AI Daily News — 2026-07-09 (Thu)

AI industry trend report (last 12 hours)

## 1. [Overview] xAI released Grok 4.5, and Elon Musk himself revealed his rivalry with bullish remarks calling it an "Opus-class model." OpenAI is preparing for the public release of GPT-5.6, and is active in raising funds, with BofA providing a 52 billion yen loan before the IPO. At the same time, geopolitical tensions have also surfaced, with Chinese authorities expressing concerns about a "security backdoor" regarding Anthropic's Claude Code.

## 2. [Anthropic / Claude trends] It has been reported that Chinese authorities have expressed concerns about a "security backdoor" regarding Claude Code. The sense of caution between the US and China regarding AI technology is also spreading to coding tools. Meanwhile, enhancements for developers continue, such as enhanced Claude integration in VS Code 1.128. FTI Consulting has published a report analyzing the unintended consequences of AI from a cyber risk perspective. The NYT also reports that AI-driven concentration of wealth is causing "excitement" in San Francisco's housing market, with the presence of AI companies including Anthropic extending into the local economy.

## 3. [OpenAI trends] The main focus will be on the public release of GPT-5.6, which multiple media outlets reported will finally be rolled out after several weeks on hold. BofA has provided a loan of 52 billion yen in anticipation of an IPO, and the evaluation from the capital market is beginning to appear in concrete numbers. In the education field, we are also focusing on expanding the range of social implementation, such as officially announcing support for building AI skills for K-12 educators. Policy regarding government-national security partnership was also published on the official blog.

## 4. [Google AI trends] Google has won a lawsuit over the data tracking of its Gemini AI assistant. On the product side, Google Photos' "Video Remix" feature now allows video stylization with Gemini Omni. On the other hand, Ars Technica points out that despite the addition of the new LLM, Gemini still lags behind in the Android Bench evaluation, and issues remain regarding its competitive advantage in terms of performance.

## 5. [Trends of other AI companies] (Meta, Microsoft, xAI, DeepSeek, etc.) xAI has released Grok 4.5, positioning it as the "most powerful model" for coding and agent tasks. Microsoft has integrated its Copilot strategy to strengthen its defense against ChatGPT and Claude, and announced a $25 billion AI implementation support initiative (6,000 people). Meta is planning its first large-scale AI data center in Canada and is also rolling out Magnifying Glass, an AI-generated image detection feature. DeepSeek is reported to have increased its traffic share to 17%, and is accelerating its move to develop its own AI chip.

## 6. [Remarks and trends of AI celebrities] Sam Altman (OpenAI CEO): An advanced model of GPT-5.6 is reportedly in direct competition with a new release of Grok. According to Meta CTO, Altman has made moves to counter the company's huge signing bonus strategy. Elon Musk (xAI Founder / Tesla CEO): Self-praises Grok 4.5 as an "Opus-class model." Meanwhile, there are reports that AI-related employment projections are undergoing a harsh reality check, with the EPA joining the DOJ's legal defense of xAI. Jensen Huang (NVIDIA CEO): Axios points out the problem of AI chip supply lines being concentrated in narrow pipelines. He was also reported to have recommended free AI tools that everyone should use. Dario Amodei (Anthropic CEO): Although it may be "impossible" to warn people about AI risks, Dario Amodei is reported to be willing to continue his efforts. Sundar Pichai (Google/Alphabet CEO): Multiple media outlets, including the Motley Fool, have analyzed that Google is the clear winner in the AI economic war between Google and Microsoft. Mark Zuckerberg (Meta CEO): There have been a number of reports that Meta's AI strategy has stalled even after cutting 8,000 jobs. Remarks by Magnificent Seven suggesting that the $700 billion AI capital investment in 2026 may end up being excessive investment are also attracting attention. Yann LeCun (Meta Chief AI Scientist): Reportedly criticized xAI as a "failure" and claimed the existence of an AI bubble. Geoffrey Hinton (AI Godfather): Predicts that most jobs will be replaced by AI by 2045, but sex workers, politicians, and sports coaches will survive.

