AI Daily News — 2026-07-06 (Mon)

AI industry trend report (last 12 hours)

## 1. [Overview] While Anthropic is on the defensive due to copyright lawsuits and accusations of distillation attacks from Alibaba, there is also a tailwind from the lifting of US export restrictions on the Fable 5 model. OpenAI is moving closer to the government and making GPT-5.6 available exclusively to government-approved partners, accelerating regulatory compliance and integration into national strategies. The rise of Chinese open source AI (DeepSeek) shook the industry structure, and xAI faced a serious internal collapse with all 11 co-founders leaving the company.

## 2. [Anthropic / Claude trends] Claude is facing a new $75 million class action lawsuit alleging copyright infringement, reigniting legal risks surrounding Claude's training data procurement. Meanwhile, US export restrictions on the Fable 5 model have been lifted, easing some of the barriers to international expansion. Dario Amodei has publicly accused Alibaba of a "distillation attack" and warned against competitors copying its model without permission. We continue to lobby Congress to make AI safety testing mandatory.

## 3. [OpenAI trends] The biggest move is the decision not to release GPT-5.6 to the public, but instead to provide it only to government-approved partners. Sam Altman visits the White House and Congress to advance policy coordination following the Trump administration's AI executive order. Regarding IPOs, the company has declared that a valuation of less than $1 trillion is "undisputed," raising investors' expectations. It has also been reported that the US government will invest 5% in OpenAI, and the focus is on strengthening capital ties with the nation.

## 4. [Google AI trends] A review of the Gemini-equipped smart speaker has been published, and the development of consumer hardware is taking shape. Google Maps is testing AI food ordering feature with Gemini integration. On the other hand, a situation has arisen where Gemini supply is allocated to Meta, and it has become clear that demand exceeds supply. The Founding Fathers-themed Workspace ad was poorly received and a marketing failure.

## 5. [Trends of other AI companies] (Meta, Microsoft, xAI, DeepSeek, etc.) xAI was described as a "failure" by Yann LeCun, with all 11 co-founders leaving the company, and an organizational crisis surfaced. The maturity of Grok Imagine's image generation capabilities contrasts with the instability of the company's infrastructure. While Meta's "Watermelon" model is reported to have matched GPT-5.5 in benchmarks, Zuckerberg himself admits that progress in AI agent functionality has lagged. Microsoft has announced that it has sped up P&G's SAP operations by 40% on Azure, and is also in the process of introducing smart mode to Copilot and investing an additional $2.5 billion in its AI business. In the DeepSeek series, GLM-5.2, Kimi K2.6, etc. compete with DeepSeek V4 on the SWE bench, and the depth of the Chinese open source model is noticeable.

## 6. [Remarks and trends of AI celebrities] - Dario Amodei (Anthropic CEO): Publicly accused of conducting distillation attacks against Alibaba's AI models. Scheduled to attend G7 France with Sam Altman and Demis Hassabis. - Sam Altman (OpenAI CEO): Visited the White House and Congress. We have made it clear that we will not accept an OpenAI IPO valuation of less than $1 trillion. - Elon Musk (xAI Founder): All 11 co-founders of xAI have reportedly left the company, forcing them to break their silence. Coverage of SpaceX-related AI-equipped terminal prototypes continues. - Yann LeCun (Meta Chief AI Scientist): In addition to clearly calling xAI a "failure," he reiterated his position that current AI lacks understanding of the real world. - Jensen Huang (NVIDIA CEO): Mentioned in connection with positive news for Micron stock, and reported in connection with investment trends in the $40 trillion humanoid robot market by NVIDIA and Tesla. - Mark Zuckerberg (Meta CEO): Remarks admitting the slow progress of AI agent functions. He also became a hot topic when he was seen making business calls while wearing Meta Glasses on a jet ski.

## 7. [Breakthroughs and innovations in AI technology] The Straits Times pointed out that the Chinese open source model is "eating its own industry", highlighting the intensification of competition within the industry by the DeepSeek model. While high scores on SWE benchmarks such as GLM-5.2 and Kimi K2.6 have been reported one after another, concerns have been raised about the growing risk of model collapse due to Meta's massive AI ad generation. There are no notable stand-alone breakthrough announcements in this news group.

