AI Industry: google and amd Moves

Google recently introduced Private AI Compute, a new cloud platform powered by its Gemini AI models, on November 11 and 12, 2025. This platform aims to deliver the same high level of security and privacy as processing AI directly on a device, allowing users to access the full power of Gemini cloud models while keeping their personal data private from Google and others. The technology, which relies on AMD-based hardware, custom TPUs, and Titanium Intelligence Enclaves, uses strong encryption and encrypted connections to ensure a "zero access assurance," meaning Google itself cannot access user data. This initiative mirrors privacy efforts from companies like Apple with its Private Cloud Compute and Meta. Initially, Private AI Compute will roll out to Pixel devices, including the Pixel 10 and the Recorder app, before expanding to Google's business products to help sectors like healthcare and finance meet privacy regulations such as GDPR. An audit by NCC Group confirmed the system offers strong protection, though Google is addressing some identified issues. Beyond this, Google is making a significant investment of 5.5 billion euros in Germany between 2026 and 2029 to expand its cloud and AI services, including new data centers in Dietzenbach and Hanau, which will support AI tools like Vertex AI and Gemini for businesses, with a focus on environmental sustainability. The broader discussion around AI also highlights critical ethical and security considerations. The National Academy of Medicine released a new AI Code of Conduct for healthcare, co-authored by Penn professor Kevin B. Johnson, to ensure AI tools are safe, fair, and helpful, addressing concerns like bias and data breaches. Red Hat Inc. is championing "zero-trust AI" and confidential computing, as explained by experts Anjali Telang and Roman Zhukov, to secure AI systems by strictly controlling access and encrypting data even during use. The real-world impact of AI was starkly illustrated on November 11, 2025, when an AI-generated photo of a fake fire at Bellaire High School in Texas caused widespread panic, despite police and fire officials confirming no actual blaze. Meanwhile, leading AI scientist Yoshua Bengio expressed increased optimism about controlling superintelligence, planning to build AI systems that function as "smart encyclopedias" without their own goals or consciousness. Conversely, Wikipedia co-founder Jimmy Wales believes big tech companies should compensate Wikipedia for using its content to train their AI models, citing concerns about accuracy and the impact of AI bots. In the business world, Suprema launched BioStar X on November 12, 2025, a new unified AI security platform for large enterprises, offering centralized control and AI-driven threat detection. Investment firm Alpine Macro, through chief equity strategist Nick Giorgi, suggests that utility stocks like Vistra and PG&E are a better investment for the AI trend than technology stocks, given the massive electricity demand from new data centers, with the utilities sector already gaining over 17% in the S&P 500 this year. Additionally, the US law firm Honigman partnered with Hotshot to provide practical AI training for its clients, using real-life scenarios to enhance legal professionals' understanding of AI tools.

Key Takeaways

• Google launched Private AI Compute on November 11 and 12, 2025, a new cloud platform utilizing Gemini AI models to offer on-device level privacy for cloud-based AI processing.
• Private AI Compute uses AMD-based hardware, custom TPUs, and strong encryption, ensuring Google cannot access user data, similar to Apple's Private Cloud Compute.
• The platform will first be available on Pixel devices and then expand to Google's business products, targeting sectors like healthcare and finance to meet privacy regulations.
• Google is investing 5.5 billion euros in Germany between 2026 and 2029 to expand its cloud and AI services, including new data centers, supporting tools like Vertex AI and Gemini.
• The National Academy of Medicine released an AI Code of Conduct for healthcare, focusing on safety, fairness, and preventing bias in AI applications.
• Red Hat Inc. promotes "zero-trust AI" and confidential computing to enhance AI system security by strictly controlling data access and encrypting data during use.
• An AI-generated photo of a fake fire at Bellaire High School on November 11, 2025, caused widespread panic, highlighting the real-world impact of AI misuse.
• AI scientist Yoshua Bengio is more optimistic about controlling superintelligence, aiming to build AI systems that act as "smart encyclopedias" without self-interest or consciousness.
• Wikipedia co-founder Jimmy Wales believes AI companies should pay for using Wikipedia's content to train their large language models, citing concerns about accuracy.
• Alpine Macro suggests utility stocks, such as Vistra and PG&E, are a better investment for the AI trend due to the significant electricity demand from new data centers.

