Vertical Compute
Vertical Compute is a Belgian deep-tech semiconductor startup spun out of imec that develops Vertical Integrated Memory (VIM) chiplets integrating high-density magnetic memory directly above compute logic. It targets semiconductor foundries, memory chip vendors, AI hardware providers and cloud providers seeking to address the AI memory wall.
- Company typePrivate
- Founded2025
- HeadquartersLouvain-la-neuve, Belgium
- Headcount11–50
- GTM typeB2B
- OfferingHardware or Manufacturing
What Vertical Compute does
Vertical Compute is a Belgian deep-tech semiconductor startup spun out of imec in October 2024 that develops Vertical Integrated Memory (VIM) technology — a proprietary approach to integrating vertical magnetic memory stacks directly above compute logic within a standard 300mm wafer manufacturing flow. The technology targets the so-called 'memory wall' bottleneck limiting AI and high-performance computing workloads, claiming >Gb/mm² data density (exceeding DRAM) while maintaining sub-10ns SRAM-like access latency. It is delivered as modular chiplets intended to drop into existing CPU, GPU and custom AI accelerator architectures rather than replacing compute cores.
The company was co-founded by CEO Sylvain Dubois (previously Google Cloud Infrastructure, AI hardware acceleration sourcing) and CTO Sébastien Couet (formerly imec Program Director for MRAM development, with 16+ patents and 130+ publications in magnetic memories). In just over a year since founding, Vertical Compute grew to approximately 25 employees across Belgium (Louvain-la-Neuve, Leuven) and France (Paris, Grenoble, Nice), and successfully completed its first 300mm VIM tape-out in March 2025. The VIM chiplet is designed to be manufacturable by standard semiconductor foundries, positioning the company as an IP and chiplet supplier rather than a vertically integrated chipmaker.
Vertical Compute is currently pre-revenue and operates a B2B go-to-market targeting semiconductor foundries, memory chip vendors, AI hardware providers and cloud providers through direct enterprise sales and OEM/chiplet integration. The company has raised €57M total seed funding across two tranches (€20M led by imec.xpand in January 2025, plus €37M led by Quantonation in March 2026), with non-dilutive capital from Flemish agency VLAIO and French public bank BPI France. Imec retains a strategic co-development relationship alongside its investment, and the company was featured at LAM Research's Future of AI and Semiconductors Event in May 2026, signaling early positioning within the advanced-packaging ecosystem.
Vertical Compute firmographics
Firmographics- Name
- Vertical Compute
- Legal name
- Vertical Compute SPRL
- Website
- https://verticalcompute.com
- Company type
- Private
- Founded year
- 2025
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Vertical Compute is a Belgian deep-tech semiconductor startup spun out of imec that develops Vertical Integrated Memory (VIM) chiplets integrating high-density magnetic memory directly above compute logic. It targets semiconductor foundries, memory chip vendors, AI hardware providers and cloud providers seeking to address the AI memory wall.
- Ownership category
- akta.pro rank
Vertical Compute industry classification
Industry- Product category
- Semiconductor Memory and AI Compute Chiplets
- NAICS
- Semiconductor and Other Electronic Component Manufacturing (3344)
- SIC
- Semiconductors & Related Devices (3674)
- akta.pro primary industry
- Silicon IP Subsystems & Chiplet IP (Die-to-Die, UCIe, NoC, I/O chiplets) (HDAHAFAH)
- akta.pro secondary industry
- AI Hardware IP & Chiplet/Packaging for Accelerators (EDA/IP, advanced packaging) (HDAAAAAL)
Keywords
Where Vertical Compute is headquartered
LocationHeadquarters
- HQ city
- Louvain-la-neuve
- HQ country
- Belgium
- HQ region
- Europe
Offices5 records
Markets served
Vertical Compute business model
Business model- GTM type
- B2B
- Offering type
- Hardware or Manufacturing
- Cost components
- Personnel, Technology or R&D, Supply Chain, Operations, Marketing or Sales, Infrastructure
Revenue model
- Chiplet Sales and Licensing: Vertical Compute manufactures and sells chiplets that integrate as enabling layers into existing processor architectures (CPUs, GPUs, AI accelerators). The technology is designed to be manufacturable by standard semiconductor foundries and integrates seamlessly into the existing supply chain.
