage and the final low-voltage GPU supply perform different functions and should be evaluated separately.
Other News Related to the Event
1. Open Rack architectures are increasing demand for specialised power solutions
The 2026 technical programme covered Open Rack Wide, high-density GPU power, 800 VDC and direct-to-server power delivery. For power-supply vendors, open rack specifications can help define interfaces and support interoperability. They do not, however, remove the need for system-level qualification.
Power shelves, backup modules, connectors, busbars and conversion stages must still be evaluated for electrical compatibility, transient performance, thermal limits, fault isolation, serviceability and redundancy. The commercial implication is that suppliers capable of delivering validated modules and complete reference architectures may be better positioned than vendors competing on nominal wattage or conversion efficiency alone.
Source: OCP Global Summit 2026 — Rack & Power Tracks
2. Power management is extending towards the grid
OCP’s programme included grid-interconnection requirements, solid-state transformer integration, low-voltage DC frameworks and power-system safety. This expands the scope of the power-supply market beyond equipment installed inside a server or rack. Large AI deployments increasingly require coordination between the utility interface, facility distribution, energy storage, rack power and dynamic computing loads.
For suppliers, the opportunity may extend to monitoring, protection, energy storage interfaces and power-management systems alongside conventional AC-DC and DC-DC conversion.
Source: OCP Global Summit 2026 — Recorded Content
3. AMD’s Helion reference design points towards rack-scale standardisation
AMD introduced its Helion reference rack design for next-generation AI infrastructure, based on OCP ORv4 and DC-MHS standards. The design illustrates the direction of travel towards modular, interoperable rack architectures with options for air and direct liquid cooling. The wider implication is that rack-scale compute increasingly depends on coordinated decisions across processors, networking, cooling, power and system management.
Source: Tom’s Hardware — AMD Helion Reference Design
4. Open standards are becoming commercially significant
OCP’s work on modular hardware, power distribution, network platforms and system management reflects a growing need for interoperability across increasingly complex infrastructure.
For the power-electronics supply chain, open specifications can broaden the addressable market for compatible modules. At the same time, suppliers must demonstrate that their designs meet system requirements under realistic electrical, thermal and operational conditions.
Source: Open Compute Project — Official News
5. Arm-based CPUs and Google Axion: compute choices influence rack economics
The summit’s broader agentic-AI discussion also highlighted the role of CPU platforms alongside accelerators. Google Axion is Google Cloud’s Arm-based CPU family, while Arm’s AGI CPU is a separate platform aimed at agentic workloads. These are distinct products and should not be attributed to Flex. Their relevance to power architecture is that CPU performance per watt, memory capacity, server count and workload orchestration influence how much useful work a rack delivers within a fixed power envelope.
Sources: Arm — OCP Global Summit 2026 takeaways
Google Cloud — Axion processors Google Cloud Axion
OCP Architecture: From the Grid to the GPU
OCP’s emerging power architecture is best understood as a change in where conversion occurs and how power is distributed through the facility. The legacy model commonly brings AC through facility distribution and UPS equipment, then converts it again inside individual servers. The emerging AI-factory model moves more conversion upstream and distributes DC closer to the rack or row, using shared power shelves, sidecars, batteries and high-density DC distribution. The objective is to reduce redundant conversion stages, lower current for a given power level, use copper more efficiently and create modular interfaces that can be validated across vendors.
| Layer | Legacy approach | 2026 OCP direction |
| Facility distribution | 480 VAC, centralized UPS | MVAC-to-DC conversion or sidecar AC/DC conversion convergedigest |
| Rack power | Server PSUs in every node | Power shelf / power sidecar delivering DC to the rack news.futunn+1 |
| Distribution voltage | Mostly 48 VDC inside server, AC upstream | ±400 VDC bipolar or 800 VDC nominal distribution convergedigest |
| Backup | Central UPS | Rack-level batteries / ESS integrated with DC distribution opencompute+1 |
| Density target | Tens of kW per rack | 100 kW to 1 MW+ per AI rack opencompute+1 |
The Grid-to-Chip path
A future-oriented chain can include medium-voltage AC at the facility boundary; a transformer-rectifier or solid-state transformer stage; a facility or data-hall DC backbone; a sidecar or power rack with protection and energy storage; rack-level 800 VDC or bipolar ±400 VDC distribution; intermediate DC/DC conversion; and finally point-of-load regulation at the accelerator. The exact sequence will vary by deployment. In a retrofit, a sidecar can convert existing 480 VAC locally and avoid rebuilding the entire upstream electrical system. In a greenfield AI factory, medium-voltage conversion may feed a shared DC backbone directly. Neither route is universally superior: the decision depends on site constraints, utility connection, protection strategy, load profile, serviceability and the availability of qualified equipment.