Google launches Private AI Compute for secure AI

Google introduced Private AI Compute, a new cloud platform powered by its Gemini AI models. This platform offers the same high level of security and privacy as processing AI directly on a device. It lets users access the full power of Gemini cloud models while keeping their personal data private from Google and others. This move, announced on November 11, 2025, is similar to Apple's Private Cloud Compute, focusing on AI safety and user data protection.

Google's Private AI Compute boosts business AI privacy

Google introduced Private AI Compute, a new platform that combines powerful cloud AI with strong privacy. It processes sensitive data in a secure cloud space, which even Google cannot access. This technology will first come to Pixel devices and then to Google's business products. It aims to help companies in fields like healthcare and finance use AI more safely, meeting privacy rules like GDPR. Experts are hopeful but want to see independent checks to confirm its security.

Google unveils Private AI Compute for cloud privacy

Google launched Private AI Compute on November 12, 2025, a new technology for secure AI processing in the cloud. This platform uses Gemini models and provides privacy similar to on-device processing. It relies on AMD-based hardware and strong encryption to keep user data private, even from Google. An audit by NCC Group found some issues Google is working to fix, but confirmed the system offers strong protection. This move by Google is similar to privacy efforts from Apple and Meta.

Google introduces Private AI Compute like Apple

Google launched Private AI Compute, a new cloud system that brings on-device AI privacy to the cloud. It uses Google's advanced Gemini models and strong privacy features, similar to Apple's Private Cloud Compute. The system runs on Google's own tech, including custom TPUs and Titanium Intelligence Enclaves, ensuring data is processed in a secure, sealed environment. Encrypted connections and a "zero access assurance" mean Google itself cannot access user data. This will improve AI features on devices like the Pixel 10 and the Recorder app.

Google's Private AI Compute brings device privacy to cloud

Google introduced Private AI Compute, promising on-device level privacy for its Gemini AI in the cloud. This new service aims to deliver powerful AI experiences for complex tasks that phones cannot handle alone. It uses Google's own custom Tensor Processing Units and strong encryption to keep user data secure. Google states that no one, including the company, can access the data processed within this private cloud. This technology could bring advanced AI features to Pixel phones.

Google launches Private AI Compute for cloud

Google, part of Alphabet, has launched Private AI Compute. This new service uses Google's Gemini AI models to offer enhanced privacy for cloud-based AI processing. It aims to provide secure AI capabilities while protecting user data.

National Academy of Medicine guides AI in health care

The National Academy of Medicine released a new AI Code of Conduct to guide how artificial intelligence is used in health care. Penn professor Kevin B. Johnson, a co-author, explained that this framework aims to make sure AI tools are safe, fair, and truly helpful. It seeks to prevent problems like bias and unequal access to care. The framework includes six key commitments, such as advancing humanity and ensuring fairness. This guide promotes a shared system of oversight involving federal agencies and local responsibility.

AI in healthcare raises many ethical concerns

Using artificial intelligence and machine learning in healthcare brings many ethical concerns, even though it can improve patient care. Key ethical rules like doing good, doing no harm, and being fair are challenged by AI's potential for data breaches and unfair algorithms. AI can access private patient information, so strong security and patient consent are crucial. Algorithms can be biased if they learn from incomplete data, which might lead to wrong diagnoses for some groups. While AI helps predict drug reactions, doctors must still be careful about false alarms.

Red Hat champions zero trust for AI security

Red Hat Inc. is leading the way in "zero-trust AI" to make artificial intelligence systems more secure. This method means always checking who or what is accessing data and strictly controlling access to protect sensitive information. Red Hat experts Anjali Telang and Roman Zhukov explained that these security ideas can be used for AI, just like they are for other computer systems. They also talked about confidential computing, which keeps data encrypted even when it is being used. This helps ensure data remains private and safe from unauthorized access.

Fake AI fire photo causes panic at Bellaire High

On November 11, 2025, an AI-generated photo showing a fake fire at Bellaire High School in Bellaire, Texas, caused widespread panic. A real fire alarm had gone off earlier, leading to an evacuation and a message from the principal about smoke. The fake image then spread quickly on social media and messaging apps, making many parents worried. Police and the Houston Fire Department confirmed there was no actual fire. The principal later clarified the situation, explaining that conflicting messages had created confusion.

