Stealth-mode startup Neurophos has secured $110 million in funding led by Mithril Capital to develop ultra-compact optical processors for AI inference. Founded by experts in optical cloaking and metamaterials, the company aims to create chips that dramatically accelerate AI while consuming far less power than traditional silicon. This breakthrough could enable complex AI models to run directly on edge devices like smartphones and autonomous vehicles.
In a significant development for emerging technologies, stealth-mode startup Neurophos has announced closing a massive $110 million funding round led by Mithril Capital. The investment will fuel development of a new generation of optical processors specifically designed for AI inference operations. This announcement comes amid intensifying global competition to develop specialized chips capable of running complex AI models more efficiently. What sets Neurophos apart is its unconventional origins—the company was founded by a team of scientists whose previous research focused on developing optical cloaking technologies and metamaterials.
The funding round, led by Mithril Capital with participation from other prominent investors, demonstrates significant confidence in Neurophos's technical vision. The company plans to use these substantial funds to accelerate research and development, build prototypes, expand its engineering team, and prepare for commercial production. The stated goal is to create tiny optical processors specifically engineered to accelerate inference operations performed by trained AI models.
Neurophos's approach centers on converting digital data into optical signals and processing them using miniature optical components based on metamaterials. Instead of relying on electron movement through traditional silicon transistors, these chips use photons (light particles) to perform computations. This fundamental shift from electricity to light promises enormous gains in speed and reduced power consumption, particularly for matrix multiplication tasks that form the core of neural network operations.
If Neurophos successfully delivers on its technical promises, it could revolutionize how AI is deployed. Currently, the AI industry faces a significant dilemma: the energy hunger and massive infrastructure required to run advanced models, even during inference alone. Specialized optical processors could solve this problem by enabling complex AI models to run directly on edge devices like smartphones, autonomous vehicles, smart cameras, and IoT devices without requiring constant cloud connectivity.
This would not only reduce operational and energy costs but also enhance privacy (by processing data locally) and decrease latency. Competition in this space is heating up, with major players like Lightmatter and Lightelligence already active. However, Neurophos's unique background in optical metamaterial engineering might give it an advantage in miniaturization and efficiency.
Optical processors are a type of processing unit that uses light (photons) instead of electrons to transmit data and perform computations. They work by switching light beams on and off through miniature optical components, allowing complex calculations—especially those related to AI—to be performed at tremendous speeds with significantly less power than traditional electronic chips.
Training: This is the expensive, energy-intensive phase where an AI model is "taught" using massive datasets. Inference: This is the phase where the trained model is used to make predictions or produce results based on new data (such as recognizing an image or translating a sentence). Neurophos's chips are specifically designed to accelerate the inference stage, which is the most frequently repeated phase when deploying models at scale.
Optical cloaking or "invisibility cloak" technologies rely on engineered metamaterials that can manipulate light in extraordinary ways. This expertise in controlling light at microscopic scales directly translates to creating efficient optical computing components. The team's experience with metamaterials—materials engineered to have properties not found in nature—provides them with unique insights into building compact, high-performance optical processors.
Optical computing offers several key advantages for AI applications:
While Neurophos hasn't announced specific timelines, the $110 million funding round suggests they're moving toward commercialization. Typically, hardware development cycles for new chip architectures take several years from prototype to market-ready products. The company will likely focus on creating proof-of-concept prototypes before scaling to production, with initial products potentially targeting specific edge computing applications where power efficiency is critical.
Neurophos's substantial funding and unique technical approach position it as a potentially disruptive player in the AI hardware landscape. As AI models grow increasingly complex and energy-intensive, the industry desperately needs more efficient computing solutions. Optical processors represent one of the most promising frontiers, and Neurophos's combination of metamaterials expertise and AI focus could yield breakthroughs that make powerful AI accessible on everyday devices. While challenges remain in manufacturing, scalability, and software integration, the potential rewards—from smarter smartphones to truly autonomous edge devices—make this technological pursuit worth watching closely.
Source: TechCrunch AI | Analysis & Editorial: AI Tools Oasis
Bringing you the latest news and analysis in the world of Artificial Intelligence with accuracy and credibility. Follow us for all updates.
OpenAI is advancing its ambitious super app project, aiming to integrate advanced AI capabilities into a single, multifunctional platform. This development is part of the company's strategy to expand services and deliver a unified user experience. Discover the full details and expected impact of this move.
Notion has restored access to its Anthropic AI integration after a 4-hour outage disrupted users relying on Claude-powered features. The incident highlights the growing dependency on AI productivity tools and raises questions about infrastructure stability. All user data remained secure during the disruption.
A new report from TechCrunch AI warns of a potential 'Tokenpocalypse'—a massive collapse of digital tokens due to oversupply. With over 80% of new tokens losing 90% of their value, the market faces a crisis reminiscent of the dot-com bubble. This analysis explores the risks, impacts, and how investors can protect themselves.
