All news

SSecurities Times09:41

AI score58

AI short drama concept stock Duiba surged 40.94%, with its AI short drama business profitable for three consecutive months

Duiba, a Hong Kong-listed AI short drama concept stock, surged 40.94% on October 21, with intraday gains once exceeding 50%, closing with a market capitalization of HK$1.1 billion and cumulative gains of over 660% in the past two months. Duiba said its AI short drama business has grown rapidly since its launch in January 2026 and is expected to achieve profitability for three consecutive months in July, August, and September 2026; the business has surpassed SaaS to become its largest revenue source. At the industry level, a total of 490,000 micro short dramas were released from January to September 2026, nearly 15 times the full-year total for 2025, the vast majority of which were AI micro short dramas.

#Industry news#Video#English text is being prepared.

EEast Money09:41

AI score59

CPC Central Committee and State Council Issue Document to Fully Implement the AI+ Initiative; Institutions Suggest Focusing on AI and Data Industries

On October 21, the General Offices of the CPC Central Committee and the State Council issued opinions on promoting the deep integration of scientific and technological innovation with industrial innovation, proposing to fully implement the "AI+" initiative and drive AI to empower all industries in an all-round way. The article states that as of the first half of 2026, China's generative AI user base exceeded 700 million, with a penetration rate of over 50%. Wanlian Securities suggests paying attention to AI application and cloud service providers that benefit from falling inference costs and can embed model capabilities into high-frequency workflows, and focusing on the two major investment themes of the AI industry and the data industry in the medium and long term.

#English text is being prepared.#Industry news#English text is being prepared.

EEast Money09:41

AI score62

Jensen Huang at Berlin GTC: AI Factory Deployment Shortened to as Fast as 14 Days, OpenShell Agent Security Architecture Released

On October 21 local time, NVIDIA CEO Jensen Huang unveiled the DSX POD AI factory solution at the GTC conference in Berlin, saying that when on-site supporting conditions are in place, the deployment cycle can be shortened from about six months to as fast as 14 days; the solution is planned to begin delivery by global partners in early 2027. He also said that the AI chip platform Vera Rubin has entered full production, and under Agent workloads, Vera Rubin NVL72 can achieve up to 30 times the throughput per megawatt of the previous-generation Blackwell platform. NVIDIA also released the OpenShell reference architecture for enterprise Agents, providing runtime isolation and behavior monitoring, with Anthropic, Amazon Web Services, Microsoft, Google Cloud, OpenAI and other companies participating in the collaboration.

#Industry news#Agents#English text is being prepared.

EEast Money09:41

AI score75

Meta earnings call discloses Muse agent progress: over 10 million users in the US and Canada, plans to improve compute efficiency 10x within a year

On its Q3 2026 earnings call, Meta disclosed that its personal AI agent Muse, launched in the US and Canada a month and a half ago, has already had over 10 million users complete setup, with about 25% of users using it daily; shopping is the top use case, with an average order value of over $200. Zuckerberg said short-term compute consumption is enormous and the company plans to improve compute efficiency by more than 10x within a year; if efficiency continues to improve, Muse could achieve high profitability within a few years. Meta's total revenue for the quarter was $64.6 billion, up 26% year over year, full-year capital expenditure narrowed to $135-145 billion, and it announced the introduction of non-voting Class C common stock, equivalent to a 1-for-2 split.

#Industry news#Agents#OpenAI

EEast Money09:41

AI score55

Tesla Q3 Earnings: Revenue $28.05 Billion, Doubling Down on AI and Optimus Robots

Tesla released its third-quarter earnings report, with revenue of $28.05 billion, down 0.2% year-over-year, and adjusted earnings per share of $0.48, beating expectations, while automotive revenue fell 3% year-over-year to $20.56 billion. The company lowered its full-year capital expenditure forecast from $25 billion to $20 billion, with the funds going toward the Optimus production line, AI computing, the Cybercab production line, and the Terafab chip factory in partnership with SpaceX. Tesla said Optimus has begun production at the Fremont factory, and Musk expects it will take at least a year to reach the goal of producing 1 million units per year, and plans to achieve 400 megawatts of AI training compute by the end of the year.

