anthropic

According to Jinshi reports, Microsoft recorded a $3.2 billion gain in the fourth fiscal quarter due to its investment in Anthropic, combined with expenses related to the voluntary departure plan being lower than expected, resulting in a positive impact of 27 cents on earnings per share (EPS).

Anthropic announced that the Claude Mythos Preview model has made breakthroughs in cryptographic research, discovering improved attack methods against the post-quantum digital signature candidate HAWK. HAWK is a post-quantum signature candidate solicited by NIST to combat quantum computer attacks, which has already passed two rounds of expert review. However, Claude reduced the effective key strength of HAWK-256 from 2^64 to 2^38 in just 60 hours, significantly lowering the theoretical cost of cracking. HAWK has not yet been deployed in practice, and this attack currently does not affect any production systems.Claude also discovered an improved attack against 7-round AES, achieving a speed increase of 200 to 800 times. During the research process, Claude independently completed literature reviews, mathematical reasoning, and experimental validation with limited guidance from cryptographers. The HAWK attack took about 60 hours and had an API cost of approximately $100,000; the AES attack was completed independently by Claude, which generated about 1 billion tokens before proposing core innovative ideas. Anthropic researchers then spent hundreds of hours validating the results. Anthropic stated that AI models can now help identify significant flaws in cryptographic algorithms, and the cryptography community may face bottlenecks similar to those in the software vulnerability field—AI-generated research results far exceed human verification capabilities. Anthropic has collaborated with academic institutions to launch the CryptanalysisBench benchmark and coordinated disclosures with NIST authors and government partners. The research team has achieved preliminary results on algorithms such as LEA and Serpent-128. Anthropic warns that as AI capabilities improve, actual attacks against deployed systems may be discovered in the future, necessitating the establishment of response mechanisms in advance.

According to Jinshi reports, Anthropic CEO Dario Amodei announced that they have signed a supply agreement with Samsung Electronics and SK Hynix.Bloomberg recently reported that senior officials from the South Korean presidential office revealed that Samsung Electronics and SK Hynix will sign agreements involving substantial funds with top American technology companies during Lee Jae-myung's visit. The related cooperation is expected to include long-term storage chip supply contracts, strategic investment partnerships, and memorandums of understanding, with specific scales to be disclosed after the companies officially announce them.Lee Jae-myung will arrive in San Francisco on Friday local time and attend the artificial intelligence summit, where he will hold talks with leaders of American technology companies, including NVIDIA CEO Jensen Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Broadcom CEO Hock Tan.

Anthropic upgrades Claude's voice mode. Paid users can now choose Opus, Sonnet, and Haiku, no longer limited to the faster but less capable Haiku.Users can switch models during conversations and can also switch to text at any time. Claude defaults to the model selected in the last text chat and calls the fastest version in that series. The voice mode can also access Gmail, Google Calendar, Google Docs, and Slack. Users can have Claude check emails, change schedules, or draft replies, and will request authorization before executing.The new version is now available for testing on mobile, desktop, and web. Free users can only use Haiku and one connection tool. Voice now supports multiple languages including English, Japanese, and Korean, but does not include Chinese at this time. This upgrade mainly enhances model reasoning and tool invocation capabilities. Anthropic has not changed the voice model, so the naturalness of speech and interruption experience may not show significant changes.

According to TickerTrends tracking data, as of July, Anthropic's Annual Recurring Revenue (ARR) is approximately 74.1 billion dollars, while OpenAI is about 41.3 billion dollars, with a gap of approximately 32.8 billion dollars. Since the last reporting period, Anthropic has grown by 57.7%, and OpenAI has grown by 65.2%, with the latter's growth rate slightly faster.From the timeline perspective, Anthropic's tracked ARR rose from 10.2 billion dollars in January to 35.6 billion dollars in April, 69.6 billion dollars in June, and 74.1 billion dollars in July; OpenAI increased from 21.4 billion dollars in January to 33.0 billion dollars in May and 41.3 billion dollars in July. Anthropic still significantly leads in absolute revenue scale, but OpenAI's recent growth momentum is stronger.

