AI Tech Weekly

## Overview of the Week in AI (November 3–9, 2025)

Between November 3 and 9, 2025, the AI sector experienced a flurry of strategic partnerships, groundbreaking hardware launches, and regulatory milestones that underscore the rapid evolution and commercialization of artificial intelligence. Major cloud providers and chipmakers doubled down on infrastructure commitments; tech giants formed new alliances to integrate advanced models into consumer products; and industry events highlighted AI’s expanding role in traditionally non‐tech domains. These developments reflect both the insatiable demand for compute and the strategic imperative for companies to secure bespoke AI capabilities. ([reuters.com](https://www.reuters.com/business/retail-consumer/openai-amazon-strike-38-billion-agreement-chatgpt-maker-use-aws-2025-11-03/?utm_source=openai))

## OpenAI’s Landmark $38 B Cloud Deal with AWS

On November 3, 2025, OpenAI and Amazon Web Services (AWS) announced a seven-year, $38 billion agreement granting OpenAI immediate access to hundreds of thousands of Nvidia GPUs and UltraServer clusters across AWS’s global data centers. This deal marks OpenAI’s first significant move beyond its longstanding partnership with Microsoft Azure, following a corporate restructuring that removed Microsoft’s right of first refusal for future cloud contracts. Under the pact, full capacity is expected by the end of 2026, with expansion options into 2027 and beyond. The partnership bolsters AWS’s position against rivals Microsoft and Google and underscores the industry’s mounting appetite for high-performance compute to train and deploy next-generation AI systems such as ChatGPT and Sora. ([reuters.com](https://www.reuters.com/business/retail-consumer/openai-amazon-strike-38-billion-agreement-chatgpt-maker-use-aws-2025-11-03/?utm_source=openai))

## Microsoft and NVIDIA Launch UK “Agentic Launchpad”

On November 4, 2025, Microsoft, in collaboration with NVIDIA and WeTransact, unveiled the Agentic Launchpad program to accelerate autonomous AI (or “agentic AI”) startups in the United Kingdom and Ireland. Backed by Microsoft’s £22 billion (US$30 billion) UK AI investment, the initiative invites early-stage companies building AI agents capable of independent decision-making to apply by November 28. Selected participants receive Azure cloud credits, technical mentorship from Microsoft and NVIDIA engineers, NVIDIA Inception membership, and go-to-market support including product marketing and marketplace listings. The program deepens the Microsoft–NVIDIA partnership established by last year’s GenAI Accelerator and aims to seed real-world agentic workloads across enterprise sectors in 2026. ([itpro.com](https://www.itpro.com/business/microsoft-and-nvidia-are-teaming-up-again-to-support-uk-startups?utm_source=openai))

## Apple to Integrate Google’s Gemini AI into Siri

Amid escalating competition in voice assistants, Apple confirmed on November 5, 2025, that it is finalizing a deal to pay Google approximately $1 billion annually for access to a custom 1.2 trillion-parameter Gemini model to power a major overhaul of Siri. The arrangement, reported by Bloomberg and detailed by multiple outlets, will allow Siri to tap Gemini for summarization and multi-step planning tasks while Apple continues developing its own LLMs. The Gemini model will run on Apple’s Private Cloud Compute infrastructure, preserving user data privacy. This “stopgap” partnership reflects Apple’s urgency to match advances by Google Assistant and Amazon Alexa and is slated to launch in spring 2026 as part of iOS 26.4’s “Apple Intelligence” features. ([reuters.com](https://www.reuters.com/business/apple-use-googles-ai-model-run-new-siri-bloomberg-news-reports-2025-11-05/?utm_source=openai))

## Google Cloud Debuts Ironwood TPUs and Arm-Based Axion VMs

On November 6, 2025, Google Cloud announced the general availability of its seventh-generation Tensor Processing Unit, codenamed Ironwood, alongside new Arm-based Axion virtual and bare-metal machines. Ironwood offers up to 10× the peak performance of TPU v5p and over 4× better performance per chip than its predecessor for both training and inference, delivering 4,614 FP8 TFLOPS per chip and scaling into pods of 9,216 accelerators (42.5 ExaFLOPS). Early customers like Anthropic have reported significant cost and speed benefits. The launch completes Google’s custom silicon portfolio for its AI Hypercomputer and signals a more aggressive challenge to NVIDIA’s GPU dominance in AI workloads. ([tomshardware.com](https://www.tomshardware.com/tech-industry/artificial-intelligence/google-deploys-new-axion-cpus-and-seventh-gen-ironwood-tpu-training-and-inferencing-pods-beat-nvidia-gb300-and-shape-ai-hypercomputer-model?utm_source=openai))

## AI at the Core of AAPS PharmSci 360 Conference

The AAPS PharmSci 360 meeting, which began on November 9 in San Antonio, TX, placed artificial intelligence at the heart of its five comprehensive tracks spanning drug discovery, formulation, bioanalysis, and manufacturing. According to Pharmaceutical Technology, AI-driven tools—ranging from in silico modeling and LLM-based report generation to ML-powered analytics—are now integral to accelerating pharmaceutical R&D and ensuring regulatory compliance. The conference’s keynote and panel sessions underscored how AI is reshaping every stage of pharma development, highlighting both opportunities for efficiency gains and challenges related to validation and safety. ([pharmtech.com](https://www.pharmtech.com/view/pharmtech-weekly-news-roundup-week-of-november-3-2025?utm_source=openai))

## SoftBank and OpenAI Launch “SB OAI Japan” Joint Venture

On November 5, 2025, the SoftBank Group and OpenAI officially launched SB OAI Japan GK, a 50:50 joint venture aimed at marketing “Crystal intelligence”—a packaged enterprise AI solution integrating OpenAI’s latest models with customized implementation and integration services—to Japanese enterprises in 2026. The venture will leverage SoftBank’s domestic network and operational expertise alongside OpenAI’s technology to build AI agents tailored for corporate management and operations. SoftBank has already deployed Crystal internally, creating approximately 2.5 million custom GPTs for workforce productivity, and will use early insights to refine offerings for external clients. ([group.softbank](https://group.softbank/en/news/press/20251105?utm_source=openai))

## Microsoft to Ship 60,000 Nvidia AI Chips to the UAE

In a notable regulatory development, Microsoft announced on November 3, 2025, that it will ship over 60,000 of Nvidia’s most advanced AI GPUs—including GB300 Grace Blackwell accelerators—to data centers in the United Arab Emirates under U.S. Commerce Department-approved licenses. This shipment is part of a broader $15.2 billion investment plan by Microsoft in UAE AI and cloud infrastructure through 2029. The move underscores both the UAE’s ambition to become a global AI hub and the U.S. government’s willingness to permit controlled export of cutting-edge AI hardware under stringent safeguards. ([reuters.com](https://www.reuters.com/world/china/microsoft-invest-over-15-billion-uae-secures-us-export-licenses-ai-chips-2025-11-03/?utm_source=openai))