In the week leading up to the height of summer, the AI industry witnessed simultaneous seismic shifts across three layers: infrastructure, devices, and workforce.While Apple announced a new Siri powered by Gemini, marking its full-scale entry into the battle for dominance in on-device AI, Anthropic raised the stakes in the competition for computing resources with a multi-gigawatt TPU contract, and the Chinese government accelerated the deployment of humanoid robots as a national policy.In the capital markets, OpenAI and SpaceX are charging toward IPOs, marking the moment when AI is fully transforming from an “experimental technology” into “social infrastructure.” In this article, we’ll break down how manufacturing sites should interpret this week’s news through five key topics.
1. Apple Unveils “Siri AI” with Gemini at WWDC 2026 — The True Dawn of the Multi-AI Era
On June 8 (U.S. time), Tim Cook took the stage for his final WWDC keynote as CEO and announced the next generation of Apple Intelligence and a completely revamped “Siri AI.” The new Siri runs on a custom Gemini model with 1.2 trillion parameters, which Apple licensed from Google for approximately $1 billion annually.Additionally, a “Multi-AI Extensions” system was introduced, allowing users to choose any AI from ChatGPT, Gemini, or Claude, and Anthropic Claude became an official option on the iPhone for the first time. In terms of features, agent-like capabilities have been significantly enhanced, including instant nutritional information when photographing food in the Camera app, natural language search across the Photo Library for subjects and people, context-aware reply suggestions in Messages, and iterative editing with the image generation tool “Image Playground.”iOS 27 also notably boosted on-device performance to support AI operations, with app launches up to 30% faster, photo loading up to 70% faster, and AirDrop up to 80% faster.
Implications for Manufacturing:
The era of “BYOAI” (Bring Your Own AI)—where field workers switch between AI models on their personal smartphones in accordance with company policies—has effectively begun.Many processes that previously required dedicated devices—such as inspection records, immediate diagnosis via photos of defective products, and translation of work instructions into multiple languages—can now be replaced by commercially available smartphones and general-purpose AI.At the same time, there is an urgent need to establish MDM (Mobile Device Management) combined with AI governance to define policies for handling confidential production data on a per-model basis.
2. Anthropic Signs Multi-Gigawatt TPU Deal with Google and Broadcom — Claude’s Run-Rate Revenue to Reach $30 Billion
Anthropic announced that it has signed a major agreement with Google Cloud and Broadcom to secure several gigawatts of next-generation TPU capacity.Scheduled to come online sequentially starting in 2027, this will significantly expand the computational infrastructure for the Frontier Claude model. The company’s CFO, Krishna Rao, stated that run-rate revenue has surged from approximately $9 billion at the end of 2025 to over $30 billion in 2026,and that the number of enterprise customers spending over $1 million annually has doubled from 500 to 1,000 in less than two months. Notably, Anthropic has clarified its strategy of utilizing a three-pronged approach to hardware, leveraging AWS Trainium, Google TPUs, and NVIDIA GPUs.Claude remains the only frontier model available on all three major cloud platforms—AWS Bedrock, Google Vertex AI, and Microsoft Azure Foundry—with Amazon remaining the primary cloud provider and training partner for Project Rainier.Around the same time, Anthropic filed its S-1 documents confidentially with the SEC (U.S. Securities and Exchange Commission) and has begun full-scale preparations for its IPO.
Implications for Manufacturing:
The procurement of AI models is shifting from a “single-vendor dependency” model to one based on a “multi-cloud, multi-model” approach.IT departments in the manufacturing sector have reached a stage where they need to establish strategic dependency assessments to explicitly evaluate vendor dependency risks, exit strategies, audit rights, and exposure to price fluctuations regarding AI functions integrated into ERP, MES, and SCM systems.Furthermore, the power and water consumption of AI infrastructure is becoming a management issue, and competition between data center power and factory power will become apparent in the coming years.
3. China’s MIIT Launches National Action Plan for Humanoid Robots — Over 100 Applications to Be Implemented by the End of 2026
On June 9, China’s Ministry of Industry and Information Technology (MIIT) and the State-owned Assets Supervision and Administration Commission (SASAC) jointly announced a large-scale action plan to accelerate the implementation of humanoid robots and embodied AI technologies.The plan sets a goal of deploying 10,000 humanoid robots in industrial settings and establishing more than 100 high-value use cases by the end of 2026. The scope covers China’s core manufacturing sectors, including EV assembly, electronics assembly, logistics warehouses, quality inspection, and machinery maintenance. In parallel, FANUC America announced that it is preparing demonstrations of “Physical AI” and AI-enabled robots for Automate 2026 (scheduled for June 22–26 in Chicago).Western companies such as Figure, Apptronik, and Agility Robotics are also successively transitioning their pilot projects in automotive and logistics settings to full-scale operations, and this week, the moment when humanoids crossed the threshold from the laboratory to the factory floor was observed simultaneously at multiple locations.
