VIDEANTIS GMBH

Their core work

Videantis is a Hannover-based SME that designs processor IP and embedded software for efficient deep learning inference on resource-constrained hardware. Their core expertise is translating computationally heavy neural network workloads into implementations that fit within the power and area budgets of embedded and automotive-grade chips. In PRYSTINE, they contributed to programmable AI architectures for safety-critical automotive perception systems; in TEMPO, they worked on the hardware substrate for neuromorphic computing using emerging non-volatile memory technologies like OxRAM and MRAM. They sit at the intersection of chip architecture and applied AI — a rare combination that makes them valuable in any consortium needing someone who can bridge algorithm design and silicon implementation.

What they specialise in

How they've shifted over time

Their first project (PRYSTINE, 2018) placed them squarely in automotive AI — sensors, safety-critical systems, dependable embedded architectures, and automated driving. By 2019, TEMPO shifted their focus toward the algorithmic and hardware foundations of neuromorphic computing: spiking neural networks, ASIC design, and emerging memory technologies like OxRAM and MRAM. This is not a departure from embedded AI but a deepening into its next generation — moving from conventional deep learning on automotive-grade processors toward brain-inspired hardware that could reshape how inference is performed at the edge.

Videantis appears to be positioning itself at the frontier of post-GPU inference hardware — if neuromorphic and in-memory computing mature as expected, they will be early with relevant ASIC and architecture expertise.

How they like to work

Videantis has never led an H2020 project — they participate exclusively as a specialist partner. Their two projects both fall under RIA schemes, meaning they join research consortia to contribute technical depth rather than to manage or disseminate. Despite only two projects, they have accumulated 75 unique consortium partners, which reflects participation in large industry-driven consortia (PRYSTINE in particular was a major ECSEL Joint Undertaking project with dozens of partners across the European automotive supply chain).

Videantis has connected with 75 unique partners across 16 countries through just two projects, indicating they operate inside very large, multi-stakeholder consortia. Their network skews toward the European automotive and semiconductor ecosystem rather than academia or public research.

What sets them apart

Videantis occupies a rare niche: a small German company with genuine silicon-level expertise in neural network hardware, active in both automotive safety (a mature, regulated domain) and neuromorphic computing (an experimental frontier). Most SMEs in this space either stay close to software frameworks or focus on one market vertical — Videantis spans both. For a consortium needing someone who can speak the language of chip architects, safety engineers, and AI researchers simultaneously, they are an unusual and valuable find.

Highlights from their portfolio

• PRYSTINE
  Their largest project (EUR 222,113), part of a major pan-European automotive AI safety initiative, placing them inside the core supply chain of intelligent vehicle perception systems.
• TEMPO
  Positions Videantis at the hardware frontier of neuromorphic computing, combining ASIC design with emerging memory technologies — a high-risk, high-potential research direction with few SME participants.