Close Menu
  • AI
  • Content Creation
  • Tech
  • Robotics
AI-trends.todayAI-trends.today
  • AI
  • Content Creation
  • Tech
  • Robotics
Trending
  • Black Forest Labs Releases FLUX 3 Motion: A 7B Open-Weights World Motion Mannequin That Tops RoboLab-120
  • Fastino Releases GLiNER2.5-Resolve: A 340M Open-Weight Determination Mannequin That Runs on CPU
  • BottleCap AI Releases ThinkingCap-Qwen3.8-27B: 37.2% Fewer Considering Tokens at a 0.86pp Accuracy Price
  • What if I find an AI agent that is worth the risk?
  • Google’s Gemini Can Now Make Requires You on Pixel Telephones
  • An OpenAI Agent Hacked Australia’s Well being Service. Their Authorities Discovered Out Months Later
  • Vibe Coding for Inexperienced persons — Easy Information for Creators, Entrepreneurs, and Non-technical People
  • Contrastive-LM Releases CLM-8B: An Open System One Mannequin That Scores Agent Actions As much as 9× Quicker Than Jev
AI-trends.todayAI-trends.today
Home»AI»Stealth Startup Believes it Has Solved the Memory Crisis

Stealth Startup Believes it Has Solved the Memory Crisis

AI By Gavin Wallace09/09/20262 Mins Read
Facebook Twitter LinkedIn Email
GPT-4o Tells Jokes about AI • AI Blog
GPT-4o Tells Jokes about AI • AI Blog
Share
Facebook Twitter LinkedIn Email

Kepler says that it also has increased the density of SRAM by using ferroelectrics. read and write data at lower voltages It is more efficient than traditional mechanisms for processing and storing data. It achieved this feat by creating a composite low-voltage material that worked with the ferroelectric method.

SasiManipatruni is a Kepler Cofounder and Chief Technology Officer. He says that the Kepler Team went through 35 different composites to find a type of material they believed would be easier to manufacture and more cost-effective.

“Once we found a way to solve the physics problem—as in the physical limitations for HBM—we came up with a material innovation that helps with the amount of memory you can have between chips,” He says.

Kepler claims that in its initial build-outs, it was able convert a factory into a “next-generation” In just eight months you can achieve fab, as opposed to the usual 24-month period.

“Out goal is to take the fabs and architectures already built, and push them to the limits of physics,” Olaosebikan says.

Kepler bets that the additional cost of retooling or using new materials will far exceed the 20 billion to 40 billion dollars it would take to construct new fabs and equip them with hundreds of millions in equipment.

The Waiting Game

Kepler Computing still has a long road ahead before it reaches full-scale production—assuming it gets there. The company’s technology has been tested on roughly 2,000 wafers. The company says that it plans to ship the first HBM chip samples this year. Next year production will increase in Singapore, while production in America is expected to begin in 2028.

Kaste says that he is confident about the future of GlobalFoundries. “fundamental breakthroughs have happened. What remains is getting good results on thousands of wafers and millions of devices.”

He highlights that iron can be a problem with material systems like Kepler’s, because it is part of its composite. “Iron is a tough contaminant to introduce into a production facility. So Kepler’s solution has to run on dedicated equipment, or be fully encapsulated so that it can’t escape,” He says.

“The art is in keeping that material really well isolated through our production flow,” Kaste says.

Kepler’s dedicated testing fab uses automated wafer handling equipment.

Kepler Computing is the source of this photograph

AI artificial intelligence chips memory semiconductors startups
Share. Facebook Twitter LinkedIn Email
Avatar
Gavin Wallace

Related Posts

What if I find an AI agent that is worth the risk?

24/09/2026

Google’s Gemini Can Now Make Requires You on Pixel Telephones

24/09/2026

An OpenAI Agent Hacked Australia’s Well being Service. Their Authorities Discovered Out Months Later

24/09/2026

Meta VR Glasses, Ray-Ban Meta Audio, Ray-Ban Meta Gen 3: Specs, Options, Costs

24/09/2026
Top News

Intel takes major step in its plan to purchase chip startup SambaNova

OpenAI Confidentially files for IPO in the wake of SpaceX, Anthropic

OpenAI hires 4 high-ranking engineers from Tesla, xAI and Meta

Be a part of the WIRED World Truthful in Miami on November 4

‘Pretty Crazy’ Token Usage Is Testing Bosses’ Bet on AI

Load More
AI-Trends.Today

Your daily source of AI news and trends. Stay up to date with everything AI and automation!

X (Twitter) Instagram
Top Insights

Develop an End-to End Document Intelligence pipeline with docTR, including OCR, Layout Analysis and KIE. Benchmarking is also possible.

17/08/2026

Trump and Energy Industry are Eager to Use Fossil Energy for AI

16/07/2025
Latest News

Black Forest Labs Releases FLUX 3 Motion: A 7B Open-Weights World Motion Mannequin That Tops RoboLab-120

25/09/2026

Fastino Releases GLiNER2.5-Resolve: A 340M Open-Weight Determination Mannequin That Runs on CPU

25/09/2026
X (Twitter) Instagram
  • Privacy Policy
  • Contact Us
  • Terms and Conditions
© 2026 AI-Trends.Today

Type above and press Enter to search. Press Esc to cancel.