Same story · 11 reports

#Industry news#Embodied AI#English text is being prepared.

EEast Money09:41

AI score82

Anthropic is said to be poised to publicly file its S-1 as soon as Friday, with IPO valuation expectations lowered to the $1.5 trillion range

According to media reports citing people familiar with the matter, Anthropic will publicly file its S-1 prospectus as soon as this Friday, clearing a key procedural hurdle for a potential November listing. The report said its revenue in the first half of 2026 was $16.3 billion, with a net loss exceeding $80 billion, of which $72 billion came from accounting charges related to changes in the value of instruments such as convertible notes; third-quarter revenue was $16.8 billion, up nearly 50% from $11.5 billion in the second quarter. Wall Street had previously expected it to raise up to $100 billion at a valuation of about $2 trillion, but after monthly revenue growth slowed, some investors lowered their IPO valuation expectations to the $1.5 trillion range.

Same story · 4 reports

#Industry news#Anthropic#OpenAI

CCailian Press08:56

AI score82

Anthropic is said to be poised to publicly file its S-1 for an IPO as soon as this Friday

According to media reports citing people familiar with the matter, Anthropic will publicly file its S-1 prospectus as soon as this Friday, clearing procedural obstacles for a potential November listing. The report said its revenue in the first half of 2026 was $16.3 billion, with a net loss exceeding $80 billion, and third-quarter revenue of $16.8 billion; some investors have lowered their IPO valuation expectations from about $2 trillion to a range of $1.5 trillion.

Same story · 4 reports

#Industry news#Anthropic#OpenAI

CCailian Press08:56

AI score57

Tesla Q3 revenue $28.05 billion, capital expenditure $4.73 billion far below expectations

Tesla's third-quarter revenue was $28.05 billion, down 0.2% year over year, with adjusted earnings per share of $0.48 beating the expected $0.44; capital expenditure was $4.73 billion, far below the estimated $7.95 billion, and the full-year capital expenditure forecast was cut from $25 billion to $20 billion. The funds will go toward the Optimus production line, AI computing, the Cybercab production line, and the Terafab chip factory in partnership with SpaceX. Optimus has begun production at the Fremont factory, and Musk said reaching an annual output of 1 million units may take a year or longer. Tesla plans to achieve 400 megawatts of AI training compute by the end of the year, and active FSD subscribers have risen to 1.74 million.

Same story · 11 reports

#Industry news#Embodied AI#English text is being prepared.

CCailian Press08:56

AI score67

Meta Q3 2026 Earnings Call: Muse Tops 10 Million Users in North America, Plans to Boost Compute Efficiency 10x Within a Year

Meta's total revenue for Q3 2026 was $64.6 billion, up 26% year over year, with earnings per share of $3.83 and Q4 revenue guidance of $73 billion to $76 billion. Zuckerberg said on the earnings call that the Muse personal agent has been configured by more than 10 million users in the US and Canada within a month and a half of launch, with about 25% of users using it daily; shopping is the top use case, with an average order value of over $200; the company plans to improve Muse's compute efficiency by more than 10x within a year and achieve substantial profitability within a few years. Meta also expects Reality Labs' losses to narrow to about $15 billion next year, and full-year capital expenditures to narrow to $135 billion to $145 billion.

Same story · 11 reports

#Industry news#Agents#OpenAI

336Kr · AI08:35

AI score85

PHP Exposes Two High-Severity Vulnerabilities Rated 9.2; GPT-5.6 Sol Participated in Discovery

PHP has exposed two high-severity vulnerabilities, CVE-2026-109618 and CVE-2026-109619, with a severity rating of 9.2. In the worst case, attackers can execute code remotely without logging in, and almost all versions since PHP 5.4 are affected. The official fix versions PHP 8.5.12, 8.4.27, 8.3.36, and 8.2.35 have been released, and PHP 8.1 and earlier versions are no longer supported. OpenAI said that the security company Calif, under the Patch the Planet initiative, used GPT-5.6 Sol to discover one of the critical vulnerabilities, while Wiz also independently discovered and reported both vulnerabilities. OpenAI led coordination of the patch with PHP maintainers, Wiz, and Cloudflare.