The architecture also has to support dynamic loads. Training and inference can produce rapid, correlated changes in power demand across many accelerators. Energy storage, power smoothing, fault ride-through, control-loop response and telemetry therefore become architectural functions rather than optional accessories. The design must coordinate grid behaviour, conversion stages, bus protection and load transients without compromising uptime.
Wide-bandgap semiconductors are another enabling layer. Silicon carbide (SiC) can be attractive in higher-voltage, high-power conversion stages, while gallium nitride (GaN) can support high-frequency, compact conversion in suitable voltage and power ranges. Their use is topology- and application-dependent; material choice alone does not guarantee better system efficiency. Magnetics, switching frequency, insulation, thermal design, packaging, control and EMI performance determine the realised benefit.
Two deployment models are emerging
• Brownfield / retrofit: place a sidecar or power rack beside the IT rack to convert existing AC to ±400 VDC or 800 VDC locally, with modular protection, monitoring and optional battery integration.
• Greenfield AI factory: consider medium-voltage-to-DC conversion through transformer-rectifier or SST-based systems, feeding a common DC backbone for high-density compute zones.
For WAWT, the market opportunity spans multiple layers: front-end conversion, SSTs and transformer-rectifiers, power shelves, bus converters, rack DC/DC stages, backup energy, solid-state protection, busbars, connectors, sensors, power controllers and firmware. A vendor’s position should be assessed against the layer it addresses and the interfaces it supports, not simply its maximum wattage.
Sources: OCP Data Center Facility Power Distribution project
OCP Global Summit 2026 technical tracks OCP 2026 tracks
Schneider Electric — 800 VDC rack-level architecture principles White Paper 213
OCP Security & Safety Agenda: High Voltage and Trust Must Scale Together
OCP’s security and safety agenda addressed two connected but distinct risks. On the electrical side, 800 VDC and high-density energy storage require coordinated grounding and bonding, touch-safe interfaces, fault isolation, arc-energy management, stored-energy discharge and clearly defined service procedures. At the platform level, AI infrastructure needs stronger device identity, secure boot, firmware attestation and recovery mechanisms across heterogeneous components. The SCMP (Secure Control and Management Plane) workstream announced at the summit aims to make trust and management less vendor-specific across devices, racks and pods, building on related efforts such as Caliptra, DMTF SPDM and OCP S.A.F.E. Electrical safety protects people and equipment, while platform security helps ensure that components boot and operate from trusted firmware and configurations.
Source: SCMP workstream announcement
OCP S.A.F.E. gains momentum
OCP S.A.F.E. (Security Appraisal Framework and Enablement) is gaining relevance as the ecosystem looks for repeatable, independent review of device firmware and security-sensitive components. Its significance is greater when connected to system-level trust: a secure rack or pod cannot rely only on a single device passing a review if firmware identities, update policies, recovery behaviour and component provenance are not linked across the platform. SCMP’s reference to S.A.F.E. suggests a direction in which device-level appraisal can contribute to a wider trust chain. Adoption still depends on implementation details, coverage and buyer requirements, so S.A.F.E. should be viewed as an assurance mechanism rather than a blanket guarantee that a complete system is secure.
Source: IOActive — Key takeaways from the 2026 OCP APAC Summit
Creating OCP Chiplets Ecosystem
The Open Chiplet Economy (OCE) work at OCP targets interoperability across design, packaging, assembly, testing and supply-chain participants. Chiplets can let designers combine compute, I/O, memory and specialised functions using different process technologies, but die-to-die electrical connectivity is only one part of interoperability. Secure boot, component identity, package-level thermal behaviour, power delivery, test coverage and firmware coordination must also work across multi-vendor assemblies. Summit demonstrations covered system modelling, high-radix network-on-chip switching, die-to-die interconnect, memory bandwidth and advanced packaging. For AI infrastructure, chiplet modularity may improve design flexibility, but it shifts more responsibility to package, platform and ecosystem-level validation.
Sources: OCP Global Summit 2026 programme
Arm — OCP Global Summit 2026 takeaways Arm newsroom
OCP Roadmap: From 800 VDC to a Broader AI-Factory Platform
OCP’s roadmap is not a single voltage-change project. Its Data Center Facility power-distribution work is aimed at low-voltage DC architectures up to 1,500 VDC, while the 800 VDC ecosystem is the nearer-term focus for high-density AI. A possible longer-term move toward 1,500 VDC would depend on component availability, insulation and clearance requirements, fault interruption, standards, certification and service procedures. It should therefore be treated as a direction under consideration, not as a confirmed timetable for universal rack adoption.