## 7. [Breakthroughs and innovations in AI technology] NTT Data and HYMH claimed a breakthrough in "physical AI." SK Hynix reports progress in AI memory technology through partnership with South Korea's TetraMem. Built In also analyzes that non-engineers are becoming the biggest beneficiaries of the AI revolution, suggesting that the benefits of technological innovation are expanding beyond the traditional engineer-centered structure.

## 8. [AI Regulation/Policy/Social Impact] The FTC has proposed a new policy that would make it a violation of federal law to hide how AI systems are being "manipulated." Compliance and international enforcement challenges surrounding the EU AI Act also remain a topic of discussion. Illinois passed a strict AI law, and Connecticut senators are pushing for a similar response in the state's Legislature in 2027. Lawfare also covered the debate surrounding the enactment of comprehensive AI legislation at the U.S. level (the Great American AI Act).

## 9. [Integrated analysis: What is happening now] The current balance of power is centered around frontier model competition between four companies: OpenAI, Anthropic, Google, and xAI. xAI's release of Grok 4.5 and Musk's "Opus-class" remarks are symbolic moves that use Anthropic's brand position as a reference point, and the competitive framing itself is changing. OpenAI is currently in the process of preparing its financial structure for an IPO with the release of GPT-5.6 and a 52 billion yen loan from BofA.

In terms of investment trends, concerns about excessive investment in Magnificent Seven's $700 billion AI capital investment (as symbolized by Zuckerberg's comments) are beginning to surface, while moves toward vertical integration at the hardware layer are accelerating, such as SK Hynix's AI memory technology and DeepSeek's own chip development. DeepSeek's 17% traffic share shows that there is still a large segment of the population that values cost efficiency.

Geopolitically, the Chinese authorities' expression of concern about a "backdoor" to Claude Code is a new signal that the U.S.-China AI friction is extending to the infrastructure layer (coding tools). On the regulatory front, regulations are being tightened at each country and state level, such as the FTC's transparency requirements regarding AI output manipulation and Illinois' stricter state laws, and the costs for companies to comply are expected to increase further in the future.

Yann LeCun's criticism of xAI and the AI bubble theory indicate that there is a sense of caution about overheating even from within the industry, and sentiment is increasingly polarized among both investors and researchers.

## 10. [Future points of interest] - Benchmark comparison after public release of GPT-5.6 (performance difference with Grok 4.5, Claude, Gemini) - Market evaluation of whether Magnificent Seven's 2026 AI capital investment ($700 billion) is excessive investment - Development of security concerns about AI tools in the US and China (Claude Code backdoor issue) and effectiveness of each country's regulations (FTC, Illinois law)

AI news analysis report (last 12 hours)

## 1. [Overview] The biggest topic is xAI's official release of Grok 4.5 at an unbeatable price of $2/M, and the move to "destroy the price" of Claude Opus and GPT-5.6. Anthropic is reportedly valued at $1.2 trillion in private markets and is increasing its presence in the capital markets. On the other hand, there are analytical reports saying that Google has fallen behind in the AI race, and the positioning of each company is in flux.

## 2. [Anthropic trends] Anthropic announced its own off-switch mechanism for "dual use knowledge." The company is valued at $1.2 trillion in the private market, and demand is reportedly so high that it is "nearly impossible" to acquire shares. UST announced that it will integrate Claude into its platform and provide training for 20,000 employees. On the other hand, there is a noticeable trend in the media that the Grok 4.5 is taking aim at Claude Opus in terms of both price and performance.

## 3. [OpenAI trends] It is reported that GPT-5.6 has started to be released to the public after repeated delays. ChatGPT's audio function has been updated as "GPT-Live," allowing for more natural conversations. As for Sam Altman, in addition to the controversy over his AI wealth plan, there have also been security reports in which the defendant appeared in court after a Molotov cocktail was thrown at his San Francisco home. Remote work support for Codex was also announced.

