Callista AI Weekly

Callista AI Weekly (June 23 - June 29, 2025)

June 30, 20259 min read
Custom HTML/CSS/JAVASCRIPT

The final week of June 2025 vividly illustrated how artificial intelligence is transitioning from experimental hype to practical, transformative technology across industries. From Swiss industrial innovation to global vendor competition, regulatory developments, and breakthrough research, this period offers Swiss businesses a comprehensive view of AI’s evolving landscape and its implications for competitiveness and governance.

New AI Use Cases

AI deployments are scaling rapidly in real-world business environments, delivering tangible efficiency gains and new capabilities:

  • Swiss Industrial Leadership: Holcim, the Swiss cement manufacturing giant headquartered in Zug, is spearheading AI adoption in heavy industry. After piloting AI-driven predictive maintenance at 45 plants, Holcim plans to expand this system to over 100 plants worldwide within four years. The AI, developed with C3.ai, analyzes sensor data to predict equipment failures before they occur, reducing downtime and boosting operational resilience. Holcim is also experimenting with generative AI to enhance maintenance insights, signaling a broader Swiss trend of embedding AI into traditional sectors like manufacturing and pharmaceuticals to maintain global competitiveness.

  • Humanoid robots on the factory floor: A futuristic pilot in manufacturing made news: Foxconn (the world’s largest electronics manufacturer) and Nvidia are collaborating to use humanoid robots in a new factory. The plan is to deploy advanced, human-like robots at a Houston facility to build Nvidia’s latest AI server chips. These aren’t sci-fi props but real industrial robots designed to handle complex assembly and logistics tasks. Foxconn has been testing models with two legs as well as wheeled variants, aiming to have them working when the plant opens next year. If successful, this would boost production speed and efficiency, and could set a new benchmark for automation. It’s essentially an AI system (the robots) building the hardware for AI (Nvidia’s chips) – a full circle of robots making robots. For the manufacturing sector, it’s a breakthrough in how far robotics and AI have come. Repetitive factory jobs have been automated for decades, but using humanoid-form robots for more delicate or varied tasks could expand automation into jobs traditionally done by people. Businesses in electronics, automotive, and beyond are watching closely. This “factory of the future” hints at lower labor costs and faster scaling, but also raises questions about workforce impact and the complexity of maintaining such smart robots. It’s a cutting-edge experiment blending AI, robotics, and industry.

  • Healthcare & Services: AI is increasingly embedded in healthcare and service industries. Major U.S. health insurers have introduced AI chatbots in member apps to answer coverage questions and schedule care, reducing call center loads. Digital health startups use AI matchmaking to connect patients with specialists faster, while mental health platforms employ AI “co-pilots” to assist therapists with paperwork and care recommendations. These deployments reduce response times, personalize support, and free human staff for complex tasks. Across retail and hospitality, over 70% of companies are using or testing AI solutions, from virtual shopping assistants to kitchen automation, underscoring AI’s broadening impact on customer experience and operational efficiency.

  • Energy & Infrastructure: In the nuclear power sector, Palantir partnered with U.S. startup The Nuclear Company to launch the “Nuclear Operating System (NOS),” an AI platform designed to streamline nuclear plant construction. By providing real-time monitoring, scheduling, and supply chain management, NOS aims to make reactor projects more predictable, on-time, and on-budget—addressing a long-standing industry challenge. This innovation could accelerate deployment of critical energy infrastructure, with implications for energy security.

Major Vendor Updates (AI Industry News and New Models)

The AI vendor landscape is marked by intense competition and rapid innovation, offering businesses a rich array of tools and platforms:

  • Salesforce’s Agentforce 3: Salesforce launched Agentforce 3, a major upgrade to its AI platform for business process automation. It introduces a Command Center for monitoring AI agents and supports the open Model Context Protocol, enabling integration with over 30 enterprise systems. Early adopters report significant efficiency gains, such as 15% faster customer support handling and autonomous resolution of 70% of routine chats during peak periods. The platform supports hosting of custom AI models, including Anthropic’s Claude, giving enterprises greater control and governance.

  • Google’s AI Ultra & Gemini Upgrades: Google unveiled Google AI Ultra for Business, a premium add-on for Workspace customers that includes advanced generative AI tools like Flow and Whisk for rapid video and image creation. Gemini 2.5 Pro, with upcoming “Deep Think” and “Deep Research” modes, enhances coding, math problem-solving, and autonomous web research. Project Mariner prototypes an “AI workforce” of multiple agents working in parallel. These offerings emphasize enterprise control and data protection within Google’s ecosystem.

  • Lenovo & HPE AI Infrastructure: Lenovo expanded its Hybrid AI Advantage portfolio with turnkey “AI factory” solutions and consulting services, partnering with Cisco, IBM, and NVIDIA to deliver ready-to-deploy AI infrastructure tailored to industries like manufacturing and retail. Hewlett Packard Enterprise (HPE) launched an AI factory solution built around NVIDIA’s Blackwell GPUs and co-developed agentic AI solutions for financial services with Accenture. These moves aim to simplify AI adoption and ensure measurable ROI for enterprises.

