AI Event Mag ·
Europe's October arc: from Berlin's GTC keynote to robot football in Rome
Issue 07 stretches across a dense October: NVIDIA's GTC in Berlin with a Jensen Huang keynote at the Tempodrom, the free-to-attend AI Congress Barcelona, the Nordic Data 2030 Summit in Stockholm, Maker Faire Rome with its first humanoid RoboCup football tournament and — as a compact opener — the KI-Forum Wien on AI & cybersecurity. Framed by an AI ticker on the now actively enforced GPAI oversight under the EU AI Act and the early-September release wave.
Europe's AI autumn is being celebrated, for the first time, under fully enforced rules. Since 2 August 2026 the European Commission's AI Office and the member states are responsible for implementing, supervising and enforcing the AI Act — including real powers over providers of general-purpose models: requesting technical documentation, evaluating models, demanding corrective measures, imposing fines. By that same date every member state was expected to have set up at least one national AI regulatory sandbox. Governance is no longer optional; it is an operating condition.
The model side, meanwhile, kept moving: in early September the major labs shipped new systems within days of each other. OpenAI unveiled GPT-6 “Astra”, promoting it as its best-aligned model yet and drawing an open safety debate; Google, Anthropic and OpenAI announced explicitly cybersecurity-oriented models almost simultaneously, complete with tiered safeguards and trusted-access programmes. For teams, the guiding question shifts from “which model?” to “which workflow — and who is accountable for it?”.
That very shift — from capability to accountability, from demo to sustained operation — can be seen live over the next four weeks. Five stops, five perspectives: the concentrated compute-and-robotics stage in Berlin, the publicly accessible strategy congress in Barcelona, the data-practitioner round in Stockholm, the maker festival in Rome and the compact compliance day in Vienna.
01
NVIDIA GTC Berlin — Jensen Huang's keynote and three days of compute
· Berlin, Germany · STATION-Berlin
From 20 to 22 October, NVIDIA brings its GTC to Berlin: workshops and training on 20 October, the main conference at STATION-Berlin, and at its heart the keynote by CEO Jensen Huang on Wednesday 21 October from 11:00 at the Tempodrom. According to the catalogue the programme bundles more than a hundred sessions across eight tracks — from Agentic AI and AI for Science through AI Infrastructure and Data Processing to Industrial AI, Open Models and Physical AI & Robotics.
The speaker line-up is strikingly European and industry-facing: Robin Rombach (Black Forest Labs) represents generative image research, Jörg Burzer (Mercedes-Benz) and Christian Senger (Volkswagen) the car industry, Pallavi Mahajan (Nokia) and Anton Repushko (Revolut) telco and fintech. With Jörg Bienert of the Center for Sovereign AI and Christian Piechnick (Wandelbots), sovereign AI and robotics are on the bill too. Beyond the talks there are hands-on labs, certifications and an exhibition floor.
Pricing runs from an exhibit-hall pass at 357 euros to a full conference pass around 1,755 euros, with discounts for academia, government and non-profits. NVIDIA does not state an official attendance figure for the Berlin edition — whether the trip pays off is decided by the session catalogue rather than a headcount.
Speaker highlights
- Jensen Huang — CEO, NVIDIA (Keynote)
- Robin Rombach — Co-Founder & CEO, Black Forest Labs
- Jörg Burzer — Vorstandsmitglied & CTO, Mercedes-Benz Group
- Christian Senger — CEO Autonomous Mobility, Volkswagen
- Pallavi Mahajan — Chief Technology and AI Officer, Nokia
- Jörg Bienert — Managing Director, Center for Sovereign AI (CESAI)
- Christian Piechnick — CEO & Co-Founder, Wandelbots
02
AI Congress Barcelona — three free days between strategy and governance
· Barcelona, Spain · Auditori L'illa
From 20 to 22 October, the twelfth AI Congress Barcelona takes over the Auditori L'illa on Avinguda Diagonal. It is run by CIDAI (Centre of Innovation for Data Tech and AI), coordinated by Eurecat as part of the Catalan government's AI strategy — attendance is free after registration, which makes the congress one of the most open large-scale formats of the European autumn.
The programme is clearly structured across three days: “Strategy and Trends” to open (enterprise AI readiness, GenAI with privacy and cost control, multi-agent security, agentic AI and GDPR), “Technology & Development” on day two (AI agents across industries, governance against “agent sprawl”, responsible-AI self-assessment) and “Inspiring Projects” to close (synthetic data and data spaces, auditable AI, multimodal as well as vertical versus general-purpose models).
The speaker list mixes startups, corporates and public administration: Ariadna Font (Alinia AI) opens with a keynote on preparing organisations for AGI with human oversight, and Jezabel R. Garcia (Multiverse Computing) speaks on model compression and sovereign AI. Among the fifty-plus speakers are organisations such as CaixaBank Tech, Orange/APDCAT, Sopra Steria, Nestlé and HP; gold sponsors including CaixaBankTech, Conversia and Forvis Mazars Infini back the congress.