## 8. [AI Regulation/Policy/Social Impact] The first regulatory law targeting AI-generated advertising has been passed in New York state, making it mandatory to disclose AI-generated content. In Illinois, schools are facing challenges in enforcing a new law against cyber bullying and deep fakes using AI. Anthropic is lobbying Congress to legislate AI safety testing, continuing a tug-of-war between regulatory advocates and industry.

## 9. [Integrated analysis: What is happening now] What emerges from this series of news is the shrinking technological gap between US and Chinese AI companies, and the simultaneous occurrence of defensive actions on the part of US companies. Anthropic's accusations of distillation attacks and the performance improvements of DeepSeek-based open source models are two sides of the same coin, and the simultaneous occurrence of lawsuits in the name of intellectual property protection (Anthropic's copyright lawsuit) and accusations of technology plagiarism (Alibaba criticism) can be read as an indication that the US is trying to compensate for its diminished technological superiority through legal and discursive means.

Meanwhile, OpenAI and Google are getting closer to the government. The government-only release of GPT-5.6, Altman's visit to the White House, and the US government's proposed investment in OpenAI show that the AI industry is becoming national security. Google's move to narrow its Gemini supply to Meta confirms that computational resource scarcity remains an industry bottleneck.

The departure of all of xAI's co-founders is the biggest organizational change, and it exposes the limits of Elon Musk's personal brand-based management. LeCun's statement that "xAI is a failure" is a direct criticism from the Meta camp, and it is also worth noting that mutual criticism among AI companies is becoming more public.

## 10. [Future points of interest] 1. Developments in Anthropic’s copyright lawsuit and its impact on industry practices for AI learning data procurement 2. Details of the U.S. government's OpenAI investment proposal and whether similar state involvement will be expanded to other companies 3. Status of organizational restructuring after all xAI co-founders left and Musk's response

AI industry news analysis report (last 12 hours)

## 1. [Overview] Reports of accelerated IPOs for OpenAI and Anthropic have been the main focus, with OpenAI's huge deficit ($39B) leaked and skepticism about its funding structure growing. At the same time, signs of a slowdown or review of AI investment have surfaced in Meta/xAI (Tesla). On the geopolitical front, the rise of DeepSeek and the falsification of model origins between the US and China are continuing themes.

## 2. [Anthropic trends] Dario Amodei gave an interview to STAT about the impact of AI on the biotech field, and changes to Claude's terms of service have been reported (Security Boulevard). Provision of AI tools to the California state government is also underway, and penetration into the public sector continues. Meanwhile, emails between Amodei and his associates regarding an IT contract with the Pentagon were revealed, and the circumstances surrounding the deterioration of relations were reported. Regarding the $45B AI computing resource contract with SpaceX, Elon Musk has warned that ``if capacity becomes tight, we may withdraw,'' raising the risk of Anthropic's infrastructure supply.

## 3. [OpenAI trends] Ahead of its IPO, its financial details were leaked, and it was reported that it had posted a loss of $39 billion last year. The Financial Times analyzes that ``OpenAI and Anthropic may struggle to go public.'' The Daily Upside also reported on the accelerated IPO of both companies and increased regulatory oversight of token payment schemes. On the legal front, a new lawsuit has been filed alleging that ChatGPT promoted religious delusions and led to suicide attempts. UK regulators are calling for greater oversight of AI companies.

## 4. [Google AI trends] Android's new feature "Android Halo" was announced as an AI agent for smartphones. It has been reported that Gemini's data usage has expanded beyond advertising purposes, and the privacy debate continues. Tom's Guide introduces 13 new features of Gemini as of 2026. The impact of the DMA (Digital Markets Act) on the European competitive environment in the AI field is also being discussed. Regarding Sundar Pichai, Alphabet stock recorded its best monthly performance in about 22 years.

## 5. [Trends of other AI companies] Meta: AI cloud deal with Amazon Chips led to a sharp decline in semiconductor stocks. Zuckerberg acknowledged that AI agent development was slower than expected, saying, ``We won't be given research ideas, but we will be given excellent human resources and abundant computational resources.'' It was also reported that Meta is restricting engineers from using Claude. xAI/Tesla: A weekly limit of 19,000 rupees (equivalent to approximately 19,000 yen) has been set for Tesla employees' use of AI tools, in a move to curb in-house AI use. Microsoft: Announces plans to unify enterprise and consumer versions of the Copilot app. DeepSeek: Chinese-made models have a higher share than Claude (13.3% share) in OpenRouter. The introduction of API price tier system for DeepSeek V4 is also a hot topic. It was reported that Gu Yuxian from Tsinghua University has joined DeepSeek.

