AI News on September 8st
1. Europe's Fastest Supercomputer Jupiter Launched in Germany
#Jupiter #Supercomputer #EuropeAI
Europe's first exascale supercomputer Jupiter was inaugurated in Germany, capable of performing at least one quintillion calculations per second. The system contains 24,000 Nvidia chips and occupies 3,600 square meters. German Chancellor Friedrich Merz stated this could help Europe catch up with US and China in the global AI race, though Europe currently lags with only 3 "notable" AI models compared to 40 from US and 15 from China.
2. Switzerland Releases Apertus Open-Source AI Model
#Apertus #OpenSource #SwitzerlandAI
Switzerland launched Apertus, a national open-source AI model built by EPFL, ETH Zurich and CSCS. Trained on 15 trillion tokens across 1,000+ languages, it features EU-aligned transparency standards and is accessible via Swisscom and Public AI platforms, representing Europe's growing open-source AI capabilities.
3. Microsoft Launches MAI Models as OpenAI Partnership Strains
#Microsoft #MAI #AIModels
Microsoft introduced MAI-Voice-1 and MAI-1-preview models, trained on approximately 15,000 Nvidia H100s with GB200 compute active. The launch comes as Microsoft's partnership with OpenAI shows signs of strain, with Microsoft developing more independent AI capabilities.
4. Google Expands AI Language Tools with Real-Time Translation
#GoogleTranslate #AI #LanguageLearning
Google added live translation and language learning tools to Google Translate, processing 1 trillion monthly translated words across Google services. The update includes AI-personalized practice and Gemini-based live conversations, significantly expanding its language capabilities.
5. Elon Musk's xAI Open-Sources Grok 2.5, Promises Grok 3 in Six Months
#xAI #Grok #OpenSource
Elon Musk's xAI open-sourced its Grok 2.5 model and pledged to release Grok 3 within six months. The move highlights xAI's commitment to transparency and enterprise benefits while intensifying the global AI competition among tech giants.
6. US Tech CEOs Commit Billions to AI at Trump Dinner
#Trump #TechCEOs #AIInvestment
Tech CEOs including Meta's Zuckerberg and Apple's Cook attended a White House dinner with President Trump, promising massive AI investments. Zuckerberg committed $600 billion by 2028, while Cook announced $100 billion for domestic manufacturing. Trump emphasized energy and permitting reforms to support AI data center growth.
7. Meta Partners with Midjourney for AI Image Technology
#Meta #Midjourney #AIImaging
Meta partnered with Midjourney to license AI image and video technology, expanding its creative AI portfolio. The deal includes licensing Midjourney's V7 and V1 models as Meta pursues its superintelligence ambitions and enhances its content creation capabilities.
8. EU Commits €200 Billion for AI Investment Plan
#EU #AIInvestment #DigitalEurope
The European Union announced a €200 billion AI investment initiative, including €20 billion for European AI super factories to train complex, large-scale AI models. The investment aims to boost EU competitiveness in global AI race despite current trailing position behind US and China.
9. OpenAI and Anthropic Begin Cross-Model Safety Evaluations
#OpenAI #Anthropic #AISafety
OpenAI and Anthropic opened APIs for cross-lab audits, evaluating AI models for safety, alignment and misuse resistance. The collaboration examines sycophancy, misuse cooperation, jailbreak resistance differences, and refusal-helpfulness trade-offs between their respective AI systems.
10. Google Debuts Gemini 2.5 Flash Image for Advanced Editing
#Gemini #Google #ImageAI
Google launched Gemini 2.5 Flash Image featuring narrative storytelling, style transfer, and real-world reasoning capabilities. Priced at $0.039 per image with advanced safety filters, the model represents Google's latest advancement in visual AI technology.