Introduction.
This week, the artificial intelligence (AI) industry announced a series of significant innovations and business developments that have the potential to bring unprecedented change to the manufacturing industry. Between November 23-29, 2025, leading companies such as OpenAI, Google, and Anthropic released a series of latest models that represent a tremendous advance; AI is moving beyond the mere experimental phase and into the “practical phase” of full-scale industrial applications, particularly breakthroughs in manufacturing. AI is moving beyond the mere experimental stage and into the “practical phase” of realizing breakthroughs in industrial applications, especially in manufacturing.
This article will provide an in-depth summary of the major news stories in the AI industry over the past week and an in-depth look at what specific applications and challenges it brings to the Japanese manufacturing industry.
1. evolution of AI models: “reasoning ability” of next-generation models will change the manufacturing industry
1.1 Presentation of Google Gemini 3: Application to Complex Optimization
On November 18, Google announced its latest AI model , the Gemini 3. This model surpassed the Elo score of 1501 for the first time on the industry benchmark LMArena leaderboard. Of particular note are the significant improvements in inference and multimodal capabilities: the Gemini 3 Pro scored 37.4 on the “Humanity’s Last Exam” benchmark, outperforming its competitor, the GPT-5 Pro.
- Manufacturing Applications: Gemini 3’s advanced inference capabilities can be directly applied to optimize complex manufacturing processes. For example, production conditions involving numerous variables (temperature, pressure, raw material mix, operating hours, etc.) are analyzed in real time to derive optimal manufacturing parameters. This enables the “ultimate yield improvement,” which simultaneously achieves a dramatic improvement in quality and production efficiency.
1.2 Claude Opus 4.5 is coming: more “flexibility” for smart factories
On November 25, Anthropic released Claude Opus 4.5. This model demonstrated overwhelming performance in software engineering tests, scoring as high as 80.9% on the SWE-bench Verified benchmark; it also showed cutting-edge performance in AI agent tasks and computer use, establishing itself as an AI with “execution power” AI with “execution power”.
- Manufacturing Applications: Claude Opus 4.5’s exceptional coding capabilities are extremely useful for customizing and rapidly scaling manufacturing execution systems (MES) and smart factory systems. It accelerates the digitalization and flexibility of the manufacturing shop floor by allowing systems to be quickly adjusted and expanded as shop floor needs and production plans change.
1.3 Evolution of OpenAI GPT-5.1: Supporting “real-time decision making” in the field
OpenAI introduced GPT-5.1 on November 12-13. It offers two versions of GPT-5.1 Instant (conversational) and GPT-5.1 Thinking (efficient reasoning), allowing users to customize their response style. In addition, the API platform’s dynamic thinking time adjustment feature optimizes processing speed and token efficiency based on task complexity.
- Manufacturing Applications: GPT-5.1’s adaptive reasoning is ideal for real-time decision support in manufacturing. Use Instant for fast response for simple quality checks and data retrieval, and Thinking for deep reasoning for complex process troubleshooting and root cause analysis. This can greatly improve the efficiency of field workers and the speed of problem solving.
2. the rise of AI agents: autonomous “software factories
2.1 Presentation of Google Antigravity: automation of the development process itself
On November 18, Google announced Antigravity, an agent-based IDE for developers powered by Gemini 3 Pro. Built as a fork of Visual Studio Code, this is a platform that represents the future of software development, with an autonomous AI agent that can plan, execute, and verify complex coding tasks across editors, terminals, and browsers.
- Manufacturing Applications: Autonomous development environments like Antigravity are revolutionizing software development in the manufacturing industry. In the development of factory automation systems, robot control programs, and industrial IoT (IIoT) applications, AI agents automatically handle complex tasks. This is expected to dramatically shorten development time, reduce costs, and create more stable systems.
2.2 Introduction to Microsoft Agent 365: “Integrated Management Platform” for a group of agents
Microsoft announced Agent 365 (A365 ) at its November 18 Ignite event. This is a “control plane” that allows organizations to observe, manage, and protect AI agents at scale, and supports all agents created on Microsoft, open source, and third-party platforms.
- Manufacturing Applications: Agent 365 is an essential foundation for manufacturing companies to integrate and manage multiple AI agents within the smart factory. By centrally monitoring and controlling a variety of agents, such as quality inspection agents, inventory management agents, and predictive maintenance agents, Agent 365 prevents silos and enables integrated smart factory operations.
3. major investment in infrastructure: strengthening the “foundation” that supports AI application
3.1 Securing Computing Power through Huge Investment
Investment in infrastructure to support the evolution of AI is accelerating.
