AI Austria monthly: July'25 Update
Over the summer, spread knowledge about AI!
Vienna's ambitious AI Gigafactory bid, what a new study reveals about our AI ecosystem, events recap and outlook and your chance to influence the EU AI Act.
🚀 Latest Developments
🏭 Vienna in the run for a European AI Giga factory
Austria is stepping up its commitment to AI by pledging up to €1.75 billion to attract a major AI Gigafactory to the country. This strategic investment highlights Austria’s dedication to becoming a global hub for cutting-edge AI technologies and high-tech manufacturing.
The planned gigafactory, which could generate thousands of new jobs and significantly boost economic growth, aims to position Austria prominently on the international AI landscape. The government’s ambitious support package underscores Austria’s focus on innovation, sustainability, and technological leadership. » official statement
🇪🇺 Insights from the 11th European AI Forum in Gdansk
Last week, our European umbrella organization, the European AI Forum, held its flagship conference in Gdansk, offering a sharp look into the continent's AI future. For those who couldn't attend, the key discussions revolved around the practical, national-level implementation of the AI Act, with a special focus on the development of regulatory sandboxes. Another major theme was Europe's digital sovereignty, with panels exploring the EU's ambitious plans for "AI Gigafactories" and the "Apply AI Strategy." The forum also highlighted the crucial role of open-source AI as a key driver of innovation for startups and SMEs, celebrating the official launch of AI Poland, a new coalition of over 40 organizations. In a move to broaden the network, the forum warmly welcomed AI Albania and the Georgian AI Association as its newest members.» Watch the full event recording on YouTube
🎨 How Top AI Artists Are Redefining Creativity
What happens when successful AI artists reveal their creative secrets? Our new event series, "The New Creative Wave," kicked off on June 12th at the Home of Innovation X, pulling back the curtain on what it truly takes to master creative AI. Hosted by Andreas Fraunberger and Ruben Hetfleisch, the launch featured viral creators like Tanja Schalling and Manuela Klauser alongside artist Wilhelm Steiner. The key takeaway from the evening was clear: in an age where everyone has access to the same powerful tools, the artists who succeed are the ones who impose strict, human-led constraints. They treat AI not as a chaos engine, but as a precise instrument to be directed, blending traditional skills with cutting-edge tech to create truly unique work. » Read the full playbook on the key takeaways from the event
📊 Study Reveals Austria's "Informed Optimism" on AI
How ready is Austria for the AI revolution? The new "AI Readiness Study 2025," a comprehensive survey of 1,500 consumers and 400 businesses by the Handelsverband and Google, provides a detailed snapshot of where we stand. The report reveals a sense of "informed optimism," with 60% of Austrians showing interest in AI. On the business side, 68% of Austrian companies are already using AI, primarily in marketing and customer service, with 37% already investing in AI training. However, the study also highlights a critical challenge: a significant skills gap that needs to be addressed for the ecosystem to mature. » Get the full study here (in German)
🏨 Austrian TravelTech Startup Chatlyn Raises €8 Million
Vienna-based hospitality AI startup, chatlyn, has successfully closed an €8 million Series A funding round. The company is developing an "AI brain for hotels," a sophisticated communications hub designed to unify and automate guest interactions across more than 10 channels and 35 languages. This investment, led by Smedvig Ventures, will fuel chatlyn's international expansion and further development of its platform, showcasing strong investor confidence in Austrian AI-driven solutions for global industries. » Read the full story on EU-Startups
🍏 Relaunch in Graz: The AI Styria Meetup is Back!
Get ready, Styria! After a long break, the Graz ML Meetup is returning on July 8th, bigger and better than ever. We're teaming up with Wirecube to reignite the local AI scene, uniting researchers, builders, and industry professionals. The evening is packed with deep dives into real-world use cases, including the pitfalls of on-premise LLMs, successes in predictive maintenance, and the latest on RAG agents. Join us for an evening of learning and connect with the community over pizza and drinks. Save your spot and join the meetup
🇪🇺 Defining High-Risk AI: A Chance for Clarity
The European Commission is seeking expert feedback to refine the classification of "high-risk" AI systems under the AI Act. This is a crucial opportunity to address an issue AI Austria has consistently highlighted: the need for precise, technically-sound definitions instead of vague, sweeping categories. To ensure the voice of the Austrian AI community is heard on this critical point, we strongly encourage our members to contribute their practical expertise to this consultation. The outcome will directly influence compliance and product development for all, and it's vital that the final rules are both effective and practical. The deadline for submissions is July 18, 2025. » Submit your feedback and contribute to clearer regulation
📚 A Guide to AI Education in Austria
With a growing number of AI-related educational programs, finding the right one can be a challenge. Our AI Literacy Landscape Austria platform has been launched to provide a clear overview of the options. This directory lists a wide range of courses, workshops, and certifications available across the country. By organizing these offerings, the platform aims to help individuals and companies identify suitable training and skill development opportunities in the field of AI. » Explore the directory at AI Literacy Landscape Austria
✨ Generative AI News ✨
👉The Future of AI-Powered Development?
Cursor 1.0, released on June 4, 2025, marks a major upgrade in AI-assisted development. Key features include BugBot for automated code reviews, Background Agent for remote tasks, and Memories for persistent project context. It also adds one-click MCP integration, Jupyter support, and a redesigned dashboard with advanced analytics.
