Three ex-tech leaders from Google and Meta are building AI servers with unprecedented memory capacity — a breakthrough that could reshape data center infrastructure and cut cloud computing costs worldwide.
In a bold move to reshape the future of AI infrastructure, three former Google and Meta silicon executives have launched Majestic Labs, a new AI server startup that has raised an impressive $100 million in funding.
Their mission? To develop a next-generation server architecture that promises massive memory capacity and unprecedented efficiency — potentially replacing up to 10 traditional server racks with just one Majestic system.
Building the Future of AI Hardware
Founded by Ofer Shacham, Sha Rabii, and Masumi Reynders, Majestic Labs is setting out to solve one of the biggest challenges in artificial intelligence: memory limitations in large-scale data processing.
Their patent-pending architecture boasts up to 1,000 times more memory than a typical enterprise-grade server — a leap that could transform how companies manage AI workloads.
“We’re not trying to replace GPUs across the board,” said Shacham, Majestic’s CEO. “We’re solving for memory-intensive AI workloads where the fixed compute-to-memory ratio becomes a constraint.”
From Silicon Valley to Startup Powerhouse
The founding trio’s tech roots run deep. Before creating Majestic Labs, they led advanced silicon product design at both Google and Meta, contributing to key projects like Google’s Argos video chip and Facebook’s Agile Silicon Team (FAST).
Their shared experience gave them the insight — and confidence — to strike out on their own. “We’ve been friends and colleagues for a long time,” said Reynders, now Majestic’s COO. “The idea of doing something groundbreaking together was always in the periphery.”
A $100 Million Bet on Smarter Servers
Majestic Labs recently closed its $71 million Series A round, led by Bow Wave Capital and joined by Lux Capital. Combined with earlier seed funding, the startup has now raised $100 million to accelerate development.
The company is designing complete chip systems — much like Nvidia — but optimized for high-memory, high-efficiency performance. Each Majestic server is designed to deliver the compute power of several racks, while consuming less energy, space, and cooling resources.
This could significantly reduce data center costs for hyperscalers and industries like finance, healthcare, and pharmaceuticals, where massive AI models demand ever-expanding infrastructure.
A Market Hungry for Change
Majestic’s launch comes as major tech giants — including Alphabet, Microsoft, Amazon, and Meta — are collectively projected to spend $380 billion this year on data center expansion.
While Nvidia’s GPUs currently dominate AI workloads, memory bottlenecks remain a serious constraint. Majestic’s breakthrough could help companies run larger AI models faster and more efficiently, offering a strong complement to existing GPU-based systems.
From Vision to Reality
Prototypes of Majestic’s AI servers are expected by 2027, with early customer trials already in discussion. The company currently employs fewer than 50 people, split between Tel Aviv and Los Altos, California, and plans to expand both teams in 2026.
The founders are also tapping into their powerful network — more than 1,500 former colleagues from Google and Meta — to build a world-class team of engineers and technologists.
“Every large company now needs AI,” Shacham noted. “This is the right time to build the systems that make that possible — efficiently, affordably, and at scale.”
The Big Picture
Majestic Labs isn’t just another startup — it’s a sign of where the AI hardware revolution is headed. By combining deep technical expertise with a mission to cut data center costs, this ex-Google and Meta trio is positioning their company as a key player in the next generation of AI infrastructure.
Can startups like Majestic Labs out-innovate tech giants in the AI hardware race? Share your thoughts below — we’d love to hear your take on the future of smart computing.