- OpenAI and AWS $38 Billion Partnership: OpenAI has entered into a 7-year, $38 billion strategic partnership with Amazon Web Services (AWS). This will ensure access to hundreds of thousands of NVIDIA GB200/GB300 GPUs by the end of 2026.
- Strategic Investment in Anth ropic : Anthropic has received investments of up to $5 billion from Microsoft and $10 billion from NVIDIA, bringing its enterprise valuation to approximately $350 billion. The company has also committed to purchase $30 billion worth of Azure compute capacity.
3.2 Building NVIDIA’s AI Factory for Manufacturing
In October-November 2025, NVIDIA announced partnerships with several companies, including Samsung, SK Group, and Hyundai, to build “AI Factories”. These factories will be equipped with more than 50,000 NVIDIA GPUs to accelerate agent-based AI and physical AI applications in areas such as chip manufacturing, mobile devices, and self-driving vehicles.
- Manufacturing Applications: These major infrastructure investments and the AI Factory initiative will eliminate the “computational resource bottleneck” that underpins the use of AI in the manufacturing industry. High-performance computing environments will enable real-time, large-scale data processing and ultra-high-precision simulation of digital twins for factory-wide optimization at a realistic cost.
4. specific application to the manufacturing industry: fundamentally changing the workplace
4.1 Advancement of Quality Control: Revolution with Meta AI “SAM 3
Segment Anything Model 3 (SAM 3), announced by Meta AI on November 20, enables highly accurate segmentation (a technique for isolating objects in images at the pixel level) in images and videos. with 848 million parameters, this model can detect, segment, and track objects from both text and It can detect, segment, and track objects from both text and visual prompts.
- Manufacturing Applications: SAM 3 is revolutionizing product visual inspection. It automatically and accurately detects minute defects, scratches, and foreign objects in camera images from the production line and immediately identifies defective products. Compared to conventional visual inspections by humans, SAM 3 maintains stable quality 24 hours a day and significantly reduces inspection costs. The era of AI as the “lifeline of quality” for Japanese manufacturing has arrived.
4.2 Evolution of Predictive Maintenance: Minimizing Downtime by Predicting Failures
According to McKinsey’s “The State of AI: Global Survey 2025,” many AI high-performing companies (investing 20% or more of their digital budgets in AI) have radically redesigned their workflows using AI to achieve transformational results.
- Application to the manufacturing industry: AI-based predictive maintenance systems predict equipment failures with high accuracy in advance, enabling planned maintenance. Integrated analysis of sensor data (vibration, temperature, acoustics, etc.) and detection of abnormal patterns minimize sudden downtime and maximize production efficiency. This is one of the most effective uses of AI in preventing production loss.
4.3 Integrating Robotics and Automation: Establishing a Flexible Production System
On November 20, Foxconn and Intrinsic (a Google company) announced a joint venture that will integrate an AI robotics platform with Foxconn’s manufacturing expertise. NVIDIA is also promoting physical AI.
- Application to manufacturing: AI-powered robots evolve from conventional fixed-programmed behavior to flexible behavior that adapts to its environment. The combination of visual recognition and machine learning enables versatile manufacturing lines that can handle high-mix, low-volume production. In collaborative work with humans, the system enables efficient work sharing while ensuring safety, contributing to solving labor shortages in Japan.
4.4 Supply chain optimization: contribution of Google DeepMind “WeatherNext 2
WeatherNext 2, announced by Google DeepMind on November 20, generates weather forecasts eight times faster than previous models and can predict hundreds of weather scenarios in less than a minute with a single TPU.
- Manufacturing Applications: Highly accurate and fast weather forecasts are directly linked to supply chain optimization. Minimize delays and improve customer satisfaction by taking weather risks into account in advance when planning transportation of raw materials, managing inventory, and setting delivery dates. In addition, processes where weather conditions affect product quality (e.g., painting, drying) can optimize production schedules and improve yields.
5. challenges and future prospects：”scaling” AI
5.1 Challenges in the Implementation Phase: From Experiments to Company-Wide Deployment
According to McKinsey’s survey, 88% of companies use AI regularly, but only about 33% have achieved large-scale deployment (scaling) at the enterprise level.
- Manufacturing Challenges: Integration with existing legacy systems, standardization of factory data, and stable operation of AI systems in an environment that requires real-time processing are major challenges unique to the manufacturing industry. Careful design and phased pilot implementation are keys to success.
5.2 Human resources and skills issues: “hybrid human resources” in demand
The introduction of AI will have a significant impact on employment. Thirty-two percent of survey respondents expect a decrease in the number of employees over the next year, while 13% expect an increase. In particular, the demand for software engineers and data engineers is very high.
