AI Daily News — 2026-07-11 (Sat)

AI industry trend report (last 12 hours)

## 1. [Overview] The legal battle between Big Tech has entered a new phase, with Apple suing OpenAI for stealing trade secrets. Anthropic has received polarized treatment, with Bernanke serving as its supervisory body, and conversely facing pressure to shut down the Air Force case. Meta was the week's winner, with Muse Spark 1.1's low-cost API launch pushing its stock higher.

## 2. [Anthropic / Claude trends] The addition of former Federal Reserve Chairman Bernanke to Anthropic's AI oversight body is notable as an incorporation of traditional financial authority into AI governance. Meanwhile, the U.S. Air Force has notified contracting companies to exclude Anthropic by September 1st, creating uncertainty about Claude's position in the defense field. The enterprise partnership with UST and TeraWulf's $3.5 billion plan to raise debt for data centers demonstrate growing demand for infrastructure. Although Anthropic ranked first in the AI Safety Index, it only received a C+ rating, highlighting the low safety standards of the industry as a whole.

## 3. [OpenAI trends] The biggest factor is the lawsuit filed by Apple. Axios, CNBC, CNN, WSJ, and Reuters have all reported on the incident, with Apple claiming that commercial secrets were systematically stolen "at all levels," and also pointing to the use of commercial secrets for AI devices under development. In addition, the NYT has asked the court for sanctions regarding ChatGPT's learning data, and copyright litigation is intensifying at the same time. The release of GPT-5.6 is also impacting the outlook for Agentic AI-related ETFs.

## 4. [Google AI trends] At Google I/O, Gemini's positioning was redefined and a new model was mentioned. AlphaEvolve GA has been rolled out as an official version, and algorithm optimization using Gemini has been extended to GPU design and drug discovery areas. There is also a bullish theory (Cramer) in the market that ``Google will win the entire AI war,'' citing Apple's 2.5 billion device base as evidence.

## 5. [Trends of other AI companies] (Meta, Microsoft, xAI, DeepSeek, etc.) Meta released Muse Spark 1.1 API at a quarter of the price of Anthropic/OpenAI, and the stock price rose significantly in a week. Zuckerberg appeared for the first time in three years and announced a new agent model that specializes in programming. While Microsoft's report revealed that Copilot is not widely used, there is also a trend for the latest model of OpenAI to be adopted as the engine for M365 Copilot. xAI introduced Grok to the Japanese market and changed its company name to SpaceXAI. Regarding DeepSeek, it has been pointed out that its proprietary chips could shake up the power balance of NVIDIA and Huawei, and it is increasingly being mentioned in the context of counterterrorism measures through AI abuse.

## 6. [Remarks and trends of AI celebrities] - Sam Altman (OpenAI CEO): Said at the Sun Valley conference, "What everyone wants to know is how to make AI cheaper." There were also reports of organizational turmoil, with a close executive retiring before the IPO due to chronic illness. - Elon Musk (xAI Founder / Tesla CEO): Sharply criticized rival companies upon launching Grok 4.5. He also proposed a future vision in which AI will be responsible for funding government benefits and ``labor will become voluntary.'' - Jensen Huang, NVIDIA CEO: Says software engineers prefer building AI agents to writing Python, and also mentions new bottlenecks. - Dario Amodei (Anthropic CEO): While he said that ``coding will disappear first, followed by software engineering,'' he also said that ``there is a shortage of human resources that we can't keep up with,'' showing a sense of the field that contradicts concerns about job loss. - Ilya Sutskever (SSI Co-Founder): Reported claims that AGI is already here and will bring "infinite energy, infinite calories, and infinite knowledge." - Yann LeCun (Meta Chief AI Scientist): It has been revealed that he is in negotiations with LPs to establish a new VC fund, and his movements outside of Meta will be closely watched.

## 7. [Breakthroughs and innovations in AI technology] The start of general availability of AlphaEvolve GA has great technical significance in terms of expanding applications based on Gemini in the fields of scientific computing, GPU design, and drug discovery. Meta's breakthrough in reducing ``hidden AI costs'' is also a factor in stock prices, and improving inference cost efficiency has emerged as a competitive axis. The issue of whether Embodied AI will be the "next computing revolution" was also raised.

## 8. [AI Regulation/Policy/Social Impact] The Trump administration has moved to target state-level AI legislation as "ideological." The Harvard Gazette discussed ``Responsibility in the event that an AI agent goes out of control and causes damage,'' and the lack of accountability was pointed out as a challenge in high-risk environments (JD Supra). Educational settings are also divided, with the University of Chicago Law School banning the use of cellphones and computers to curb the use of AI by students. An investigation into the misuse of AI tools by ISIS-related organizations is also a source of security concern.