Speaker highlights
- Ariadna Font — Co-Founder & CEO, Alinia AI
- Jezabel R. Garcia — AI Research Manager, Multiverse Computing
- Joan Mas — CIDAI
- Mireya Saludes — CIDAI
- Jaume Clotet — Forvis Mazars Infini
03
Data 2030 Summit Stockholm — the Nordic round of data leaders
· Stockholm, Sweden · Hotel Birger Jarl
On 21 and 22 October, the eleventh Nordic edition of the Data 2030 Summit gathers at Hotel Birger Jarl, hosted by Hyperight. The format is hybrid — capped, according to the organiser, at 250 on-site delegates, with unlimited online access via the Agorify platform. The small in-person cap is deliberate: the summit favours proximity between data leaders over a trade-hall operation.
The content runs across three parallel stages — Data Strategy & Governance, Data Platform/Context & Architecture, and Data Products, Data Quality & AI Enablement. Added to this are two interactive roundtables, an expert panel, half-day pre-event workshops on the morning of 21 October (limited to 30 participants) and a MasterClass with Samir Sharma (Datazuum) on the “Data & AI Strategy Canvas”. The working language throughout is English.
The speaker list reads like a who's-who of European data practice: Ulrika Jägare (SAS – Scandinavian Airlines), Ela Osterberger (Booking.com), George Linton (Chief Data Officer, Electrolux Group), Frank J. Kartmann (Deutsche Bank) and Alex Astrogold (Ericsson) are among the thirty-plus presenters. After the conference day, a cocktail reception rounds off the networking programme.
Speaker highlights
- Ulrika Jägare — VP Data and AI, SAS (Scandinavian Airlines)
- Ela Osterberger — Director of Experimentation & Measurement, Booking.com
- George Linton — Chief Data Officer, Electrolux Group
- Frank J. Kartmann — VP Platform Architecture & AI, Deutsche Bank
- Samir Sharma — CEO, Datazuum (MasterClass)
04
Maker Faire Rome — where humanoid robots play football for the first time
· Rome, Italy · Gazometro Ostiense
From 23 to 25 October, the Gazometro Ostiense, a former gasworks in the south of Rome, turns into the fourteenth Maker Faire Rome — The European Edition. Organised by the Rome Chamber of Commerce through Innova Camera under a Make: licence, the Faire is an exhibition-and-festival format: “Nobody's pitching, everyone's showing.” An independent scientific committee selects the projects; the themes span robotics, AI and IoT through digital fabrication and sustainability to agritech, food-tech and space-tech.
The 2026 highlight is a first: an Italian Open RoboCup tournament for humanoid robot football comes to Rome — an official RoboCup regional event staging the first international football tournament for humanoid robots. Universities and research centres compete in 3-on-3 and 5-on-5 matches, accompanied by technical tests, public demos and a RoboCup humanoid-soccer workshop; among the participants is Team SPQR of Sapienza University of Rome.
Because the Faire is a public festival rather than a speaker-centred conference, exhibits and demonstrations take centre stage instead of a keynote line-up. The figures cited by the organiser — a cumulative 895,000-plus visits and more than 8,000 projects from over 40 countries — are explicitly historical totals since 2013, not a forecast for this edition.
05
KI-Forum Wien — a compact day on AI, compliance and cybersecurity
· Vienna, Austria · DC Tower
The earliest date in this issue belongs to the KI-Forum Wien on 6 October — a one-day, German-language format hosted by PwC on the 41st floor of the DC Tower in Vienna's Donau City. It is the series' second Austrian stop of 2026 after Linz and deliberately foregrounds the less glamorous but operationally decisive questions: how AI actually works, where the EU AI Act bites, and how to keep deepfake risks out of the organisation.
The morning (09:00–12:00) is an AI foundations training including a certificate under Article 4 of the EU AI Act — precisely the proof of AI literacy the regulation requires of organisations. The afternoon (13:00–17:15) brings six practice-oriented talks, among them EU AI Act compliance (Philipp Summereder), AI and data security (Roland Pucher, PwC) and concrete implementations in accounting, HR and IT.
The day is framed by all-day access to the 41st-floor sky lounge and a closing networking session over drinks. For smaller teams looking for a structured entry into the AI Act's obligations, the Vienna forum is a manageable counterpoint to this issue's large congresses.
Speaker highlights
- Manuel Wolfsteiner — KI-Basisschulung (ki-fit.at)
- Philipp Summereder — EU AI Act / Regulierung
- Roland Pucher — KI & Datensicherheit, PwC
- René Zeilinger — Sovereign AI for Organizations (headwai.org)
AI ticker: models by the week, oversight with teeth
September delivered one of the densest release windows of the year. OpenAI unveiled GPT-6 “Astra”, promoting it as its best-aligned model yet; the announcement came with an open safety debate in which researchers stressed the still “jagged” performance curve of today's AI. Almost simultaneously, Google, Anthropic and OpenAI announced explicitly cybersecurity-oriented models — with tiered safeguards and trusted-access programmes that keep the most sensitive capabilities behind review processes.
A signal from research stood out too: Anthropic set up a life-sciences group and reported that Claude had autonomously identified a previously uncharacterised enzyme system in bacteriophages whose architecture recalls CRISPR arrays. Such news underlines that the most interesting AI frontier right now is less the chatbot than domain-specific, application-near assistance.
And the regulatory foundation is keeping pace: since 2 August 2026 the AI Office and member states have been actively enforcing the EU AI Act's GPAI rules, with powers ranging from documentation requests to fines; the national regulatory sandboxes were due in parallel. That simultaneity — fast models, oversight that means it — is the thread through October: in Berlin, Barcelona, Stockholm, Rome and Vienna, the theme is less the next demo than reliable, auditable operation.