#Safety / alignment#OpenAI#English text is being prepared.

JJiwei.com08:35

AI score53

After NVIDIA's de facto acquisition of Groq, Groq's next-generation chip is reportedly set to abandon SRAM in favor of 3D stacked DRAM

Foreign media reports say that Groq, which has been de facto acquired by NVIDIA, is evaluating the introduction of 3D stacked DRAM packaging technology in its next-generation chips. The approach is called 3DS and uses TSVs to drill thousands of microvias inside DRAM chips for vertical interconnection. Unlike HBM, which pursues data exchange rates, 3DS focuses on expanding storage capacity while maintaining the existing standard DRAM interface protocols. Groq's LPU architecture relies on SRAM with very small capacity and faces a capacity bottleneck when dealing with models with hundreds of billions of parameters. 3D stacking is intended to solve the problem of not being able to store enough while maintaining speed advantages.

Same story · 2 reports

#Industry news#English text is being prepared.#English text is being prepared.

CCaixin08:34

AI score73

Shares of AI compute cloud provider CoreWeave fell more than 30% in two weeks, and its high-leverage compute expansion model is being questioned

Shares of U.S. AI compute cloud provider CoreWeave fell more than 30% in two weeks, closing at $62.13 per share on October 20 and still down more than 20% from a week earlier; Oracle's stock fell 9.20% over the past week. According to S&P data, on October 15 CoreWeave's bonds maturing in 2032 traded at 77% of face value, and five-year CDS spreads for Oracle, SoftBank, Amazon, Meta and Nvidia all reached historic highs. Australia's NextDC fell more than 40% in a single day, and the listing processes of data center developers Firmus and Nscale also ran into obstacles.

#Industry news#English text is being prepared.#English text is being prepared.

TThe New Stack08:33

AI score73

AI discovers PHP core vulnerability PHP2Shell, affecting about 70% of websites, taking two months to fix

Wiz Research and Calif.io, with the help of frontier models such as GPT-5.6 Sol and Gemini 4 Argon, discovered two critical vulnerabilities in the PHP engine, collectively named PHP2Shell, with a CVSS v4.0 score of 9.2, allowing unauthenticated code execution on WordPress, Drupal, and Joomla; at least one vulnerability affects all PHP versions over the past decade. Wiz said a single researcher using frontier models completed in about a week work that previously required a team several months; the vulnerabilities were reported on August 24, and fixed versions PHP 8.5.12, 8.4.27, 8.3.36, and 8.2.35 were released on October 21. OpenAI participated in the coordination under the Patch the Planet program and proposed a five-phase AI vulnerability disclosure response framework, while committing to provide $1 billion in subsidized access credits through Daybreak.

Same story · 6 reports

#Safety / alignment#OpenAI#Google

SSiliconANGLE08:32

AI score63

One-Year-Old AI Research Lab Flapping Airplanes Raises $1 Billion, Reportedly Valued at $6 Billion

AI research lab Flapping Airplanes announced it has closed a new $1 billion funding round, with Bloomberg reporting a valuation of $6 billion; the company is about a year old and has no commercial product yet. In its first year the company focused mainly on developing new GPU primitives, saying its megakernel framework can complete full training, inference and game running within a single kernel, allowing memory-bound inference and compute-bound training to overlap on the same chip. The Series A round was co-led by Index Ventures and Lux Capital, with Sequoia Capital and GV, which participated in the $180 million seed round in January, continuing to invest, and Nvidia also among the institutional investors.

#Industry news#English text is being prepared.#English text is being prepared.