Implications for Manufacturing:
The cautious view that “it’s still too early for humanoids” is rapidly becoming a moot point. For Japanese manufacturers with production bases in China in particular, if local suppliers boost productivity by 10–30% through the introduction of robots, a decisive gap in price and lead time competitiveness will emerge within one to two years.Now is the time to begin shortlisting the “processes that humanoids will replace first” on the shop floor—such as pallet handling, parts supply, fixture changes, and nighttime inspections—and to secure a budget (in the range of 5 to 20 million yen) for proof-of-concept (PoC) evaluations.
4. OpenAI Files IPO Documents with the SEC — Also Announces Major Partnership with Oracle Universal Credits
On June 8, Reuters reported that OpenAI had filed its IPO documents confidentially with the SEC, following in the footsteps of Anthropic.Based on media reports, the company’s estimated valuation exceeds $850 billion, with monthly revenue reaching approximately $2 billion. Furthermore, on June 11, OpenAI and Oracle launched a partnership that allows enterprise customers to access OpenAI’s Frontier models and Codex using their existing Oracle Universal Credits (UCM) allocation. This is more than just an expansion of sales channels. As part of its “Stargate” project, Oracle has announced $500 billion in data center investments, with dedicated facilities currently under construction in various U.S. states, including Texas, New Mexico, Wisconsin, and Michigan.The ability to procure AI models using existing ERP budgets marks a turning point that will dramatically lower both the psychological and contractual barriers for procurement departments in the manufacturing sector.
Implications for Manufacturing:
The traditional model—where “AI implementation” meant “adding a new vendor,” “extended PoCs,” and “endless approval processes”—is gradually being restructured to allow companies to purchase AI through existing core system vendors such as Oracle, SAP, and Microsoft.In areas already running on Oracle or SAP—such as SCM, quality control, production planning, and demand forecasting—there will likely be a rapid increase in cases where evaluating AI add-ons through core system vendors takes precedence over in-house development, offering advantages in terms of TCO, operations, and auditing.
5. The Launch of NVIDIA RTX Spark and the AI Semiconductor Market — Edge AI Comes to the Manufacturing Floor
Following the announcement at Computex 2026, NVIDIA’s new “RTX Spark” was officially launched in the edge AI market this week.Featuring Blackwell-generation RTX GPUs, 6,144 CUDA cores, and 5th-generation Tensor Cores (supporting FP4 precision), it enables local inference of large-scale models on laptops and desktops.Meanwhile, while the AI semiconductor sector experienced a correction in early June, with the market shrinking by approximately $1.4 trillion, the outlook for the overall semiconductor market remains unchanged, with projections indicating it will reach $1.3 trillion for the full year of 2026. That same week, Foxconn announced an initiative to build a digital twin on NVIDIA Omniverse and optimize the AI server production line itself within a simulation environment.Siemens is also moving forward with the release of “Digital Twin Composer” on the Siemens Xcelerator Marketplace, and the trinity of AI chips, digital twins, and Physical AI is becoming the standard stack for factories.
Implications for Manufacturing:
For factories struggling with persistently high cloud inference costs, a “return to edge AI”—deploying RTX Spark-class workstations on the shop floor—has become a viable option.In areas requiring low latency and data confidentiality—such as visual inspection, abnormal noise diagnosis, and robot path optimization— a hybrid design combining cloud AI and edge AI will become the mainstream architecture over the next 12 months. When planning capital expenditures, companies should incorporate edge AI GPUs into their server rack and industrial PC replacement schedules.
General Considerations for Manufacturing
— A three-pronged strategy combining “AI Governance × Physical AI × Return to Existing Vendors”
Looking at this week’s news, the key issues that the manufacturing industry needs to address over the next one to two years can be organized into three main areas. First is making AI governance actionable.With developments such as Apple Siri AI’s Multi-AI Extensions, Anthropic’s multi-cloud TPU strategy, and regulations like the U.S. NSPM-11 and the EU AI Act (with a deadline of August 2—55 days away), we have entered an era where the “where, who, and why” of AI usage must be defined at the level of technical agreements.Manufacturers must prioritize AI adoption policies, model selection criteria, cross-border data rules, and audit log requirements in their roadmaps with the same weight as ERP updates. Second, preparing for the production deployment of Physical AI.China’s plan for 10,000 humanoids, FANUC’s Physical AI demo, Foxconn’s digital twin factory, and Siemens’ xcelerator integration are all moves to shorten the “simulation → physical verification → mass production deployment” cycle from 2–3 years to 6–12 months.For Japan’s manufacturing sector to avoid falling behind this wave, it is essential to establish permanent in-house evaluation environments (test cells) for humanoids and collaborative robots and foster a culture of uncovering use cases led by the shop floor. Third, reducing procurement friction through a “return to existing vendors.” The OpenAI–Oracle partnership, the integration of Claude into Microsoft Foundry, and Anthropic’s rollout across the three major cloud platforms have paved the way for purchasing AI not as “adding a new vendor” but as an “extension of existing ERP/cloud contracts.”This significantly reduces the burden on finance, legal, and IT departments and creates an environment where operational departments can test AI features on a monthly basis. This serves as a powerful tailwind for overcoming the lengthy approval processes characteristic of Japan’s manufacturing sector. The overarching message is clear. AI has completely moved beyond the stage of debating “whether or not to adopt it” and has entered a phase where the focus is on the precision of implementation design: “which model to use, in which process, through which vendor, and under what governance.”
