Powering the Future: Why Thermal Management Defines the Future of Electronics
Standard of Excellence: Engineering Is the New Sales—How Technical Collaboration Wins Business
SMT Perspectives & Prospects: Artificial Intelligence, Part 7—Data Module 2
Jon Peddie Research Releases New Photonic AI Processors Report
April 7, 2026 | Jon Peddie ResearchEstimated reading time: 1 minute
Jon Peddie Research (JPR), the leading research and consulting firm for graphics and AI technologies, has released its latest report on AI, the Photonic AI Processors report covering developments during the third quarter of 2025.
Data centers power artificial intelligence, and AI is breaking data centers. GPU clusters now consume hundreds of watts per chip, and scaling compute faster than efficiency has become unsustainable. Photonic AI processors address this at the physics level — replacing electrons with photons to perform AI matrix operations faster and at a fraction of the energy cost.
This report covers seven companies building photonic compute processors. Four of the seven companies are German, French, or British. The US contingent, Lightmatter and Neurophos, are the best-funded. Overall, the companies have received over $1.1 billion in funding.
The Photonic AI Processors report defines photonic AI processors and how they are used. It outlines two distinct product categories: photonic compute processors, where light performs arithmetic, and photonic interconnects, which move data between conventional chips and replace copper wires.
"In electronic AI processing, avoiding latency means staying in memory,” said Dr. Jon Peddie, president of Jon Peddie Research. “In photonic AI processing, avoiding latency means staying out of it."
While the companies covered in the main body of this report focus on classical photonic AI acceleration (matrix multiplication, inference, HPC), a distinct set of players is building photonic quantum computers that use single photons as qubits and linear optical elements to implement quantum gates and entanglement. These systems target quantum simulation, optimization, and cryptography workloads, not near-term neural network acceleration, and operate on much longer commercialization timelines. Nonetheless, they have attracted over $2.6 billion in investment.
“This is an exciting and opportune time for investors as well as public companies to get involved in this emerging market,” said Peddie. “It’s not yet gold rush days, and highly leverageable bets can still be made.”
Share on:
Testimonial
"In a year when every marketing dollar mattered, I chose to keep I-Connect007 in our 2025 plan. Their commitment to high-quality, insightful content aligns with Koh Young’s values and helps readers navigate a changing industry. "
Brent Fischthal - Koh YoungSuggested Items
Future-proof Laser Depaneling for PCBs with Photonics
04/10/2026 | Real Time with... APEX EXPOBill Solari of Photonics Systems USA outlines the advantages of laser depaneling for circuit board manufacturing. We learn that this advanced technology addresses rising material costs by increasing panel yield and improving edge quality. This process provides precision, suitability for medical applications, and cost-saving benefits, including the elimination of cleaning steps. There have been significant advancements in laser depaneling, offering a competitive and efficient solution for modern electronics production.
PIC & Mix: How Quantum Technologies are Shaking up the Photonic Integrated Circuit Market
04/08/2026 | IDTechExPhotonic integrated circuits (PICs) are optical systems fabricated on semiconductor wafers, allowing complex optical processes to be performed on a chip-scale device.
CEA-Leti, NcodiN Partner to Industrialize 300 mm Silicon Photonics
03/11/2026 | NcodiNCEA-Leti and NcodiN, a French deep-tech startup pioneering nanolaser-enabled photonic interconnects, announced a strategic collaboration to industrialize NcodiN’s optical interposer technology on a 300 mm integrated photonics process.
NVIDIA Architecture Drives Optical Interconnects; CPO in AI Data Centers Rising
03/11/2026 | TrendForceNVIDIA’s next-generation AI compute rack architecture indicates that future GPU designs will increasingly prioritize higher chip-to-chip interconnect density and faster data transmission, according to TrendForce’s latest research on the high-speed interconnect market.
STMicroelectronics Expands Silicon Photonics Production for AI Demand
03/09/2026 | STMicroelectronicsSTMicroelectronics, a global semiconductor leader serving customers across the spectrum of electronics applications, is now entering high-volume production for its state-of-the-art silicon photonics-based PIC100 platform used by hyperscalers for optical interconnect for data centers and AI clusters.