Vertical Compute offers Vertical Integrated Memory (VIM™), a 3‑D chiplet technology that stacks magnetic memory directly above processing cores to provide ultra‑dense, non‑volatile storage with sub‑10 ns latency and unlimited endurance. By eliminating the separate memory hierarchy and data bus, VIM™ reduces power consumption and bandwidth bottlenecks, enabling efficient on‑chip inference for large AI models across data‑center, edge, and consumer applications.
Funding
$20.9M 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
Modern AI and high‑performance computing workloads are limited by the “memory wall” – the latency, bandwidth, and energy costs of moving data between separate memory and compute units. This bottleneck prevents large models such as LLMs from running efficiently on existing chips and forces costly, power‑hungry architectures.
Solution
Vertical Compute addresses the memory wall with its proprietary Vertical Integrated Memory (VIM™) technology. VIM™ stacks magnetic bit‑strings vertically above transistors, creating ultra‑dense, non‑volatile memory directly on the same chiplet as the processing cores. This 3‑D integration delivers >10 Gb/mm² density with sub‑10 ns SRAM‑like latency and unlimited endurance, eliminating the need for a separate memory hierarchy and data bus. By shortening the physical distance between data and compute, the architecture reduces power consumption dramatically while maintaining high performance, enabling on‑chip inference for large language models and other AI workloads. The solution is manufacturable in standard semiconductor foundries, allowing seamless integration into existing supply chains for data‑center accelerators, edge devices, and consumer electronics.
Target Audience
Primary customers are semiconductor foundries and chip designers seeking high‑density, low‑latency memory‑compute integration, as well as AI hardware vendors, cloud and edge computing providers, and enterprises that require efficient on‑device AI inference.
Features
- Vertical Integrated Memory (VIM™) stacks magnetic memory cells directly above CMOS transistors, achieving >10 Gb/mm² density
- Sub‑10 ns access latency with SRAM‑level speed and 3D‑NAND‑level cost efficiency
- Unlimited endurance and non‑volatile storage, removing the need for refresh cycles
- No external memory bus, eradicating bandwidth bottlenecks and reducing system power
- Compatible with standard semiconductor fab processes, enabling chiplet‑based deployment
- Supports on‑chip inference for large language models and other data‑intensive AI workloads
- Designed for seamless integration across cloud data centers, edge devices, and consumer products