Observed Signal · May 15, 2024 · Product Launch · Source: Tech.eu · Impact: 2/5 · Sentiment: Positive
Silo AI Launches Viking 7B Multilingual LLM for Nordic Languages
Silo AI, Europe's largest private AI lab, released Viking 7B, the first multilingual large language model covering all Nordic languages. Developed with the University of Turku's TurkuNLP research group and HPLT, the model also supports English and programming languages. Silo AI reports best-in-class performance across Nordic languages compared to other open models without compromising English performance. Viking 7B is trained on the EuroHPC LUMI supercomputer using up to 4096 AMD MI-250X GPUs, powered entirely by hydroelectricity. The release is part of Silo AI's effort to strengthen European digital sovereignty and reduce English linguistic dominance in AI. Training progress stands at 100% for Viking 7B, 85% for the 13B model, and 65% for the 33B model. The company also offers a model-serving platform for organisations to customise and deploy models using their own data.
Introduces an open multilingual LLM covering Nordic low-resource languages, supporting AI digital sovereignty, but does not directly shift AdTech or major platform dynamics.
Track AMD Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- Silo AI, with University of Turku's TurkuNLP and HPLT, released Viking 7B, the first multilingual LLM for all Nordic languages.
- Viking 7B covers Danish, Finnish, Norwegian, Icelandic, Swedish, English, and programming languages.
- The model was trained on the EuroHPC LUMI supercomputer using up to 4096 AMD MI-250X GPUs powered by hydroelectricity.
- Training progress is 100% for Viking 7B, 85% for the 13B model, and 65% for the 33B model.
- Silo AI claims best-in-class performance versus open models such as Falcon, GPT-SW3, Llama, Mistral, and MPT.
Connected Companies & Entities
2 Entities mapped“Viking is trained on the EuroHPC supercomputer LUMI, utilising up to 4096 AMD MI-250X GPUs....”
“Today Europe’s largest private AI lab Silo AI is releasing, together with the University of Turku’s research group TurkuNLP and HPLT, the fi...”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
US suspends Microsoft, Adobe from green card labor program
The Trump administration has suspended Microsoft, Adobe, and six other major tech firms—Capgemini, Cognizant, HCL, Infosys, Tata, and Wipro—from the U.S. Permanent Labor Certification program, which facilitates green cards for skilled foreign workers. Secretary of Labor Keith Sonderling cited active federal investigations and alleged fraud, noting that no new or pending applications from these companies will be accepted. Vice President JD Vance accused Microsoft of replacing laid-off workers with H-1B visa holders. Microsoft defended its practices, stating that most U.S. employees are American and that 80% of its H-1B petitions were for existing employees. The announcement was made during a White House summit on H-1B fraud, and the administration also plans to investigate nine universities, including Harvard, Yale, and Stanford, over student visa program abuse.
Tuesday's Top Analyst Calls: Nvidia, AMD, Meta & More
CNBC compiled Tuesday's biggest analyst calls, featuring upgrades, downgrades, and price target changes across a wide range of sectors. Notable calls include Evercore ISI upgrading Procter & Gamble to Outperform, Wells Fargo raising Meta's price target to $1,000, and Citi raising AMD's price target to $800. Goldman Sachs reiterated Buy on Netflix and SpaceX, while Morgan Stanley reiterated Overweight on SpaceX. The roundup covers technology, consumer, industrial, and financial stocks, providing insights into analyst sentiment and market expectations.
Reflection AI launches open-weight model Beam at lower compute cost
Reflection AI has launched Beam, its first frontier open-weight AI model, claiming it matches leading Chinese models like GLM-5.2 on reasoning benchmarks while using 3-4x less inference compute. The 501B-parameter MoE model (23B active) was trained on 23.8 trillion tokens and features a 1M token context window. It targets enterprises, public sector, and sovereign nations, with plans for 'AI factories' allowing customization on proprietary data. Reflection has raised ~$4.7B from backers including Nvidia and Sequoia, and signed compute deals worth over $7B (including a $6.3B deal with SpaceX) for Nvidia GB300 chips. Independent analyses place Beam around GLM-5.2 level, below DeepSeek V4 Flash on some benchmarks. Beam's weights (under Apache 2.0) and technical details will be released this month via hyperscalers and neoclouds. Additionally, Mistral released 'Mistral Large 4', a 1 trillion-parameter multimodal model. The article also covers the rise of personal AI assistants like Instinct (raising $1B in Series C) and Meta's Muse, alongside a16z's report on AI app adoption and public safety concerns.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.