summary
The week of June 8–13, 2026, was a unique one in which AI demonstrated decisive momentum simultaneously across five layers: “consumer devices (Apple),” “computing infrastructure (Anthropic),” “the physical world (China’s MIIT and FANUC),”“capital markets (OpenAI, SpaceX),” and “edge semiconductors (NVIDIA).” There are three key points that manufacturing executives must keep in mind. First, design AI governance at the technical contract level.Second, establish an in-house evaluation environment for humanoid and physical AI within this fiscal year. Third, make maximum use of shortcuts for AI procurement by leveraging existing ERP and cloud contracts. With Automate 2026 (Chicago, June 22–26) coming up next week, the full-scale deployment of Physical AI and robotics is expected to accelerate further. We can’t take our eyes off the moves each company is making.
Source List
1. Apple Newsroom: “Apple unveils the next generation of Apple Intelligence, Siri AI, and more” — https://www.apple.com/newsroom/2026/06/apple-unveils-next-generation-of-apple-intelligence-siri-ai-and-more/
2. Apple Newsroom: “Apple introduces Siri AI, a profoundly more capable and personal assistant” — https://www.apple.com/newsroom/2026/06/apple-introduces-siri-ai-a-profoundly-more-capable-and-personal-assistant/
3. Build Fast With AI “AI News Today – June 8, 2026: 16 Biggest Stories” — https://www.buildfastwithai.com/blogs/ai-news-today-june-8-2026
4. Build Fast With AI “AI News June 11, 2026: 12 Biggest Stories Today” — https://www.buildfastwithai.com/blogs/ai-news-today-june-11-2026
5. Anthropic: “Anthropic Expands Partnership with Google and Broadcom for Compute” — https://www.anthropic.com/news/google-broadcom-partnership-compute
6. Anthropic: “Anthropic confidentially submits draft S-1 to the SEC” — https://www.anthropic.com/news/confidential-draft-s1-sec
7. Reuters: “OpenAI files for US IPO after Anthropic as AI giants head to public markets” — https://www.reuters.com/technology/openai-files-us-ipo-after-anthropic-ai-giants-head-public-markets-2026-06-08/
8. Yahoo News UK / KameraOne “China launches major push to deploy tens of thousands of humanoid robots” — https://uk.news.yahoo.com/china-launches-major-push-deploy-010056635.html
9. AI Weekly, “China Pushes 10,000 Humanoid Robots Into Work by 2026” — https://aiweekly.co/alerts/china-pushes-10000-humanoid-robots-into-work-by-2026
10. PR Newswire: “FANUC America Showcases Physical AI and AI-Enabled Robotics Demos at Automate 2026” — https://www.prnewswire.com/news-releases/fanuc-america-showcases-physical-ai-and-aienabled-robotics-demos-at-automate-2026-302782870.html
11. NVIDIA “GeForce @ COMPUTEX 2026: NVIDIA RTX Spark Unveiled” — https://www.nvidia.com/en-us/geforce/news/computex-2026-nvidia-geforce-rtx-announcements/
12. CNBC: “Nvidia’s new PC chips are the CEO’s bid to ‘own’ every part of the AI stack” — https://www.cnbc.com/2026/06/02/nvidias-new-pc-chips-are-ceos-bid-to-own-every-part-of-ai-stack.html
13. Foxconn Press Release: “Foxconn to Build AI Factories with NVIDIA Omniverse Platform” — https://www.foxconn.com/en-us/press-center/press-releases/latest-news/1484
14. Siemens News: “Siemens Unveils Digital Twin Composer” — https://news.siemens.com/en-us/digital-twin-composer-ces-2026/
15. Medium / Stephen Stanley, “AI Intelligence Report for June 1–June 10, 2026” — https://medium.com/@stephen.stanley777/ai-intelligence-report-june-1st-june-10th-8a53edc5d4e2
16. The White House, “Promoting Advanced Artificial Intelligence Innovation and Security” — https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/