## 9. [Integrated analysis: What is happening now] The biggest structural change this week is that the legal battle between Apple and OpenAI has evolved from a conflict between AI companies to a conflict between Big Tech and AI companies. Apple is in a contradictory relationship, partnering with OpenAI to develop AI capabilities for its devices, but also suing OpenAI for theft of commercial secrets, creating uncertainty about the entire partnership going forward.

Meta clearly initiated the price war. Muse Spark 1.1's pricing, which is one-fourth that of Anthropic and OpenAI, indicates a competitive shift from differentiation in model performance to differentiation in inference cost efficiency. The stock market liked this, with Meta stock posting its best performance of the week.

Anthropic is simultaneously receiving contradictory signals of ``increasing support for enterprises'' and ``exclusionary pressure in the government/defense field,'' and it appears that Claude's evaluation of political neutrality and safety does not match actual procurement decisions. A C+ rating on the AI Safety Index shows the reality that even the industry leaders are unable to meet safety standards.

Among the statements by celebrities, Dario Amodei and Jensen Huang both talk about a future vision in which ``human work will shift to the side of building AI agents,'' while Amodei himself says that ``recruitment is not keeping up.'' It is interesting to see the contradiction -- the theory of job loss and the theory of talent shortage coexist from the same person's mouth.

Geopolitically, it has been suggested that the rise of DeepSeek's proprietary chips may drive a wedge into the oligopoly structure of NVIDIA and Huawei, and along with the financial disparity issue in Taiwan, pressure is increasing to reorganize the AI semiconductor supply chain.

## 10. [Future points of interest] 1. The outcome of the Apple vs. OpenAI lawsuit: Settlement or prolongation, and the impact on the partnership between the two companies 2. Will Meta Muse Spark's low price strategy trigger price revisions for Anthropic and OpenAI? 3. Will the U.S. Air Force's Anthropic Exclusion Notice (deadline: September 1st) affect other government agencies and contracting companies?

AI industry trend report (last 12 hours)

## 1. [Overview] Two major news stories shook the industry: Anthropic officially launched Claude Sonnet 5 and Apple sued OpenAI for trade secret theft. Meta has withdrawn its AI image generation tool for Instagram due to backlash, and xAI has surfaced a move to ban Tesla employees from using Claude and force them to migrate to Grok.

## 2. [Anthropic / Claude trends] Claude Sonnet 5 has been officially launched (as reported by Mashable). In parallel, Anthropic announced new developments in interpretability technology and is working to visualize Claude's internal reasoning process (Ubergizmo). Anthropic launched a "Hard Questions Portal" and published a survey showing that people trust AI companies less than government agencies (Tech Times). Regarding the IPO, it is said that the S-1 has been submitted privately, and the IPO competition with OpenAI is in full swing (MSN). On the other hand, there are reports that there are concerns about switching to Chinese AI tools in response to Claude Code's backdoor warning (South China Morning Post).

## 3. [OpenAI trends] The biggest news is the lawsuit filed by Apple. Apple filed a lawsuit accusing OpenAI of stealing trade secrets related to hardware development (multiple media outlets including the Washington Post, NYT, AP, Bloomberg, and the Guardian reported the case). The complaint alleges that "all levels of OpenAI" were involved (MSN). At the same time, it was revealed that Johannes Heidecke, head of safety at OpenAI, would be leaving the company amid organizational restructuring (WIRED, The Tech Buzz). British authorities have pointed out that GPT-5.6 has a "universal jailbreak" that could unleash dangerous cyber capabilities (Fortune). The deployment of ChatGPT Work for companies is also continuing (KuCoin, multiple implementation examples).

## 4. [Google AI trends] There were no notable technology announcements, and the focus was on consumer topics such as Gemini for Android Auto experience improvements (Tom's Guide) and Gemini/ChatGPT comparison articles. Of note is the fact that Google has dropped out of the top 5 AI labs for the first time in the Artificial Analysis Intelligence Index (OfficeChai), which should be watched closely as a signal of declining competitiveness.

## 5. [Trends of other AI companies] (Meta, Microsoft, xAI, DeepSeek, etc.) Meta pulls Instagram's AI image generation tool after backlash (TechCrunch, BBC, NBC Bay Area, Deadline). There is a history of strong criticism from CAA and SAG-AFTRA. There is a "K-shaped gap" within Meta, where AI talent receives high pay while regular employees fear layoffs (finance.biggo.com). Regarding xAI, there are multiple reports (Electrek, Startup Fortune, Yahoo Finance) that Elon Musk instructed Tesla employees to switch from Claude to Grok, acknowledging the inferior performance of Grok itself (Electrek, Startup Fortune, Yahoo Finance). Grok 4.5 is in second place on the APEX-SWE leaderboard (Crypto Briefing). The adoption rate of Microsoft Copilot remains at less than 4.5%, and it is pointed out that it is struggling in comparison with the price increase (MSN). It has been reported that DeepSeek is developing its own AI chip (The Daily Star), and the company is also accumulating in Zhongguancun, Beijing (Seoul Economic Daily).

