Introduction.
Between January 10 and 17, 2026, the AI industry reached an important turning point. A series of news releases from around the world during this period indicate that AI has evolved from an “interactive tool” to a “physical entity” and has the potential to fundamentally change the foundations of all industries, including manufacturing. This paper summarizes the major AI-related news of the past week and discusses its potential application to the manufacturing industry in particular.
1. the rise of physical AI: CES 2026 shows a new era
Robotics Revolution Driven by NVIDIA
On January 5, 2026, at CES 2026 in Las Vegas, NVIDIA CEO Jensen Huang declared, “The ChatGPT moment in robotics has arrived. This announcement marked the tipping point at which AI will move from virtual space to the physical world in earnest.
NVIDIA has announced a new set of open models for physical AI. Of particular note are the “Cosmos” series, which has the ability to understand the world and make inferences and action plans, and the Isaac GR00T N1.6, which is dedicated to humanoid robots. These models show the way for robots that have been limited to a single task and difficult to program to evolve into general-purpose robots with reasoning capabilities.
The Jetson T4000 module with Blackwell architecture was also announced, offering four times the energy efficiency and AI computing power of the previous generation for $1,999 (in units of 1,000).
Acceleration of practical application by global companies
Global companies such as Boston Dynamics, Caterpillar, Franka Robotics, LG Electronics, and NEURA Robotics announced a new generation of robots that leverage NVIDIA’s technology stack. Caterpillar, in particular, announced that it is expanding its cooperation with NVIDIA to bring AI and autonomy to construction and mining machinery, providing a concrete example of AI applications in heavy industry.
2. evolution of large-scale language models and increased competition
China DeepSeek’s Breakthrough
On January 9, Chinese AI startup DeepSeek announced plans to release its next-generation AI model, V4, in mid-February. The model specializes in coding capabilities and has outperformed Anthropic’s Claude and OpenAI’s GPT series in internal tests, according to the company.
It should be noted that DeepSeek has achieved a breakthrough in its ability to handle extremely long coding prompts. This is a significant advantage for developers working on complex software projects. DeepSeek also proposed a new AI learning method, Engram, which shows how large models can be trained on lower-performance chips.
OpenAI’s Stargate Project Expansion
OpenAI has announced five new data center sites in the $500 billion Stargate Project, a joint initiative with Oracle and SoftBank. The project will make major investments in AI infrastructure over the next four years to build the foundation for next-generation AI development.
On January 8, a partnership with SB Energy was also announced, and a structure is in place to support the expansion of AI infrastructure from an energy supply perspective.
Google’s Gemini Evolution and Personalization
On January 14, Google introduced a new feature called Personal Intelligence in its Gemini app. This is the ability to customize the AI assistant’s responses based on the user’s photos, emails, and other data, signaling that AI can now provide more personalized service.
In addition, Gemini integration is in full swing in Gmail, and the “AI Overviews” feature allows users to summarize huge email threads and instantly answer questions in natural language. This has the potential to dramatically improve the efficiency of business communications.
3. memory chip shortage and AI infrastructure challenges
Surge in HBM demand and prices
According to a January 14 article in Reuters, an unexpected side effect of the AI boom is a global shortage of memory chips. The price of high bandwidth memory (HBM) has more than doubled since February 2025, which has led to higher prices for consumer products such as smartphones and PCs.
SK Hynix and Samsung have already sold out of their 2026 capacity and are beginning to book for 2027; OpenAI’s Stargate project alone will require 900,000 wafers per month by 2029, which is about twice the current global HBM production This is about twice the current global HBM production.
This situation reflects the intensifying competition for AI development and the rapid increase in infrastructure investment, while at the same time suggesting that supply constraints may become a bottleneck for future AI diffusion.
4. manufacturing applications: 5 major trends for 2026
Trend 1: Autonomous production scheduling with agent-based AI
IDC predicts that by 2026, more than 40% of manufacturers with production scheduling systems will upgrade to AI-driven capabilities and begin enabling autonomous processes. This will enable the optimization of complex production planning with minimal human intervention.
Agent-based AI can make integrated decisions to optimize demand forecasting, inventory management, and equipment utilization, and adjust production plans in real time. This enables dynamic and adaptive production management, a departure from traditional fixed scheduling systems.
Trend 2: Evolution of Digital Twin and Simulation Technology
NVIDIA’s Cosmos model enables physics-based synthetic data generation and evaluation of robotic policies. Manufacturing companies will be able to utilize this technology to build a digital twin of their production lines and simulate various scenarios before the actual equipment is put into operation.
In addition, a new open source framework called Isaac Lab-Arena enables large-scale robot policy evaluation and benchmarking. This allows for full validation in a simulation environment prior to the introduction of a robot in a manufacturing site, greatly improving the success rate of implementation.
Trend 3: Integration of human resource development and AI
As Fortune’s January 15 article emphasizes, 2026 is “the year AI penetrates the industrial workforce. The key, however, is to use AI not as a replacement for workers, but as a tool for training the next generation of talent.
The U.S. will need 3.8 million new manufacturing workers by 2033, but up to 1.9 million positions may go unfilled due to skills gaps AI-powered, personalized training programs can shorten the learning curve for new workers and incorporate the tacit knowledge of seasoned technicians into AI systems This will help address the talent shortage.