Yoshua Bengio more hopeful about controlling AI

Yoshua Bengio, a top AI scientist, feels more positive about controlling superintelligence than he did two years ago. He plans to build AI systems that act like "smart encyclopedias" without their own goals or self-interest. Bengio believes that AI without consciousness or desires will not try to control humans. He has even started a new nonprofit group to work on this type of AI. This new view comes as other AI leaders have different ideas about the dangers of advanced AI.

Google invests 5.5 billion in German cloud AI

Google is investing 5.5 billion in Germany between 2026 and 2029 to grow its cloud and AI services. This plan includes building new data centers in Dietzenbach and Hanau, which will support AI tools like Vertex AI and Gemini for businesses. The investment also focuses on being environmentally friendly, aiming for 85% carbon-free energy by 2026 and reusing waste heat for local homes. Google is also working to restore natural areas and offer training programs to improve digital skills in Germany. This major investment connects technology growth with sustainability and workforce development.

Wikipedia founder wants AI companies to pay

Jimmy Wales, who co-founded Wikipedia, believes big tech companies should pay for using Wikipedia's information to train their AI. He noted that Wikipedia's content is a major part of the data used by large language models. While the content is free to use, Wales is concerned about the impact of AI bots on Wikipedia's systems. He also doubts that AI models can create accurate encyclopedia content because they often make up facts. Wales suggests AI should learn to say "I'm not sure" when it does not know something.

Hotshot and Honigman train lawyers on AI

The US law firm Honigman teamed up with Hotshot to offer an AI training program for its clients, including in-house lawyers. Clients wanted to learn how to use AI tools in practical ways, but creating custom training for each was difficult. So, at their Innovation Symposium in Chicago, they held a hands-on workshop where clients could directly use AI tools. Hotshot created all the training materials, letting Honigman lawyers guide the sessions. The workshop used real-life examples, such as a phishing attack and a CCPA law change, to give participants useful, immediate experience.

Utilities stocks are better for AI investing

A research firm, Alpine Macro, suggests that utility stocks are a better investment for the artificial intelligence trend than technology stocks. Nick Giorgi, their chief equity strategist, believes utilities are entering a period of faster growth because of the huge electricity demand from new data centers. The utilities sector has already performed well this year, gaining over 17% in the S&P 500. Giorgi recommends buying utility stocks and selling energy stocks, naming Vistra and PG&E as top choices.

Suprema launches BioStar X AI security platform

Suprema launched BioStar X on November 12, 2025, a new unified AI security platform for large businesses. This platform acts as a central control system, showing interactive maps, AI video, alarms, and access logs all at once. BioStar X uses AI to spot important events like intrusions or people falling, helping to prevent threats. It also allows for detailed access rules and can connect with other security systems. With strong features like AES-256 encryption, BioStar X can handle many devices and users as a company grows.

Sources

• Google Unveils Private AI Compute, a Gemini-run Cloud Platform Promising Security
• Google’s “Private AI Compute” Could Reshape How Businesses Use Artificial Intelligence
• Google Launches 'Private AI Compute' — Secure AI Processing with On-Device-Level Privacy
• Google reveals its own version of Apple’s AI cloud
• Google's Private AI Compute promises good-as-local privacy in the Gemini cloud
• Google introduces Private AI Compute for privacy in cloud
• National Academy of Medicine Issues Code of Conduct to Guide Health Care’s AI Revolution
• Ethical Considerations of Artificial Intelligence Use Abound
• Zero-Trust AI: Red Hat’s guide to securing AI workloads
• AI-generated photo of fake fire at Bellaire High School, prompts panic: police
• Am More Optimistic About Our Ability To Control Superintelligence Than 2 Years Ago: Yoshua Bengio
• Google’s €5.5B Germany investment reshapes enterprise cloud and AI
• Wikipedia founder Wales wants Big Tech to pay for training AI
• Case Study: Hotshot’s AI Training Program For Clients
• These stocks in a 'sweet spot' are a better way to play AI than tech, says research firm
• Suprema launches BioStar X, a unified AI security platform for enterprise-grade control