- Chiplet-based Solution Deployment: Delivered through a chiplet-based approach, the technology is designed to integrate seamlessly into existing processor architectures, acting as an enabling layer enhancing CPUs, GPUs and custom AI accelerators with dramatically higher memory density and bandwidth.
Go-to-market motion2 records
Distribution channels3 records
Marketing channels6 records
Vertical Compute product offering
Product offeringCore offering
Vertical Compute develops and commercializes its proprietary Vertical Integrated Memory (VIM) technology, delivering it as modular chiplets that integrate vertical magnetic memory stacks directly above transistor logic within a single 300mm wafer manufacturing flow. The VIM chiplets act as an enabling layer that enhances existing CPUs, GPUs, and custom AI accelerators with >Gb/mm² memory density and sub-10ns (SRAM-like) access latency, eliminating the data-bus bottleneck. The technology is designed to be manufacturable by standard semiconductor foundries and is sold or licensed to chipmakers, memory vendors, AI hardware providers, and cloud providers.
Product overview
Vertical Compute offers a unified product portfolio centered on its proprietary VIM™ (Vertical Integrated Memory) technology, a breakthrough semiconductor memory solution that integrates vertical memory stacks directly on compute logic. The core VIM™ technology is delivered as modular chiplets that serve as an enabling layer for existing processor architectures — enhancing CPUs, GPUs, and custom AI accelerators. The chiplet-based approach allows chipmakers to integrate high-density, high-bandwidth memory directly above processing cores without replacing existing compute cores, addressing the AI memory bottleneck while maintaining compatibility with standard semiconductor manufacturing processes.
Differentiator
Problem solved
Functional benefit
Products and services
- VIM™ (Vertical Integrated Memory) Proprietary vertical integrated magnetic memory technology that combines memory and compute into a single cohesive solution, blending vertical memory concepts with nano-magnetism to achieve <10 ns latency with unlimited endurance similar to on-chip SRAM, while providing high density and cost-efficiency comparable to 3D-NAND. Delivered as modular chiplets for chipmakers.
- Vertical Memory Chiplet A chiplet-based memory product integrating vertical memory stacks directly on compute logic within a single 300mm wafer manufacturing process. Acts as an enabling layer enhancing CPUs, GPUs, and custom AI accelerators with dramatically higher memory density and bandwidth.
Quantifiable outcome
- 100x performance gains for AI workloads including LLMs and Agentic AI by eliminating memory bottlenecks
- +2 more outcomes
Companies that use Vertical Compute
Customer profileSegments7 records
Ideal customer profiles5 records
Vertical Compute technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability4 records
Feature4 records
Vertical Compute partnerships and signals
Strategic signalPartnerships
One partnership is on record.
- imeccoreVertical Compute spun out of imec in October 2024. imec is a world-class R&D center for nanotechnologies based in Leuven, Belgium. The company maintains a strategic partnership with imec and imec also reinvested as an existing investor. This relationship provides access to decades of world-class semiconductor research.
Scale indicators4 records
Recent moves7 records
Expansion highlights6 records
Vertical Compute competitors and assessment
Company assessmentBroad incumbents
- ARM: ARM is the dominant processor IP licensor, providing CPU and system IP that integrate into SoCs from Apple, Qualcomm, Nvidia, and others. It is comparable to Vertical Compute as a chiplet/IP licensor targeting the same AI accelerator and SoC ecosystem, though at significantly larger scale and broader scope.
- Synopsys: Synopsys is a leading EDA and semiconductor IP provider with a portfolio that includes memory compilers, interface IP, and design IP for AI/semiconductor applications. It is a broad incumbent comparable to Vertical Compute in the semiconductor IP licensing model, though at vastly larger scale and with a much broader product portfolio.
- Cadence Design Systems: Cadence is a major EDA and semiconductor IP provider serving the same chip design ecosystem as Vertical Compute's foundry and accelerator customers. It is a broad incumbent with a semiconductor IP portfolio that includes memory and interface IP, comparable to Vertical Compute's chiplet IP licensing model at significantly larger scale.
- Groq: Groq develops LPU inference chips with a memory-centric architecture (TSP - Tensor Streaming Processor) explicitly designed to address AI memory bottlenecks. It is comparable to Vertical Compute as a player attacking the AI memory wall, though Groq sells complete chips/systems rather than chiplet IP.