GE Aerospace will invest $30 million over the next five years to launch a program to train 10,000 highly skilled workers beginning in 2026. These initiatives show the future of manufacturing through AI and human collaboration.
Trend 4: Smart factories and autonomous operations
According to Deloitte’s 2026 Manufacturing Outlook, the majority of manufacturers plan to invest at least 20% of their improvement budgets in smart manufacturing initiatives. This includes automation hardware, data analytics, sensors, and cloud computing.
By 2029, it is predicted that 30% of factories will use open, virtualized, software-defined automation platforms to centrally manage control systems. This will allow factory-wide configuration changes and optimization to be performed at the software level without traditional physical rewiring.
Trend 5: Predictive optimization of supply chains
As tariff uncertainty and trade friction persist, predictive supply chain management using AI is becoming increasingly important. By leveraging large-scale language models and agentic AI, companies can automatically navigate trade risks and identify potential cost-saving opportunities.
According to a McKinsey study, agent-based AI will generate up to $650 billion in additional revenue in each industry by 2030, and automate repetitive tasks to reduce costs by up to 50%. In manufacturing, the technology is expected to be used across raw material procurement, inventory optimization, and logistics management.
5. challenges and prospects
Data security and OT/IT integration
IDC predicts that 75% of large manufacturers will use AI-enabled OT (operational technology) cyber defenses by 2029 to combat the risk of data model contamination. This will allow them to autonomously flag low-level threats and reduce detection time by 60%.
In addition, by 2027, 40% of all operational data will be integrated between applications and platforms autonomously by AI agents. This is due to increased standardization and the use of AI agents built for specific data.
Human-robot skill transfer
By 2028, companies that do not design two-way skills transfer loops are projected to face 20% higher downtime and retraining costs and lower efficiency compared to competitors that do implement them.
This demonstrates the importance of a cyclical learning system in which the robot not only learns from the human, but also feeds back the data and experience accumulated by the robot to train the human.
Sustainability and Energy Transition
Energy transition is an ongoing challenge for manufacturing and energy organizations. Integrating building information modeling (BIM) and manufacturing operations management (MOM) data with point cloud scanning can improve energy efficiency in buildings, which account for 30-40% of global CO2 emissions.
Scope 3 emission regulations (accountability for environmental impacts outside the direct scope of operations) and growing customer demand are driving investments in supply chain visibility, analysis, and optimization.
summary
The week of January 2026 clearly demonstrated that AI technology has entered a new phase. The practical application of physical AI, improved performance of large-scale language models, and massive investments in AI infrastructure are all elements that will shape the future of manufacturing.
For manufacturers, these technologies are more than just efficiency tools; they have the potential to transform the business model itself. Autonomous production systems, predictive maintenance, adaptive supply chain management, and collaborative human and AI workforce development will be key determinants of manufacturing competitiveness in the coming years.
At the same time, however, issues such as data security, human resource development, and energy supply must be seriously addressed; positioning AI not as a “human replacement” but as a “human augmentation” and establishing an optimal collaborative relationship between technology and humans will be the key to sustainable growth for the manufacturing industry.
The year 2026 will be the year that AI moves from the “experimental phase” to the “implementation phase” of manufacturing. In order not to miss this wave of change, we must start acting now.
Source List
- Reuters – “Artificial Intelligencer: the AI gold rush is coming for your gadgets” (January 14, 2026)
https://www.reuters.com/ technology/artificial-intelligence/artificial-intelligencer-ai-gold-rush-is-coming-your-gadgets-2026-01-14/ - Deloitte – “2026 Manufacturing Industry Outlook” (November 13, 2025)
https://www.deloitte.com/us/en/insights/industry/ manufacturing-industrial-products/manufacturing-industry-outlook.html - Fortune – “AI will infiltrate the industrial workforce in 2026-let’s apply it to training the next generation, not not replacing them” (January 15, 2026)
https://fortune.com/2026/01/15/lets-train-workers-on-industrial-ai-not-replace-them-kriti -sharma/ - Manufacturing Dive – “5 manufacturing trends to watch in 2026” (January 8, 2026)
https://www.manufacturingdive.com/news/5- trends-watch-2026-tariffs-uncertainty-ai-workforce-chemical-investments/809109/ - IDC – “Charting the AI-driven future of manufacturing” (November 12, 2025)
https://www.idc.com/resource-center/blog/ charting-the-ai-driven-future-of-manufacturing/ - Reuters – “DeepSeek to launch new AI model focused on coding in February” (January 9, 2026)
https://www.reuters.com/technology /deepseek-launch-new-ai-model-focused-coding-february-information-reports-2026-01-09/ - NVIDIA – “NVIDIA Releases New Physical AI Models as Global Partners Unveil Next-Generation Robots” (January 5, 2026)
https:// nvidianews.nvidia.com/news/nvidia-releases-new-physical-ai-models-as-global-partners-unveil-next-generation-robots - Google Blog – “Gmail is entering the Gemini era” (announced January 2026)
https://blog.google/products-and-platforms/products/ gmail/gmail-is-entering-the-gemini-era/ - OpenAI – “Announcing The Stargate Project” (January 21, 2025)
https://openai.com/index/announcing-the-stargate-project/ - Forbes – “How Jensen Huang Won CES 2026” (January 12, 2026)
https://www.forbes.com/sites/stevenwolfepereira/2026/01/12/ how-jensen-huang-won-ces-2026/
