1. Executive Summary
Today, highly significant news that marks a major turning point in the AI industry was reported. OpenAI’s decision to discontinue the video generation app “Sora” symbolizes the consolidation and strategic realignment of generative AI products. Meanwhile, NVIDIA is promoting the standardization of AI infrastructure through technology donations to the open-source community, and Microsoft has announced new optical technology that dramatically improves data center energy efficiency. Tensions between law enforcement and Anthropic also continue, intensifying the tug-of-war between technological innovation and regulation.
2. Today’s Highlights
OpenAI Announces End of Video Generation AI “Sora”
OpenAI has decided to discontinue “Sora,” the video generation AI app released in Fall 2025 that garnered significant attention. According to an announcement on its official social media, the app and its developer API will be phased out. This decision stems from persistent concerns from Hollywood stakeholders and creators regarding copyright, as well as criticism over the proliferation of “AI slop” (valueless AI-generated content). In particular, the issue of deepfake videos of prominent individuals was a major blow. Partners such as Disney have also issued statements respecting this decision, and it is believed that OpenAI will temporarily withdraw from the video generation market to rebuild its product strategy. This move suggests a transition from the rapid expansion phase of generative AI to a phase that demands more responsible operation and sustainable business models.
Source: Sora Official SNS
NVIDIA Donates GPU Management Technology to Kubernetes Community
At KubeCon Europe, NVIDIA announced its donation of its “Dynamic Resource Allocation (DRA) Driver” for GPUs to the Cloud Native Computing Foundation (CNCF). This will enable more open and transparent management and orchestration of GPU resources on Kubernetes, which serves as the foundation for AI workloads. By placing technology that was previously managed solely by NVIDIA under community control, increased efficiency in building AI infrastructure across the industry is expected. Furthermore, NVIDIA’s “KAI Scheduler” has been adopted as a CNCF sandbox project, accelerating efforts to optimize inference for large-scale AI models. This demonstrates the company’s strategic intent to enhance its dominance not only in hardware but also in the software ecosystem within the AI competition.
Source: NVIDIA Official Newsroom
3. Other News
-
Microsoft Announces MicroLED Communication Technology for AI Data Centers Microsoft Research has announced a new data transmission technology using “MicroLED,” an alternative to copper wires and existing fiber optics. This technology is expected to reduce power consumption by approximately 50% and enhance reliability in inter-server communication within AI data centers. Commercialization is planned for 2027 and beyond, making it a crucial key to solving the energy constraint issues of AI servers. Source: Microsoft Official News
-
Conflict Between Anthropic and US Government Intensifies in Court In a lawsuit where Anthropic is protesting the US Department of Defense’s “Supply Chain Risk Designation,” a federal district court judge suggested that the government’s actions appeared to be an unfair “punishment” against Anthropic. The company claims this is retaliation for its refusal to allow AI use for autonomous weapons and surveillance, and this ruling is expected to have a significant impact on the safety policies of future AI companies and government involvement. Source: WXXI News
-
OKI Enters EMS for AI Servers Japan’s OKI Electric Industry (OKI) has announced the launch of its contract manufacturing services (EMS) for AI server equipment. With strengths in manufacturing and implementing large multi-layer boards equipped with high-heat-dissipating AI semiconductors, the company aims to meet the surging demand for AI servers. They aim for 1 billion yen in sales in fiscal year 2026. Source: Business Wire
-
Google Releases Data Pipeline with LlamaParse and Gemini 3.1 Google Developers has published a guide on building an AI workflow to extract structured data from unstructured documents (like PDFs) by combining LlamaIndex’s “LlamaParse” and “Gemini 3.1 Pro.” This is drawing attention as an implementation reference for developers, aimed at tasks such as automating financial analysis in enterprises. Source: Google Developers Blog
-
Redwood AI Presents at BIO-Europe Spring Redwood AI, which operates the AI-driven chemical synthesis platform “Reactosphere,” presented its technology at BIO-Europe Spring 2026. Discussions revolved around how to balance safety and execution speed in accelerating drug discovery and chemical research using AI. Source: ACCESS Newswire
4. Conclusion and Outlook
From today’s news overall, it is clear that the AI industry has significantly shifted its focus from “experimentation” to a phase of “real-world constraints and responsible operation.” OpenAI’s discontinuation of “Sora” highlights the weight of social responsibility and regulation that cannot be sustained by technological potential alone. Meanwhile, the technologies presented by NVIDIA and Microsoft focus on optimizing AI’s physical infrastructure (power, data transmission, management platforms), laying the groundwork for a sustainable AI ecosystem. In the coming months, hardware optimization and governance surrounding AI safety and copyright will likely become the main battlegrounds for development competition.
5. References
This article was automatically generated by LLM. It may contain errors.