👉Apple’s Privacy-First Approach to App Intelligence
Apple’s new Foundation Models framework lets developers easily add powerful, private on-device AI to their apps using simple Swift code. Running entirely offline on Apple devices, it enables features like summarization, natural language search, and personalized content, without sending data to the cloud. This might become a major shift toward smarter, privacy-first apps.
👉Anthropic’s Fair Use Victory With a Major Catch
A federal judge ruled that Anthropic's use of legally obtained copyrighted books to train Claude was fair use, but allowed claims to proceed for millions of pirated books. This sets precedent that AI companies can train on legitimately acquired copyrighted material, while facing potential damages of up to $150,000 per pirated work.
👉Which way, AI Engineer?
Anthropic embraces multi-agent systems, reporting 90% performance gains from parallel specialized agents, while Cognition argues they're fundamentally fragile and advocates for single-threaded agents instead. The core debate: performance through parallelism versus reliability through simplicity.
👉Prompt Engineering is Dead, Long Live Context Engineering
The AI engineering community is adopting "context engineering" over "prompt engineering" to better describe the complex art of assembling information, tools and formatting within LLM context windows. This shift reflects the evolution from simple "ChatGPT wrappers" to sophisticated production systems requiring careful careful orchestration of context, tools, and control flows.
👉From AlphaFold to AlphaGenome: DeepMind's Next DNA Breakthrough
Google DeepMind released AlphaGenome, an AI model that analyzes up to 1 million DNA base pairs to predict how genetic variants affect gene regulation, maintaining high resolution while processing long sequences. It achieves state-of-the-art performance on genomic benchmarks and is now available via API for research use.
🌏 Asia spotlight
🇨🇳 China | Xiaohongshu open-sources dots.llm1 Xiaohongshu has open-sourced its first large language model, dots.llm1. The model contains 142 billion parameters in a mixture-of-experts (MoE) architecture, yet only 14 billion parameters are active at inference. It achieves performance close to Alibaba’s Qwen3-32B across Chinese- and English-language tasks, mathematics, and alignment benchmarks. A defining feature of dots.llm1 is its training data: the variant dots.llm1.ins was pre-trained on 11.2 trillion tokens of non-synthetic data. As a nationwide social-content platform whose latest valuation has surged to roughly RMB 250 billion, Xiaohongshu seeks to show that efficient model design coupled with high-quality data can extend the capability frontier of large language models.
After the Labubu craze, a new “AI fluff-ball” collectible has exploded in popularity: backed by investor Zhu Xiaohu, it was priced at RMB 399 and sold out at launch. From day one, the founders designed “Fuzai” for women aged 18-35. Early research by Luobo Intelligence shows that many female users long for the kind of emotional connection and companionship one gets from a pet, friend, or even a partner—and the interactive power of large AI models offers a fresh way to satisfy that deeper need. Women are far more likely to bond with an AI toy that is “fluffy, lightweight, and able to show emotion” than with a desktop, guard-like robot assistant.
Galaxy Universal and Tsinghua University have launched OpenWBT, the industry’s first open-source, full-body teleoperation system for humanoid robots. OpenWBT works across multiple robot models and bridges virtual–physical environments, allowing deployment in just a matter of hours.
Astribot, a pioneer in embodied-intelligence robots, has entered a deep strategic partnership with Shenzhen’s first municipally run nursing-care institution, the Shenzhen Nursing Home. Leveraging their respective strengths in technology and application scenarios, the two sides will focus on developing AI-powered elder-care robots, training multimodal datasets, and deploying intelligent care systems. Together they will explore innovative uses of embodied‐intelligence technology in daily living assistance, health monitoring, and emotional companionship—advancing a new model of “smart, tech-enabled elder care” that blends cutting-edge science with human compassion.
Visual perception-driven multimodal reasoning: Alibaba Tongyi introduces VRAG, defining the next generation of retrieval-augmented generation。
Ant Group is ramping up its AI-driven healthcare ambitions and on 26 June rolled out a new app called “AQ.” Designed to meet the public’s everyday medical and health-management needs, AQ packs 100+ AI features—from health education and appointment guidance to lab-report explanations and personal health records. The platform also hooks into a vast professional network: 5,000+ hospitals nationwide, nearly one million physicians, and about 200 “digital twin” specialist avatars. The app is now gradually appearing in all major app stores.
📆 Upcoming Events
Aug 25 Prompt Engineering Essentials
Aug 28 Foundations of LLM Mastery: Retrieval Augmented Generation
Sep 08 Fine-tuning LLMs on one GPU
👉 For more information on upcoming events head over to our community calendar.
🧑💻 Featured Jobs
Once a month we feature selected openings from our job board in this section.
Data Intelligence & AI Readiness, Senior Consultant ML und LLM Engineering,
Praktikum Künstliche Intelligenz und ML/LLM Engineering, Manager:in Software Delivery mit Schwerpunkt Künstliche Intelligenz at EY
Junior AI Engineer at Deine KI Agentur
PostDoc researcher (f/m/d) for AI in Environmental Simulation at VRVis GmbH
Senior Software Entwickler:in, Junior Software Entwickler:in and Data Scientist at Blumatix
👉 Check out more listings at our Job Board