Can Google and Meta Execs Fix AI Data Centre Memory Limits?

Three former Google and Meta executives have joined forces to solve one of the biggest challenges facing AI infrastructure: the memory wall. This limits performance, scalability and efficiency in modern data centres.

Founded by Ofer Shacham, Sha Rabii and Masumi Reynders, Majestic Labs has raised US$100m to develop breakthrough memory systems designed to dramatically reduce the cost and environmental impact of AI data centres.

Tackling the AI memory wall

“Majestic is built on a simple and powerful insight: AI’s next leap forward will come from access to more powerful AI infrastructure and more powerful AI infrastructure requires a reimagination of the memory system,” says Ofer Shacham, Co-founder and CEO.

“Majestic servers will have all the compute of state-of-the-art GPU/TPU-based systems coupled with 1000x the memory.

“Our breakthrough technology packs the memory capacity and bandwidth of 10 racks of today’s most advanced servers into a single server, providing our customers with unprecedented gains in performance and efficiency while slashing power consumption.”

Customers are expected to gain access to prototype Majestic servers by 2027, with the company positioning its technology as a transformative step forward for data centre design and AI scalability.

AI systems rely on vast amounts of data being moved between processors and memory blocks. While processors continue to advance rapidly, memory technology has lagged behind, creating a bottleneck known as the memory wall. 

As processor speeds increase faster than memory bandwidth, processors spend more time waiting for data, slowing performance and increasing energy demand across large-scale AI workloads.

Solving a fundamental infrastructure challenge

Majestic Labs aims to eliminate this bottleneck through an innovative architecture that allows the memory of up to ten racks to be integrated into a single high-density server. 

This approach could deliver as much as a 50-fold performance gain over current systems, significantly improving compute efficiency for training and inference workloads.

“AI infrastructure is scaling at unprecedented speed, but the industry has not solved key fundamental architectural inefficiencies,” says Masumi Reynders, Co-founder and COO.

“Majestic addresses this by delivering immediate operational gains on today’s workloads while maintaining full programmability and flexibility to adapt as AI evolves beyond transformer-based models.”

Masumi adds that reducing the cost and footprint of AI data centres will help make advanced compute capacity more widely accessible, particularly in regions where infrastructure investment has lagged. The company’s architecture is designed to support existing software frameworks while introducing a step change in performance efficiency and sustainability.

Advancing scalability and sustainability in AI

Majestic Labs’ co-founders describe their mission as one that combines innovation in system architecture with environmental responsibility. 

By consolidating massive memory resources within fewer physical servers, the company expects to reduce total power consumption and cooling demands, easing the sustainability challenges faced by hyperscale operators.

“Majestic allows for a level of scalability and operational efficiency that simply isn’t possible with traditional GPU based systems,” says Sha Rabii, Co-founder and President.

“Our systems support vastly more users per server and shorten training time, lifting AI workloads to new heights both on-premises and in the cloud.

“Our customers benefit from tremendous improvements in performance, power consumption and total cost of ownership.”

A new foundation for AI data centres

The company’s focus on overcoming the memory wall represents a crucial turning point in data centre architecture. 

By reimagining how compute and memory interact, Majestic Labs seeks to redefine what is computationally possible, enabling the next generation of large-scale AI models to run more efficiently and sustainably.

With its experienced leadership team and significant investment backing, Majestic Labs positions itself at the centre of a growing movement to rethink AI infrastructure from the ground up. 

As the global AI race accelerates, innovations like Majestic’s could help determine which operators are able to scale responsibly and cost-effectively in the years ahead.