By Technology and Industry Desk
Published: October 2026
Main Facts
In a strategic move designed to address the colossal power demands of modern artificial intelligence infrastructure, Microchip Technology and Navitas Semiconductor have announced a collaborative partnership. The alliance has yielded a cutting-edge 800V DC-to-6V DC reference design tailored specifically for AI data centre rack power applications.
This comprehensive hardware and software platform bridges the gap between high-voltage power distribution and the ultra-low-voltage, high-current requirements of modern high-performance graphics processing units (GPUs) and AI accelerators. By combining Microchip’s advanced digital power control and hardware-based security ICs with Navitas’ high-frequency GaNFast gallium nitride (GaN) power devices, the design achieves unprecedented levels of thermal efficiency and power density.
Crucially, this engineering blueprint aligns directly with the emerging Open Compute Project (OCP) 800V DC standard. As data centre operators race to re-architect their facilities to handle dense, power-hungry GPU clusters, this turnkey reference platform—complete with reference hardware, software packages, and extensive design documentation—seeks to slash implementation risks and accelerate time-to-market for power supply unit (PSU) developers.
Chronology and Industry Context
The rapid evolution of generative AI and large language models (LLMs) has sparked a seismic shift in data centre architecture. For decades, standard server racks operated on 12V or 48V direct current distribution architectures. However, as individual server chassis begin consuming tens of kilowatts—and entire racks push well past the 100 kW threshold—traditional lower-voltage distribution networks suffer from severe copper losses, bulkier cabling, and thermal bottlenecks.
The industry-wide pivot toward an 800V DC backbone has accelerated rapidly through the mid-2020s.
- Early 2024–2025: Major hyper-scalers and semiconductor titans began drafting frameworks to transition internal data centre grids to higher voltages to minimize distribution losses.
- Mid-2026: Semiconductor giants began aligning their roadmaps with AI infrastructure needs. Competitors such as Texas Instruments (TI) and Infineon (in collaboration with Nvidia on MGX architectures) introduced respective 800V DC solutions.
- Late 2026: Microchip and Navitas cemented their collaborative response with the release of their 800V-to-6V reference design, bringing a distinct combination of digital signal control, robust hardware security, and wide-bandgap semiconductor advantages to the OCP ecosystem.
Supporting Data and Technical Specifications
The collaborative reference design is not merely a conceptual blueprint; it is a fully realized, highly optimized power distribution board (PDB) built upon elite silicon and wide-bandgap components.
1. Digital Control and Security Foundation
At the heart of the Microchip side of the architecture are the dsPIC33AK Digital Signal Controllers (DSCs) and the TA100 CryptoAuthentication security IC.
- dsPIC33AK256MPS306 Family: Powered by a robust 200 MHz 32-bit core featuring a double-precision floating-point unit (FPU), these controllers offer deterministic digital power control essential for high-frequency, high-efficiency DC/DC conversion stages. They incorporate 78 ps high-resolution Pulse Width Modulators (PWMs) and multiple 12-bit Analog-to-Digital Converters (ADCs) operating at up to 40 MSPS. Additionally, they feature native library support for Commercial National Security Algorithm (CNSA) Suite 2.0 recommended post-quantum cryptographic algorithms.
- TA100 Security IC: Provides a definitive hardware root of trust, enabling secure boot, cryptographic authentication, and protected firmware updates—vital features for safeguarding mission-critical cloud and enterprise AI servers against sophisticated cyber threats.
2. Power Conversion and GaN Technology
On the power conversion front, Navitas brings its state-of-the-art wide-bandgap semiconductor technology into play.
- GaNFast FETs: The power distribution board utilizes 16 × NV6034 devices—650V, 17 mΩ GaNFast FETs—configured in a stacked half-bridge topology on the primary side.
- Advanced Packaging: Housed in DFN8×8 dual-side-cooled packages, these transistors drastically reduce thermal resistance. This thermal efficiency enables continuous high-power operation without sacrificing conversion ratios.
- Performance Metrics: The resulting PDB achieves up to 96% peak efficiency at full load while operating at a 1 MHz switching frequency. This translates to an astonishing power density of 2,100 W/in³, saving invaluable rack space while keeping thermal management footprints manageable.
Official Responses and Industry Commentary
Industry leaders from both partnering organizations emphasized the strategic importance of this collaboration in meeting the demands of next-generation computing infrastructure.
"AI infrastructure optimisation is driving one of the most significant power architecture transitions the datacentre industry has experienced in decades," stated Joe Thomsen, Vice President of Microchip’s digital signal controller business unit.
He added: "As the ecosystem moves toward higher-voltage rack power systems, developers need proven control and security to help reduce implementation risk. Our collaboration with Navitas combines digital control, hardware-based security and advanced GaN power conversion to help customers bring 800V rack power systems to market more quickly."
Echoing these sentiments, Vipin Bothra, Vice President of Global Solution Marketing at Navitas Semiconductor, underscored the unique material advantages that gallium nitride brings to the table:
"As AI infrastructure scales to support increasingly demanding computing platforms, Navitas’ GaNFast technology is a critical enabler of higher power density, greater efficiency and improved system performance. By combining Navitas’ leadership in power semiconductors with Microchip’s digital control expertise, this collaboration accelerates the delivery of advanced power solutions tailored to the evolving requirements of next-generation AI datacentres."
Implications for the Data Centre and Semiconductor Industry
The release of the Microchip-Navitas 800V-to-6V reference design carries profound implications across multiple technology sectors:
1. Accelerated Transition to 800V Architecture
By providing a pre-tested, standards-compliant reference design adhering to Open Compute Project (OCP) guidelines, the partnership lowers the barrier to entry for power supply manufacturers. Companies looking to pivot to 800V DC infrastructure no longer have to design control algorithms, security layers, and high-frequency magnetics from scratch. This is expected to shorten design cycles significantly.
2. Validation of Gallium Nitride (GaN) in High-Power Industrial Applications
While silicon-based MOSFETs and silicon carbide (SiC) have historically dominated high-voltage applications, the successful integration of 650V GaNFast FETs in a stacked half-bridge topology demonstrates GaN’s viability in high-density, multi-kilowatt enterprise environments. Achieving a 1 MHz switching frequency at 96% efficiency proves that wide-bandgap semiconductors are ready to scale up to the rigorous demands of modern AI clusters.
3. Heightened Focus on Supply Chain Security
Integrating hardware-based roots of trust via components like Microchip’s TA100 directly into the power delivery architecture highlights a growing industry trend: security is no longer restricted to software or applications processors. As data centres become prime targets for state-sponsored and criminal cyber-espionage, securing the physical power management layer prevents malicious firmware tampering at the bare-metal level.
4. Intense Market Competition
As data centre operators seek every possible efficiency gain to mitigate escalating electricity costs and grid limitations, the race among semiconductor vendors—including Microchip, Navitas, Texas Instruments, and Infineon—to capture the 800V DC power market will only intensify. For power electronics engineers, this healthy competition yields increasingly sophisticated tools, richer reference libraries, and faster, more reliable hardware platforms designed to sustain the AI revolution.
