Powerlattice Raises $86M to Cut AI Processor Energy Use
The startup is developing monolithic voltage regulator chiplets designed to reduce power waste by 50% in data centers.
Updated on Oct. 30, 2026 in Semiconductors
Live Poll
Do you believe hardware innovations will effectively solve the power constraints of artificial intelligence?
Powerlattice has raised $86 million in new capital to scale its power delivery chiplet technology. The company, which is seeking a total of $111 million, recently demonstrated a monolithic integrated voltage regulator chiplet that aims to improve power efficiency in AI processors.
Why it matters
AI processors generate heat levels that frequently overwhelm data center cooling and power infrastructure. Reducing energy waste at the chip level is becoming a critical requirement for maintaining high-performance compute density.
The firm's monolithic integrated voltage regulator chiplet combines the controller, power transistors, and inductors onto one piece of silicon. This approach is intended to reduce energy loss by shortening the current path to the processor.
The players
Powerlattice
A startup designing integrated voltage regulator chiplets to optimize power delivery in high-heat AI processor packages.
Peng Zou
A co-founder of Powerlattice who helped establish the firm in 2023.
Gang Ren
A co-founder of Powerlattice focused on developing chiplet-based power solutions.
Sujith Dermal
A co-founder of Powerlattice currently scaling the company's chiplet research and manufacturing presence.
The details
Powerlattice uses a monolithic design to shrink the physical distance electricity travels to reach a processor, which the company claims can reduce total energy consumption by 50%. A monolithic integrated voltage regulator — a component that maintains constant output voltage despite input changes — integrates multiple power-handling elements onto a single die. This design allows for higher power density compared to multi-chip power delivery systems that rely on separate discrete components.
Timeline
2023: Powerlattice was founded.
November 2025: The company emerged from stealth with $25 million in Series A funding.
Week of October 26, 2026: The company demonstrated its integrated voltage regulator chiplet.
October 30, 2026: SEC filing for the latest funding round was submitted.
The Tech Race
Powerlattice is competing to solve power density limits in AI silicon, an area where board-level solutions have struggled to keep pace with demand. The firm's shift toward on-package monolithic integration marks a direct challenge to traditional discrete power delivery architectures.
This technology targets data center operators facing cooling capacity limits, as the integrated chiplets are designed to sit inside processor packages. Adoption will depend on successful integration into high-performance silicon roadmaps, with timeline for commercial availability remaining unannounced.
The takeaway
As AI workloads grow, companies are increasingly moving power delivery components directly into processor silicon to minimize heat-related inefficiencies. Watch for announcements regarding manufacturing partnerships or product benchmarks that validate the claimed 50% energy reduction in production environments.
Further reading
For more on the hardware driving modern AI, visit our Semiconductors section.
Source note: This article includes information reported by GeekWire.
Live Poll
Do you believe hardware innovations will effectively solve the power constraints of artificial intelligence?