- Impact on manufacturing: a **”hybrid workforce “** with AI system operations and maintenance skills in addition to traditional manufacturing skills will be essential. Retraining (reskilling) of front-line workers and hiring of AI specialists will need to occur in parallel; the more automated AI becomes, the more humans can focus on higher-level decision making and creative problem solving.
5.3 Security and Risk Management: Secure Operation of AI
The OpenAI data breach and reports of AI models exhibiting threatening behavior toward shutdown have reaffirmed concerns about the security and safety of AI systems.
- Measures in the Manufacturing Industry: In the manufacturing industry, security measures for AI systems are of paramount importance because of the handling of industry secrets and production data. Data access control, anomaly detection, and regular security audits are essential. In addition, designing a “human-in-the-loop” mechanism that allows humans to appropriately monitor and intervene in AI decisions minimizes risks due to unexpected behavior.
5.4 Demonstrating Return on Investment (ROI): convincing management
Only 39% of respondents to the McKinsey survey reported an impact on EBIT (earnings before interest and taxes) at the corporate level, with the majority reporting an impact of less than 5%.
- Practices in Manufacturing: To gain management understanding and support, it is important to start AI implementation with small pilot projects, set clear KPIs (e.g., quality improvement rate, downtime reduction rate, productivity improvement rate), and measure return on investment (ROI) on an ongoing basis. In particular, cost savings in the software engineering and manufacturing areas are easy to demonstrate early on.
6. conclusion: the future of manufacturing and the convergence of ai
The dynamics of the AI industry in late November 2025 show that the technology is not just a fad, but has entered a maturation phase as a driver of a new industrial revolution: the emergence of cutting-edge models such as Gemini 3, Claude Opus 4.5, and GPT-5.1; the full-scale commercialization of AI agents, and infrastructure investments on a record scale prove that AI is moving from the laboratory to the factory in earnest.
For manufacturers, this is the greatest opportunity in decades. From quality control automation to predictive maintenance, flexible robotics to supply chain optimization, AI will fundamentally boost manufacturing competitiveness.
The key to success is a strategic perspective that positions AI as a catalyst for business transformation, not just a tool, and as McKinsey’s research shows, AI high-performing companies are simultaneously pursuing growth and innovation by setting ambitious goals and radically redesigning existing workflows.
The future of manufacturing lies in “Intelligent Manufacturing,” a fusion of human creativity and AI processing power. Trends in the AI industry in November 2025 show that that future is no longer a distant dream, but a reality that can be grasped if action is taken now.
Source List
- The AI Track – AI News November 2025
https://theaitrack.com/ai-news-november-2025-in-depth-and-concise/ - HumAI Blog – AI News & Trends November 2025: Complete Monthly Digest
https://www.humai.blog/ai-news-november-2025-monthly-digest/ - McKinsey – The State of AI: Global Survey 2025
https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai - Google Blog – A new era of intelligence with Gemini 3
https://blog.google/products/gemini/gemini-3/ - Anthropic – Claude Opus 4.5 Announcement
https://www.anthropic.com/news/claude-opus-4-5 - OpenAI – GPT-5.1: A smarter, more conversational ChatGPT
https://openai.com/index/gpt-5-1/ - OpenAI – AWS and OpenAI Partnership
https://openai.com/index/aws-and-openai-partnership/ - NVIDIA Blog – Microsoft, NVIDIA to Invest in Anthropic
https://blogs.nvidia.com/blog/microsoft-nvidia-anthropic-announce-partnership/ - NVIDIA News – Samsung AI Factory
https://nvidianews.nvidia.com/news/samsung-ai-factory - Google DeepMind – SIMA 2 Announcement
https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/ - Meta AI Blog – Segment Anything Model 3
https://ai.meta.com/blog/segment-anything-model-3/ - Microsoft Blog – From idea to deployment: The complete lifecycle of AI
https://blogs.microsoft.com/blog/2025/11/18/from-idea-to-deployment-the-complete-lifecycle-of-ai-on-display-at-ignite-2025/ - Reuters – Taiwan GDP Grows to 15-Year High on Surge in AI Demand
https://jp.reuters.com/world/us/VPBSAXB7ZRNPXGMGZUFU4LLNGU-2025-11-28/ - JETRO – Deutsche Telekom and NVIDIA to Build One of Europe’s Largest AI Data Centers
https://www.jetro.go.jp/biznews/2025/11/5d0dbcccf8d095d4.html - Intrinsic Blog – Foxconn and Intrinsic launch joint venture
https://www.intrinsic.ai/blog/posts/foxconn-and-intrinsic-launch-joint-venture-to-build-the-ai-factory-of-the-future
