😺 🎙️ Watch: Is Brain-like Computing What's Next?

2026-04-23 · Source: The Neuron · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, long

Summary

Great Sky, co-founded by Jeff Shainline, is developing brain-inspired AI hardware to overcome the memory-processor bottleneck in conventional computing. Their Superconducting Optoelectronic Networks (SOENs) architecture integrates memory and processing, mimicking the brain's structure, and utilizes superconductors, photons, and analog computation. This approach aims to enable new AI models by moving information with light, contrasting with current GPU-heavy systems that face memory and energy constraints. Great Sky plans a Brookhaven fusion deployment later this year, specialist system availability in two to three years, and larger data-center deployments thereafter, with a 100M-parameter roadmap. This innovation challenges the "hardware lottery" by proposing a new compute substrate that could make previously impractical, brain-like AI architectures economically viable.

Key takeaway

For AI Hardware Engineers evaluating future compute architectures, you should investigate Great Sky's SOENs as a potential alternative to GPU-centric designs. This brain-inspired approach, integrating memory and processing with optical communication, could enable new AI model types and address current energy and memory bottlenecks. Consider how such a shift in compute substrate might make previously unviable recurrent or brain-like systems viable, impacting your long-term hardware strategy and model development.

Key insights

Brain-inspired hardware using light and integrated memory-compute can overcome traditional AI bottlenecks.

Principles

• AI architectures thrive on compatible hardware.
• Integrated memory-compute reduces bottlenecks.
• New substrates enable diverse AI models.

Method

Great Sky's SOENs architecture intertwines memory and processing using superconductors, photons, and analog computation to move information via light, bypassing the memory-processor bottleneck.

In practice

• Deploy specialist systems for specific workloads.
• Explore analog values for AI computation.
• Consider optical communication for data transfer.

Topics

• Brain-like Computing
• AI Hardware
• Superconducting Optoelectronic Networks
• Memory-Compute Integration
• Analog Computation
• Optical Communication
• Hardware Lottery

Best for: Research Scientist, Investor, AI Hardware Engineer, AI Scientist, Director of AI/ML

Related on AIssential

• Exploring the Convergence of Artificial Intelligence and Neuroscience — AI and neuroscience share a symbiotic relationship, driving mutual advancements in understanding cognition and developing intelligent systems.
• This Breakthrough Could Kill the GPU Race — Optical computing with active metasurfaces offers a path to dramatically reduce AI's energy footprint by performing computation directly with light.
• 2026.22: Luceing Their Mind — The future of AI compute will diverge, with agentic inference prioritizing memory capacity and cost over raw speed.
• Fixing AI’s Bottlenecks: Memory, Scale, and Sparsity — AI's future hinges on addressing memory, connectivity, and sparsity through engineered, biologically-inspired hardware and algorithms.
• The Sequence Opinion #819: How AI Chips are Made? — Hardware, not just algorithms, drives 60% of deep learning's current success.
• OpenAI GPT-5.3-Codex-Spark Shows What's Possible With Cerebras — OpenAI's new coding model, GPT-5.3-Codex-Spark, leverages Cerebras hardware, challenging Nvidia's dominance and expanding AI hardware options.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Neuron.