The startup has developed a patent-pending AI software solution that optimizes memory usage in computing environments, functioning seamlessly across on-premises and cloud configurations without requiring changes to the operating system or applications. This technology significantly reduces computing costs by enhancing efficiency in both virtualized and containerized settings.
Funding
$8.2M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Product
Problem
Server memory (DRAM) is a significant cost component in modern data centers, especially with the demands of AI and large-scale databases. Memory utilization often drops to 50% for many workloads, leading to wasted resources and unnecessary expenditure. Traditional solutions for memory limitations, such as procuring larger systems or sharding data, are complex, time-consuming, and expensive.
Solution
MEXT provides an AI-powered predictive memory software solution that addresses memory underutilization by transparently auto-tiering memory pages. The software offloads cold memory pages from DRAM to a lower-cost Flash tier, leveraging AI to predict which pages will be needed and proactively move them back to DRAM before the application requests them. This approach allows applications to run within a smaller DRAM footprint, reducing costs and expanding effective memory capacity without requiring changes to the operating system or applications. MEXT's solution works across on-premises and cloud configurations, as well as virtualized, bare-metal, and containerized environments.
Target Audience
MEXT targets organizations that deploy modern applications like AI and large-scale databases, and are looking to reduce computing costs or expand memory capacity.
Features
- AI-powered predictive engine for continuous memory page prediction and transparent auto-tiering
- Offloads cold memory pages from DRAM to a 20x lower-cost Flash tier
- Transparent memory tiering that keeps application performance intact
- Software-only solution compatible with existing hardware and software configurations
- Works across on-premises, cloud, virtualized, bare-metal, and containerized environments
- Observability tools to understand application memory requirements and utilization patterns