## 6. [Remarks and trends of AI celebrities] Sam Altman (OpenAI CEO): Regarding OpenAI's IPO valuation, he said, "If it's less than $1 trillion, it's a nonstarter" (AOL.com). He also bears the brunt of the lawsuit filed by Apple. Dario Amodei (Anthropic CEO): Leads the creation of the Hard Questions Portal and publishes survey results showing low trust in AI companies. The company is said to have privately filed an S-1 for the IPO (MSN). Jensen Huang (NVIDIA CEO): An in-house engineer said, "I prefer building agents to writing code" (MSN). His personal recognition continues, with actual sales of his jackets becoming a hot topic. Mark Zuckerberg (Meta CEO): Revealed plans to begin mass production of Meta's proprietary AI chip in September, but shares fell 5% after AI agent admitted it "isn't really accelerating very much" (AOL.com). He was also responsible for the withdrawal of the Instagram AI image tool. Elon Musk (xAI Founder / Tesla CEO): Instructs Tesla employees to migrate to Grok, a decision while acknowledging that it is inferior to Claude in terms of performance (Startup Fortune). Others have said that AI will finance government benefits and make work optional (AOL.com). Ilya Sutskever (SSI Co-Founder): Although there are no reports of his direct remarks, there have been reports in related contexts that he has been linked to the departure of the OpenAI safety officer and a drop in Google's rankings. Demis Hassabis (Google DeepMind CEO): He is only mentioned in the context of the Apple vs. OpenAI lawsuit, and his statements have not been reported.

## 7. [Breakthroughs and innovations in AI technology] Anthropic's interpretability technology is making progress toward visualizing Claude's internal thought processes (Ubergizmo). Other than that, there are no reports of clear technological breakthroughs, and there are many articles pointing out that the tone of the industry is shifting from ``enlarging models'' to ``cost efficiency and sparsity'' (CNBC).

## 8. [AI Regulation/Policy/Social Impact] The grace period for the EU's AI Act will end on August 2, 2026, and the European Commission will gain the power to impose fines on how powerful models are trained and how they operate (Space Daily). The US FTC is considering a draft policy statement on accuracy constraints (Winston Taylor). Canada's Bill C-36 attempts to address AI privacy, but its sufficiency is questionable (Al Jazeera). In the educational field, the University of Chicago Law School has taken steps to ban the use of AI technology in first-year classrooms (CBS News).

## 9. [Integrated analysis: What is happening now] The Apple vs. OpenAI lawsuit is a symbolic case in which competition between AI companies has expanded from "model performance" to a legal battle over "trade secrets and intellectual property." At the same time, OpenAI has been on the defensive as its security chief has left the company and UK authorities have pointed out a jailbreak. In contrast, Anthropic is steadily progressing with the launch of Sonnet 5 and preparations for IPO, and is working to build its corporate image by rolling out the Hard Questions Portal, which differentiates itself by "trust."

The decision by xAI's Musk camp to forcibly migrate Tesla employees to Grok was made after acknowledging the inferiority in terms of performance, and it is clear that the company is prioritizing the enclosure of the corporate ecosystem and data sovereignty over technological rationality. This can also be read as a geopolitical and strategic decision to break away from dependence on OpenAI/Anthropic.

As symbolized by the expression "K-shaped disparity," the disparity in treatment between AI personnel and regular employees is widening, and the withdrawal of the Instagram AI image tool can be said to be the first clear example of social backlash against generated AI content (pressure from copyright and portrait rights organizations) beginning to influence actual product decisions.

In terms of investment trends, the competition for giant AI models has come to a halt, and interest has shifted to cost efficiency and inference efficiency (DeepSeek's in-house chip production, MoE-based efficiency trend). The fact that Google has dropped out of the top five of the Intelligence Index casts doubt on the technological superiority of the Gemini camp, and the three-way structure of Anthropic, OpenAI, and xAI is becoming clearer.

## 10. [Future points of interest] 1. Progress in the Apple vs. OpenAI lawsuit - How much of the technology transfer process within OpenAI will be revealed at the discovery stage? 2. Responses by companies toward the end of the grace period of the EU AI law (August 2, 2026)—Movement to disclose learning data to avoid the risk of fines. 3. Anthropic/OpenAI's IPO preparation competition—S-1 submission contents and valuation negotiations.