Michael Kratsios, the Director of the White House Office of Science and Technology Policy, stated that information has been obtained showing that Moonshot AI is distilling Anthropic's Fable model by building a large-scale internal platform for the development of its K3 model, and has switched access methods multiple times to evade detection. He also pointed out that Moonshot AI has purchased GB300 servers and accessed GB300 equipment in Thailand, which may be used for training AI models. Kratsios stated that the U.S. supports legitimate distillation and open innovation but opposes the theft of U.S. technology through large-scale secret industrial distillation methods. Currently, Moonshot AI has not publicly responded to this.

According to The Wall Street Journal, AMD (AMD.O) has signed a significant chip and investment agreement with Anthropic. AMD will invest $5 billion in Anthropic as part of the chip deal.Anthropic will begin purchasing up to 2 gigawatts of AMD (AMD.O) latest generation chips in the first half of 2027.

Artificial intelligence company Anthropic announced on Wednesday that it will make an additional donation of $20 million to the bipartisan advocacy organization Public First Action, bringing its total commitment to $40 million. Anthropic stated that the decision for the additional donation stems from the continuous enhancement of AI model capabilities (including its self-developed Mythos model) and the potential risks it brings.Public First Action is a 501(c)(4) nonprofit organization dedicated to public education on AI transparency and safety. The organization is associated with three super political action committees (super PACs) and plans to support candidates advocating for AI safety across both parties. Anthropic emphasized that the donations are solely for supporting Public First Action's policy mission and cannot be used to influence the election of specific candidates.In a statement, Anthropic noted that AI policy decisions will impact various aspects of public life, including the labor market and national security, and the company hopes to promote "flexible regulation" to manage risks while reaping the benefits of AI. Currently, the super PAC "Leading the Future," which advocates for rapid AI development and light regulation, has received significant funding from supporters including OpenAI co-founder Greg Brockman and venture capital firm a16z. As AI increasingly becomes a topic in elections, public education and lobbying efforts by advocacy groups are continuing to ramp up.

According to market news: Anthropic plans to hold an IPO investor meeting in the coming weeks.

According to the Bloomberg Billionaires Index, DeepSeek founder Liang Wenfeng's net worth has risen to approximately $36 billion, more than doubling from about $16.7 billion previously, surpassing Anthropic co-founder Dario Amodei and OpenAI co-founder Greg Brockman.Previously, DeepSeek completed a $7.4 billion financing round in June 2026, with the company's valuation reaching $50 billion. Bloomberg estimates that Liang Wenfeng holds about 78% of DeepSeek's shares, and his wealth increase mainly comes from the rise in the company's valuation.

PPP market prediction tool monitoring shows that the "Anthropic highest valuation this year" prediction event data on Polymarket indicates that the probability of Anthropic reaching a valuation of $1.5 trillion by the end of 2026 has risen to 78%, up 47% in 24 hours; the probability of reaching $2 trillion has risen to 37%, up 19% in 24 hours.On June 1, Anthropic secretly submitted an S-1 registration statement draft to the U.S. SEC, officially starting the IPO process in the U.S. Recently, research institution SemiAnalysis released a report stating that Anthropic is expected to achieve $1 billion in GAAP operating profit in the third quarter of 2026, with an operating profit margin of 6%.

According to Decrypt, Bitcoin mining company TeraWulf has signed a 20-year lease with Anthropic to build an approximately 401 MW AI data center park at the Justified Data campus in Hawesville, Kentucky, USA, which is expected to generate about $19 billion in revenue.Under the agreement, the first phase of the campus is expected to go live in the second half of 2027 and reach full capacity in early 2028. TeraWulf stated that the lease is expected to receive investment-grade credit rating support. In addition, TeraWulf has sold its 50.1% stake in the Abernathy Joint Venture to a group led by partner Fluidstack, exiting its approximately $450 million investment at a premium.Following the announcement, TeraWulf's stock price surged nearly 14%, and the stock prices of Bitcoin mining companies such as IREN, Hut 8, and Cipher Digital also rose by more than 10%, respectively.