Direct peers
- NEO Semiconductor: NEO Semiconductor develops 3D DRAM and 3D NAND technologies (X-NAND, X-DRAM, X-CIM) targeting AI memory bottlenecks through novel 3D architectures. It is a direct peer to Vertical Compute in pursuing 3D-stacked memory solutions for the AI memory-wall problem, though NEO's approach focuses on DRAM/NAND extensions rather than chiplet-level integration.
- Everspin Technologies: Everspin is the leading commercial MRAM (magnetoresistive RAM) provider with spin-transfer torque memory products shipping in volume. It is a direct peer in magnetic memory technology, though focused on standalone MRAM products rather than Vertical Compute's integrated chiplet/memory-on-logic approach.
- Avalanche Technology: Avalanche Technology develops spin-transfer torque MRAM (STT-MRAM) products targeting aerospace, industrial, and AI applications. It is a direct peer to Vertical Compute in MRAM/spintronic memory technology, though Avalanche focuses on discrete memory products rather than chiplet-level compute integration.
- Crossbar: Crossbar develops Resistive RAM (ReRAM) memory technology for AI, IoT, and embedded applications. It is comparable to Vertical Compute as a non-volatile memory IP/startup targeting AI workloads with novel memory architectures, though ReRAM differs from Vertical Compute's magnetic memory approach.
Emerging players
- Mythic: Mythic develops analog in-memory compute AI processors that perform neural network inference directly within memory arrays to address AI memory bottlenecks. It is comparable to Vertical Compute as an emerging player attacking the memory-wall problem, though Mythic focuses on analog compute-in-memory rather than 3D-stacked magnetic memory.
- Tenstorrent: Tenstorrent develops AI accelerators with a RISC-V based architecture and custom memory hierarchy targeting LLM inference workloads. It is comparable to Vertical Compute as an emerging AI hardware player attacking the memory-wall challenge, though Tenstorrent focuses on full-stack accelerator systems rather than chiplet-level memory IP.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat4 records
Key risks6 records
Key highlights6 records
Customer concentration
Vertical Compute social profiles
Digital presenceVertical Compute financial estimates
Financial estimateRevenue estimate
Valuation estimate
Vertical Compute leadership team
Management profileNumber of profiles
Profiles2 records
Vertical Compute funding detail
Funding detailFunding overview
Funding rounds3 records
Investors15 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Vertical Compute M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Vertical Compute
What does Vertical Compute do?
Vertical Compute develops and commercializes its proprietary Vertical Integrated Memory (VIM) technology, delivering it as modular chiplets that integrate vertical magnetic memory stacks directly above transistor logic within a single 300mm wafer manufacturing flow. The VIM chiplets act as an enabling layer that enhances existing CPUs, GPUs, and custom AI accelerators with >Gb/mm² memory density and sub-10ns (SRAM-like) access latency, eliminating the data-bus bottleneck. The technology is designed to be manufacturable by standard semiconductor foundries and is sold or licensed to chipmakers, memory vendors, AI hardware providers, and cloud providers.
Is Vertical Compute a public or private company?
Vertical Compute is a private company. It is classified as venture growth investor backed and is currently operating.
When was Vertical Compute founded?
Vertical Compute was founded in 2025. It employs 11 to 50 people.
Where is Vertical Compute based?
Vertical Compute is headquartered in Louvain-la-neuve, Belgium, in the Europe region.
How does Vertical Compute make money?
Two revenue lines are on record. Chiplet Sales and Licensing is the primary driver. The others are chiplet-based Solution Deployment.
Who are Vertical Compute's main competitors?
Broad incumbents on record are ARM, Synopsys, Cadence Design Systems and Groq. Direct peers are NEO Semiconductor, Everspin Technologies, Avalanche Technology and Crossbar. Emerging players are Mythic and Tenstorrent.
Does Vertical Compute have an API?
No public API is recorded for Vertical Compute.
What industry is Vertical Compute in?
Vertical Compute's product category is Semiconductor Memory and AI Compute Chiplets. Its primary akta.pro industry code is HDAHAFAH, Silicon IP Subsystems & Chiplet IP (Die-to-Die, UCIe, NoC, I/O chiplets), with a secondary code of HDAAAAAL, AI Hardware IP & Chiplet/Packaging for Accelerators (EDA/IP, advanced packaging). Its NAICS code is 3344 and its SIC code is 3674.