top of page

Decoding the Future of AI - Beyond Moore's Law

Peek into 2030: From thumbnail-sized powerhouses to materials that defy nature: Meet the startups defining the next era of artificial intelligence.

Decoding the Future of AI - Beyond Moore's Law

In reflecting upon the history of technology, one cannot help but draw parallels between the mainframe computers of the 1950s and today's cloud-based LLMs, such as ChatGPT and Brad. These behemoths, though transformative in their own right, have presented obstacles to scalability, with their voracious energy consumption, significant network latency, and substantial operational costs.

Fast forward to 2030, and if AI's growth mirrors Moore's Law, we can anticipate models that are not just 100x more potent, but dramatically more energy and bandwidth efficient. Achieving this necessitates a comprehensive overhaul: from the foundational design of the chips propelling these models, to a reexamination of the materials constituting these chips and the manner in which data courses through them.

There are handful of startups around the world that are focused on solving this exact problem. Let's look at a couple of such groundbreaking startups Rain Neuromorphics and Neurophos (both are my portfolio companies),

Rain Neuromorphics:
-------------------------------
Pushing towards a future where a chip as compact as a thumbnail can accommodate 100 billion-parameter models. Rain Neuromorphics, uses an astute blend of algorithm and hardware, leveraging neuromorphic chips which emulate the human brain's intricate efficiency to achieve mangnitude more higher efficiency. Their gen 1 chip promises an astounding 100x energy efficiency leap beyond the benchmarks set by industry titans like Nvidia, AMD, and Intel.

https://rain.ai/

Neurophos:
-----------------
Aims to revolutionize AI processing by harnessing the prowess of optical AI processors. Their chips aim to outpace today's leading AI processors by a factor of 100 in both speed and energy efficacy. The secret? A seamless fusion of optical computing's efficiency with avant-garde metamaterials research. Metamaterials are new type of engineered materials that manipulate energy forms, including light, on an almost infinitesimal scale, boasting properties seldom seen in naturally occuring materials.

https://www.neurophos.com/

By pushing the boundaries of applied science, we stand on the cusp of an AI epoch that promises not just intelligence, but unmatched efficiency and scale.

Lead! Don't Follow!

A meditator's perspective on AI - insights for the Modern Leader!  

Signup and Receive the latest updates straight to your inbox.

Thanks for subscribing!

bottom of page