## 6. [Remarks and trends of AI celebrities] - Sam Altman (OpenAI CEO): Calls remote work "one of the worst mistakes in tech." A comment comparing her child's word combinations to the mathematical reasoning of GPT-5.6 also became a hot topic. - Dario Amodei (Anthropic CEO): ``A 100 million word context window is technically possible,'' he said, noting that it will also spread to the AI-crypto field. - Elon Musk (xAI Founder/Tesla CEO): While it is reported that SpaceX has invested $60B in AI, he restrained the computing resource contract with Anthropic by saying, ``We could withdraw if capacity is tight.'' - Jensen Huang (NVIDIA CEO): Pointed out the existence of new AI bottlenecks. He predicted that ``AI supercomputers will become commonplace in homes.'' The next generation AI rack system is reportedly delayed until 2028 due to manufacturing issues. - Mark Zuckerberg (Meta CEO): Reportedly, he is building an "AI CEO" to assist him in his work. Officially acknowledged delays in AI agent development. - GeoffrHinton (AI Godfather/Nobel Laureate): Mentioned potential for 83% cost reduction for self-hosted Frontier AI. It is reported that he called on Ghana to be wary of AI risks.

## 7. [Breakthroughs and innovations in AI technology] Progress is being made in the use of AI in cancer treatment, from early detection to tailored treatment (RTTNews). Bosch Global emphasizes leadership in robotics. Asia Times analyzes its AI development philosophy, which is different from that of the West, saying, ``China is building AI to serve as a compass, not a map.'' Hugging Face has announced an educational coding agent called Tau.

## 8. [AI Regulation/Policy/Social Impact] A legal roadmap for AI implementation on U.S. college campuses is discussed (Inside Higher Ed). It was pointed out that the Consumer Protection Act does not cover robotic shopping agents. There is a divided debate over the feasibility of the UK's ``AI growth zones'' initiative (The Guardian). In China, major chatbots are discontinuing their AI persona functions, as Beijing authorities tighten regulations (Nikkei Asia). 28,000 monthly job losses were reported in the tech and financial sectors as a result of AI.

## 9. [Integrated analysis: What is happening now] As OpenAI and Anthropic prepare to go public in earnest, it is clear that doubts about the financial standing of both companies are erupting at the same time. OpenAI's $39B deficit leak and FT's observation that it will struggle to go public suggests that investors are becoming more cautious about AI companies' profit models. On the other hand, Elon Musk's comments that treat the supply of computing resources to Anthropic as a hostage expose the interdependence and vulnerability of companies regarding AI infrastructure.

It is not a coincidence that both Meta and Tesla have simultaneously moved to curb the internal use of AI (restricting Claude usage, capping AI tool budgets), and can be seen as a sign that internal management of generated AI costs has emerged as a management issue. Zuckerberg's statement about ``delays in AI agent development'' is in line with this, and it appears that there is a gap between Meta's huge infrastructure investment (Amazon chip contract) and actual product results.

On the geopolitical front, DeepSeek's increasing share of OpenRouter and Claude's relative decline, as well as China's "compass-type AI" philosophy (Asia Times), indicate that the divergence of AI development approaches between the US and China is accelerating. In China, on the other hand, there is a parallel movement to strengthen controls by tightening regulations on AI personas, and we can see a two-pronged strategy of external technology export offensive and domestic control.

The new bottleneck pointed out by NVIDIA's Jensen Huang and the report that rack systems will be postponed to 2028 confirm the reality that the hardware supply network is not keeping up with the rapidly expanding demand for AI. Google's report that ``AI demand exceeds supply'' is consistent with this, and the scarcity of computing resources is becoming structured as a bottleneck for the entire AI industry.

## 10. [Future points of interest] 1. OpenAI/Anthropic's IPO trends and financial disclosures — How do large deficits affect investor sentiment and valuations? 2. Future of the AI computing resource contract between Anthropic and SpaceX — Will Musk's remarks lead to an actual review of the contract? 3. Expansion of Meta/Tesla's restraint on internal AI usage — Will it spread to other companies and cost consciousness become the industry standard?