SSiliconANGLE08:32

AI score78

OpenAI to Launch GPT-6 Astra and Other Models on Google Cloud

OpenAI announced it will offer its latest large language models through Google Cloud's Model Garden, with the flagship model GPT-6 Astra launching first, followed by GPT-6.1 Sol and GPT-6 Luna. Astra achieved a perfect score on ExploitBench and has been security-optimized for enterprise use; Sol and Luna are priced at 5x and 100x lower token cost than Astra, respectively. The three models will be generally available in Model Garden within weeks, Google Cloud customers can pay with existing long-term consumption commitments, and the two sides also plan to subsequently integrate Confidential Compute, Antigravity, and Gemini Agent.

Same story · 13 reports

#Industry news#OpenAI#Google

TTechCrunch AI07:49

AI score85

OpenAI and Google Wiz use AI to discover high-severity vulnerabilities that have existed in PHP for over twenty years

The U.S. National Security Agency warned that PHP contains two high-severity vulnerabilities that could allow attackers to break into a server's file system or database. Patches have been released by the PHP Foundation, and websites that cannot be patched immediately should be taken offline temporarily. OpenAI disclosed that under its "Patch the Planet" initiative, security testing firm Calif used GPT-5.6 Sol to discover one of the vulnerabilities; the Wiz team used the Wiz Atlas security agent, Gemini 4 Argon, and GPT-5.6 Sol to discover multiple PHP vulnerabilities. These vulnerabilities have existed in all versions since PHP 5, released in 2004. The official recommendation is to upgrade to the current version, PHP 8.5.12; PHP 8.1 and earlier versions are no longer supported.

#Safety / alignment#OpenAI#Google

TThe Verge AI07:48

AI score62

Report says Texas data center expansion could erase decades of air quality improvements

A report by the Houston Advanced Research Center (HARC) says that if all planned data centers in Texas are connected to the grid and powered under the expected grid structure, fine particulate pollution from the power sector will more than triple within five years, reaching the highest level since 2006; haze-forming NOx emissions will increase by 136%, approaching 2004 levels; and carbon dioxide emissions will be 90% higher than in 2025. The report models three data center energy consumption forecasts and notes that if data centers are paired with self-built renewable energy, NOx emissions could be as much as 10 times lower, but if gas-fired generation is built on-site to avoid grid interconnection queues, pollution could instead be the highest. Texas currently ranks first in the U.S. data center market with 26 GW of under-construction and existing capacity.

#Industry news#English text is being prepared.

TThe Paper07:47

AI score51

arXiv Implements Submission Limits Starting October 1: Each Author Limited to Two Papers per Month

The preprint platform arXiv has implemented submission limits starting October 1, capping each author at a maximum of two papers per month. arXiv scientific director Steinn Sigurðsson said the surge in submissions is mainly driven by the large volume of low-quality manuscripts produced by AI tools; uploads reached 40,000 in September, compared with about 30,000 per month from March to July. The platform says the move is intended to protect nearly 300 volunteer moderators; currently nearly 4,000 manuscripts are still queued for release, and some scholars have noted that co-authors taking turns submitting can circumvent the limits.

#English text is being prepared.#arXiv#English text is being prepared.

TTechCrunch AI05:48

AI score59

Meta says its AI app Muse has over 10 million users, with shopping among the main uses

Meta said at its third-quarter earnings call that its AI app Muse now has more than 10 million users, with about 25% of registered users using it daily. Zuckerberg said that among active users who shop using the Muse agent, more than a quarter spend an average of over $200 per transaction, and that Muse enables agentic shopping through a Plaid connector. He also said Muse has achieved product-market fit, and that the next step is to scale up, grow revenue, and optimize compute costs, with cost efficiency expected to improve more than 10-fold over the next year.

Same story · 11 reports

#Industry news#Agents#OpenAI

TThe Verge AI05:46

AI score62

Meta plans to issue a new class of non-voting shares to finance AI investment

In its third-quarter earnings report, Meta announced it will issue a class of non-voting shares to support AI initiatives, potential acquisitions, and employee compensation, with transactions beginning in 2027. The company expects capital expenditure this year to be between $135 billion and $145 billion, higher than its previous forecast of $130 billion to $145 billion. CFO Susan Li said the new shares will expand financing tools and provide investment flexibility for the business momentum brought by AI.

Same story · 11 reports

#Industry news#English text is being prepared.#OpenAI