According to reports from South Korean media, Samsung Electronics' foundry business is accelerating its expansion in the global AI chip market. Following last year's acquisition of Tesla's AI chip orders, Samsung is currently advancing production cooperation for customized chips (ASIC) with Meta and Anthropic.It is reported that Meta is negotiating a next-generation ASIC design and production contract with Samsung worth over 10 trillion Korean won. Meta's third-generation AI accelerator "MTIA" plans to shift from TSMC to large-scale production using Samsung's advanced 2-nanometer process, with early chip architecture design being jointly conducted by Samsung's System LSI division to align with its rapid six-month development cycle. In addition, the American AI company Anthropic is also evaluating the use of Samsung's 2-nanometer process to develop exclusive ASICs, aiming to advance the internalization of its AI infrastructure.Industry insiders point out that with the influx of AI chip orders from global tech giants, along with ongoing negotiations for automotive chip foundry agreements with BYD, Samsung's foundry business is expected to approach a medium to long-term order reserve of 50 trillion Korean won, and is likely to turn a profit in operating income in the fourth quarter of this year.

According to reports from ZhDongxi, Alibaba announced today an全面 ban on the Claude series products under Anthropic. All employees are required to uninstall Agent products, including multiple series models such as Sonnet, Opus, Fable, and Claude Code. This ban will officially take effect on July 10. Previously, Alibaba had encouraged employees to freely use external models, including Claude, through a large reimbursement policy.Insiders pointed out that this "reverse ban" is closely related to a series of recent frictions between the two parties. In mid-June, Anthropic accused Alibaba of using a large number of fake accounts to conduct "industrial-grade model distillation attacks" in a letter to the U.S. Senate, and subsequently implemented large-scale account bans and other risk control measures targeting Chinese users. Additionally, developers recently exposed that Claude Code had previously embedded covert detection mechanisms in earlier versions, secretly collecting and tagging data from Chinese time zones and domain names of several Chinese tech companies, including Alibaba. Although Anthropic officials later acknowledged this "experimental" measure and announced its rollback and deletion in the latest version, it still sparked widespread controversy and a trust crisis within the industry.

According to the Financial Times, the U.S. government is in discussions with several AI companies and plans to introduce voluntary industry standards for the release of cutting-edge AI models as early as next week, to prevent advanced technologies from being misused by other countries. The standards will clearly define regulations regarding model performance benchmarks, release timelines, and domestic and international access permissions.Affected by recent tightening of regulatory policies, several leading AI companies have adjusted their operations. OpenAI has postponed the full release of GPT-5.6 at the government's request, allowing access only to a few partners that have passed qualification reviews; two top models from Anthropic have just had their export restrictions lifted after nearly three weeks; Google is also in close communication with the government while preparing a new generation of code models. Additionally, the report mentions that both OpenAI and Anthropic, which are currently under regulatory scrutiny, are actively preparing for their initial public offerings (IPOs).

Anthropic's latest Sonnet series masterpiece, Claude Sonnet 5, has officially been made available for developers through the B.AI API platform. This model has achieved breakthrough progress in code generation and complex agent tasks. With a 1M large context window and innovative adaptive thinking capabilities, Claude Sonnet 5 can more stably and efficiently manage large complex projects. At the same time, to meet the needs of different business scenarios, B.AI has taken the lead in the industry by launching a dual-channel access model of "official direct connection" and "self-selected service providers," fully assisting developers in flexibly balancing extreme performance and operational costs. Starting today, users can log in to the B.AI console to apply for an API key and quickly access the experience. For detailed technical guides and documentation, please refer to the B.AI official website.