The Verifiable Stack: Confidential Compute & Private AI
AI is accelerating faster than ever and is gradually moving into Danish companies.
With this progress come questions that most forums fail to answer concretely: What can you actually document about the system you’ve put into production? Which architectural choices have real security consequences? When are cryptographic solutions mature enough to be relevant for your organization?
The Verifiable Stack is a monthly meetup for everyone working with AI who is responsible for the systems they deploy. Each event brings together practitioners, engineers, and researchers who work with verifiable and private AI on a daily basis.
The Focus
We revolve around two core disciplines:
Verifiable AI: Documenting what a system does (and doesn't do) through cryptographic attestation mechanisms, zero-knowledge proofs, and trusted execution environments.
Private AI: Reducing data exposure during processing and inference through local models, federated learning, and encrypted computation.
Both fields are maturing rapidly. Part of this series' purpose is to provide a realistic picture of what is usable now, what is on the horizon, and what is still at the research stage.
Practical Information:
Speakers: Luis Bezzenberger (Brainbot), Kristoffer Nærland (RedHat) & Peter Al Humaidan (SAIBA)
Location: Niels Bohr Bygningen, KU Jagtvej 128 - 2200 København N
Date: April 23, 2026, from 16:00–18:00
The Verifiable Stack is initiated by SAIBA in collaboration with AI Denmark and Nordic Blockchain Association.