## 4. [Google AI trends] An analytical article titled ``Why Google was late in the AI race'' was published on NPR, and the tone of the external evaluation was somewhat harsh. On the other hand, the company continues to steadily expand its products, including the launch of the Gemini iPhone version and the launch of Gemini for Home. There have been several announcements about the adoption of Cognizant and Renault, and the company's penetration into enterprises continues.

xAI released Grok 4.5 to the public on July 9th, with a price tag of $2/M input, significantly lower than Claude/GPT-5.6. Elon Musk boasts that it's "Opus-class, but fast." There is analysis that Meta has lost Llama's position to Qwen in the open source area, and it has been reported that the appeal of the copyright lawsuit has been postponed. DeepSeek is reported to have started developing its own AI chip, and has taken first place in the Apple App Store. Microsoft has been a hot topic for expanding 365 Copilot's AI collaboration and for GitHub Copilot's sharp increase in usage fees ($800 per month).

## 6. [Remarks and trends of AI celebrities] - Sam Altman (OpenAI CEO): AI wealthy scheme sparks controversy; defendant in house firebombing appears in court. Commenting on the ChatGPT audio update: ``It's magical and yet realistic.'' - Elon Musk (xAI founder/Tesla CEO): Leading the release of Grok 4.5 and calling it "Opus-class." A conversation with Texas Governor Abbott about space-based AI. Optimus and AI will change healthcare, Cathie Wood predicts $50 billion in AI computing demand. - Jensen Huang (NVIDIA CEO): Said that their engineers prefer building agents to writing code. He said he would be "deeply concerned" if an engineer with an annual salary of $500,000 did not spend $250,000 worth of tokens, underscoring a shift in the way people view AI costs. - Dario Amodei (Anthropic CEO): The focus is on the private market valuation of $1.2 trillion. There are analytical reports that major tech companies are changing their tone regarding job loss due to AI. - Mark Zuckerberg (Meta CEO): At one point, Meta stock fell 5% after the AI agent admitted it was "not really accelerating." - Geoffrey Hinton (AI Godfather): Warns that growth supremacy is dangerously accelerating the existential risks of AI. He commented that ``no one knows'' the impact on employment.

## 7. [Breakthroughs and innovations in AI technology] Grok 4.5 appeared as a coding agent specialized model trained on Cursor, and the price destruction of $2/M input became the biggest technical topic. There are also reports that Google's AI compression technology will help reduce power consumption in data centers, and the trend toward improving infrastructure efficiency continues. There are no big announcements about Sparse Attention or new architectures in this news set.

## 8. [AI Regulation/Policy/Social Impact] Illinois' AI Safety Measures Act (SB 315) stipulates obligations required of frontier AI development companies by January 2028, and law firms are calling for action. Microsoft President Brad Smith has criticized Washington's AI policy, calling it "regulations that lack transparency and integrity." An article has been published urging members of Congress to sort out the issues in AI policy, and the process is moving toward the implementation stage of regulations.

## 9. [Integrated analysis: What is happening now] The core of this news set is xAI's price offensive. Grok 4.5's $2/M setting clearly targets Claude Opus and GPT-5.6, and signals a turning point from competition in model performance differentiation to price competition. At the same time, Anthropic has received a high evaluation from the capital market with a valuation of $1.2 trillion in the private market, and the gap between product price competition and investor evaluation is noticeable.

Although Google has been criticized for being a ``latecomer,'' it is steadily gaining ground by expanding its base with Gemini for Home, iPhone version deployment, and enterprise implementation, so it cannot be simply said that it is at a disadvantage. It has been pointed out that Meta has been overtaken by Llama and Qwen in the open source area, and it is becoming clear that Chinese companies (DeepSeek, Qwen) are catching up technologically. DeepSeek's announcement to develop its own chip is a move that should be watched closely as a strategic shift by Chinese AI companies aiming to break away from dependence on NVIDIA.

Among the comments by celebrities, Jensen Huang's comment on the amount of token consumption is symbolic, and shows that the standard of value for AI talent is shifting from ``the amount of code written'' to ``how much AI is used.'' Zuckerberg's frank admission that "AI agents are not accelerating" exposed the gap between the market's overheated expectations and the reality of implementation.

## 10. [Future points of interest] 1. Will Anthropic and OpenAI respond to the Grok 4.5 price offensive with price cuts or differentiation strategies? 2. Official release date of GPT-5.6 and the technical issues behind the delay 3. Starting with Illinois SB 315, how far will frontier AI regulations take shape in other states and at the federal level?