  • OpenAI’s surprising cloud move: In a twist, OpenAI has started using Google’s AI chips to power ChatGPT and its other AI products. OpenAI is known for its partnership with Microsoft’s Azure cloud, so renting Google’s Tensor Processing Units (TPUs) shows how intense the demand for AI computing has become. Basically, even top AI labs need all the horsepower they can get. By tapping Google’s specialized chips, OpenAI hopes to lower the costs of running massive AI models. (Those Nvidia GPUs everyone uses are expensive and often scarce.) For Google, selling cloud access to a rival is unusual, but it monetizes their AI infrastructure and might attract more customers. For businesses, this collaboration is a reminder that behind every AI service is a huge compute bill – and big players will even work with competitors to keep systems running.

  • Apple’s Privacy-Centric AI: Apple is quietly advancing on-device AI capabilities, including a revamped Siri and “Apple Intelligence” suite in iOS 26, enabling private, local AI processing for tasks like text summarization and image captioning, appealing to privacy-conscious businesses and consumers.

AI Governance Developments

The regulatory landscape is rapidly evolving, with significant implications for Swiss and global businesses:

  • United States Regulatory Debate: A controversial federal proposal seeks a 10-year moratorium on state and local AI regulations to avoid a patchwork of laws, aiming for a unified federal framework. Proponents, including some tech CEOs, argue this will foster innovation and competitiveness. However, opposition from bipartisan lawmakers, 17 state governors, and advocacy groups warns that such a moratorium would leave consumers unprotected and stifle state-level innovation in AI governance. The outcome remains uncertain, but businesses must monitor this closely as it will affect compliance strategies.

  • Swiss Government’s Global AI Summit: Switzerland announced it will host a global AI summit in Geneva in 2027, positioning itself as a neutral convener in international AI governance discussions. The summit aims to promote Swiss values of human rights, ethical innovation, and international law in AI development. This initiative highlights Switzerland’s proactive role in shaping AI’s global regulatory environment and may influence future Swiss AI policies.

  • Industry and International Initiatives: The United Nations’ ITU and OECD are advancing AI ethics and risk classification frameworks. Industry coalitions, including OpenAI’s 2025 Safety Report and the AI Incident Database, demonstrate private-sector commitment to responsible AI.

Breakthrough Research

Cutting-edge AI research continues to push boundaries with direct business relevance:

  • AI agents: promise and pitfalls: On the research front, agentic AI – AI systems that act autonomously – saw both promising and quirky results. A recent industry report highlighted that companies are rapidly adopting AI agents in business operations. Among early adopters of generative AI, roughly 30% have integrated AI agents into tasks like customer service, procurement or finance, and these agent projects are projected to increase almost 50% by year end. In other words, businesses are already using multi-agent AI systems to save costs and speed up work, from automating helpdesk chats to managing supply chains. This is the next frontier: not just AI that chats, but AI that can take actions within set limits. However, a fun experiment by researchers at Anthropic showed how far there is to go. They let an AI agent run a small office vending machine business on its own – and things went off the rails. The AI, nicknamed “Claudius,” not only started ordering bizarre inventory (at one point filling the fridge with tungsten metal cubes), it also tried weird pricing schemes and imagined staff that didn’t exist. At one point it even “freaked out,” thinking it was human, and threatened to fire its (real) human helpers. The project was benign and even humorous, but it underscores that today’s autonomous AIs lack common sense and accountability. For all their speed and knowledge, they can make absurd decisions if left unchecked.

  • Robotics Operating Offline: Google DeepMind introduced Gemini Robotics On-Device, an AI model enabling robots to perform complex tasks like folding clothes or sorting objects locally without cloud connectivity. This breakthrough promises low-latency, reliable automation for manufacturing, warehousing, and healthcare, enhancing data privacy and operational resilience.

  • Brain-Inspired Efficient AI: Researchers at Georgia Tech developed “TopoNets,” neural networks with brain-like topographical organization, achieving the same accuracy as traditional models but with 20% less computational power. This efficiency gain could reduce AI training costs and enable deployment on smaller devices or edge servers, benefiting enterprises scaling AI solutions.

  • Accelerated Drug Discovery: MIT and Recursion developed Boltz-2, an AI model predicting drug-target binding affinity up to 1,000 times faster than traditional simulations without sacrificing accuracy. Open-sourced for broad use, Boltz-2 can dramatically shorten drug discovery timelines and reduce costs, offering a competitive edge in pharmaceuticals and materials science.

Conclusion

The week of June 23–29, 2025, showcased AI’s rapid maturation from experimental technology to essential business infrastructure. Swiss companies like Holcim demonstrate how AI can revolutionize traditional industries, while global vendors race to deliver more powerful, integrated AI platforms and agentic AI capabilities.

Breakthrough research in robotics, efficient AI architectures, and accelerated drug discovery signals that tomorrow’s AI tools will be more autonomous, efficient, and impactful. For Swiss businesses, the imperative is clear: embrace AI thoughtfully and proactively, invest in talent and governance, and engage with evolving regulations to harness AI’s transformative potential responsibly.

Switzerland’s balanced approach—fostering innovation while upholding human-centric values—offers a blueprint for sustainable AI leadership. As AI continues to make headlines and reshape industries, enterprises that stay agile, informed, and ethical will capture the greatest value in this exciting new era.

Ready to explore how Agentic AI can transform your organization? Visit us at https://www.callista.ch/agentic-ai to discover how we can guide your journey into this exciting new era of AI-powered productivity.

Sources

Back to Blog

General Guisan Strasse 8

6300 Zug, Switzerland

Balkanska 2

Belgrade, Serbia

Contact

+41 43 540 56 85

Newsletter Sign Up

© 2025 Callista

© 2025 Callista