In parallel, OCP’s broader Open Data Center Ecosystem vision connects facilities and grid interfaces with power distribution, energy storage, cooling, silicon, servers, networking and systems management. Roadmap progress will be measured through specification maturity, reference designs, validation methods and deployable supply chains—not just headline rack-power targets. Open Rack Wide, modular hardware, grid-resilience work, solid-state transformer specifications and secure management initiatives each address a different dependency in this transition.
For vendors, the likely sequence is uneven: existing 48/50 V and ORv3 systems will coexist with new sidecar and 800 VDC solutions; power shelves and DC/DC conversion will evolve alongside rack cooling; and grid-to-chip architectures will advance at different speeds depending on facility readiness and local regulation. WAWT’s view is that suppliers should plan for multiple architectures during the transition and demonstrate interoperable products, transparent efficiency curves, protection behaviour, telemetry and serviceability.
Sources: OCP Data Center Facility — Power Distribution
OCP Global Summit 2026 — official technical tracks OCP tracks
OCP event theme — Scaling Innovation for the AI Era OCP 2026 theme
WAWT’s Key Takeaways from OCP 2026
Our Subject Matter Expert Dinesh Kithany’s participation at OCP Global Summit 2026 provides a valuable perspective on the relationship between AI infrastructure growth and the evolving power-supply market. From WAWT’s standpoint, five implications deserve particular attention.
1. The addressable power-supply market is becoming more technically segmented.
The transition towards 800 VDC does not simply increase demand for higher-wattage supplies. It creates differentiated requirements across front-end conversion, intermediate bus conversion, rack power shelves, backup power and point-of-load regulation. Market analysis should therefore distinguish products by architecture, voltage class, application and deployment stage.
2. System integration is becoming a competitive differentiator.
Suppliers that can combine power conversion, protection, monitoring, redundancy and thermal integration may have a stronger proposition in high-density AI infrastructure. This does not eliminate the market for standalone merchant power supplies, but it increases the importance of interoperability and validated system performance.
3. Efficiency must be evaluated across the complete power chain.
A high-efficiency AC-DC stage does not guarantee an equally efficient rack. Intermediate conversion, voltage regulation, backup systems, distribution losses and cooling requirements all affect the total outcome. WAWT considers end-to-end power architecture an increasingly important lens for evaluating vendor positioning.
4. The wireless-power opportunity requires a separate market assessment.
OCP 2026 was primarily a wired-power event. Wireless charging and wireless energy transfer should not be conflated with high-voltage DC distribution. Their potential lies in applications where contactless energy delivery offers a practical advantage, with commercial viability determined by power level, range, efficiency, reliability and deployment economics.
5. Standards development can influence vendor opportunities before full market adoption.
Emerging specifications can shape product roadmaps, qualification requirements and investment decisions well before a technology becomes widespread. Tracking standardisation, reference designs, supplier announcements and deployment timelines is therefore essential to understanding where the next opportunities in power electronics may emerge.
For WAWT, the central conclusion is that AI infrastructure is reshaping the power-supply market at the architectural level. The next competitive phase will be defined not only by how much power a supply can deliver, but also by how efficiently, safely and reliably it integrates into a complete AI computing system.
About WAWT
Wired and Wireless Technology (WAWT) is a strategic technology analyst and consultancy firm specialising in the wireless power and power supply industries.
WAWT provides market intelligence, technology analysis, strategic consultancy and bespoke research covering wireless power technologies, as well as AC-DC and DC-DC power supplies, external adapters and chargers.
Through WAWT’S Power Supply Intelligence Service, WAWT tracks technology developments, application markets, vendors, power classes, regional trends, market shares, emerging technologies and long-term market opportunities.
WAWT’s analysts regularly engage with companies, technology developers, industry associations and key events across the global power-electronics ecosystem to understand not only what technologies are being demonstrated today, but where they are likely to create commercial opportunities tomorrow.
To discuss the latest market trends, technology developments and insights from OCP 2026, please contact our subject matter expert, Dinesh Kithany, Founder & Chief Analyst, Wired & Wireless Technology (WAWT), at analyst@wawt.tech.
Contact WAWT to discuss the full OCP 2026 event-based insight report and the latest wireless-power and power-supply market intelligence.