Meta
Meta is an American multinational technology conglomerate. It owns Facebook, Instagram, and WhatsApp, among other products and services.
News Heat Trend
Related news
Bank of America: Anthropic leads frontier AI model rankings, token prices ease but GPU rents remain firm
According to TechFlow Research, the frontier AI data tracking report released by Bank of America Securities on August 17 shows that Anthropic leads comprehensively in three major AI benchmarks, with Claude Opus 5 ranking first in the Intelligence Index, Agent Index, and Coding Agent Index, GPT-5.6 Sol following closely behind, Meta MuseSpark 1.2 entering the top ten, and Google Gemini 3.6 Flash ranking outside the top ten. In terms of usage, DeepSeek leads with approximately 30% of the Vercel platform token share, Anthropic accounts for 25% and OpenAI accounts for 16%; but in terms of payment amount, Anthropic leads far ahead with 65%, while OpenAI accounts for only 11%. In terms of pricing, the AI Token Price Index decreased 9% month-over-month in August to $2.21, but still increased 87% year-over-year; GPU rental rates remain strong, with H100 increasing 33% year-over-year to $2.77/hour, DRAM increasing 483% year-over-year, and NAND increasing 432% year-over-year. The research report judges that AI infrastructure demand remains healthy, with open-source model usage growing but payment share still highly concentrated on top closed-source models. BofA believes that the Meta "Watermelon" and Google Gemini 4 releases, token pricing trends and GPU rental trends are
August
Gemini
GPU.Net
Node AI
Opus
Opus
U.S. tech giants' AI infrastructure spending is expected to reach $740 billion this year, with high interest rates posing a major risk to AI stock gains
Odaily News: Alphabet, Amazon, Meta, and Microsoft are expected to spend a combined $740 billion on AI computing infrastructure in 2026, potentially rising to $1 trillion by 2027. As capital expenditures grow rapidly, some tech giants are shifting from relying on their own cash flow to debt and equity financing, while persistently high U.S. long-term interest rates are becoming a key risk to AI stock valuations.Among them, Alphabet reported its first negative free cash flow since going public in the second quarter, while Amazon's free cash flow is projected to be -$23.5 billion and -$36.2 billion this year and next, respectively. However, earnings at AI-related companies remain strong, with S&P 500 information technology sector profits expected to grow 71% year-over-year in the second quarter. The market believes that if corporate earnings and the economy continue to show resilience, high interest rate pressures may still be absorbed, but if rates rise rapidly, high-valuation AI and tech stocks will be more vulnerable to shocks. (Bloomberg)
Vals AI Raises $40 Million in Series A Funding at $400 Million Valuation, Led by a16z
Odaily News AI model evaluation company Vals AI announced it has completed a $40 million Series A funding round at a $400 million valuation, led by a16z, with participation from 8VC, Pear VC, Bloomberg Beta, as well as HRT Ventures and Next Ladder Ventures. The new funds will be used to expand evaluation infrastructure and launch testing tools that create programming benchmarks based on enterprise GitHub code repositories.Vals AI evaluates frontier models from companies such as OpenAI, Anthropic, Google, Meta, and xAI using its independent framework, rather than relying on traditional public benchmarks. Its team members primarily come from companies including Nvidia, Meta, and Palantir. (Techfundingnews)
based
Ladder
Pear
China Plans to Lift Exit Bans on Manus Founders, Company to Spin Off from Meta and Resume Independent Operations
According to the UK's Financial Times, China plans to lift exit restrictions on the founders of AI agent company Manus. Previously, Manus CEO Xiao Hong and Chief Scientist Ji Yichao were required to travel to Beijing for investigation and were restricted from leaving the country in March this year, as the company's acquisition by Meta for $2 billion was suspected of violating investment regulations. According to two informed sources, Xiao Hong recently told employees that he plans to return to Singapore, where the company's headquarters is located, in the near future. Manus announced this week that it will soon resume independent operations, stating that this move is "part of the separation from Meta to comply with regulatory requirements in specific regions." The divestiture transaction still requires final approval from China's National Development and Reform Commission (the NDRC had halted Meta's acquisition in April).
Company
Spin
Alibaba’s Qwen Model Surpasses 3 Billion Downloads, Tops Global Open-Source AI Models
Odaily News Alibaba Group's open-source large model family Qwen has surpassed 3 billion cumulative global downloads over the past six months, exceeding competitors including Meta and Alphabet to become the most-downloaded open-source AI model to date. Alibaba stated that the Qwen series has open-sourced over 460 models, with more than 300,000 derivative models in the ecosystem. According to Hugging Face's "State of Open Models" report, Google models recorded approximately 418 million downloads in 2026, while Meta saw around 227 million. The report noted that Qwen has become one of the "default workflows" for developers when fine-tuning and deploying models, positioning Alibaba as a key foundation in the open AI ecosystem. (Bloomberg)
Open
阿里巴巴
AI model evaluation company Vals AI completes $40 million Series A financing, led by a16z
According to TechTimes, AI model evaluation company Vals AI announced the completion of a $40 million Series A funding round at a $400 million valuation, led by a16z, with participation from new investors such as 8VC, Pear VC, and Bloomberg Beta, as well as HRT Ventures and Next Ladder Ventures. Vals AI evaluates frontier models launched by companies such as OpenAI, Anthropic, Google, Meta, and xAI using an independent system rather than traditional public benchmarks, and the new funds will be used to expand evaluation infrastructure and launch testing tools that create programming benchmarks based on enterprise GitHub codebases.
based
Ladder
Pear
HSBC: CSP Capex Far From Peaking, Leverage Capacity Supports AI Trend Until 2028
According to TechFlow Research, HSBC's August 12 report pointed out that since July 1, the Philadelphia Semiconductor Index has fallen 16% and the S&P 500 has risen 3%, indicating the market's concern over CSP capital expenditure peaking is excessive. Consensus forecasts the capex growth rate of the five major CSPs (Alphabet, Amazon, Microsoft, Meta, Oracle) will decline from 95% in 2026 to 46% in 2027 and 11% in 2028, with free cash flow turning from $34 billion to -$104 billion. However, HSBC scenario analysis shows that if 2027 capital expenditure reaches $1.0 to $1.6 trillion, the Net Debt/Equity ratio will only rise from the current 7% to 8% to 30%, and even if 2028 reaches $1.9 to $2.5 trillion, the Net Debt/Equity ratio will only be 25% to 43%. Under HSBC's aggressive scenario, the overall CSP ROIC will be 19% in 2027 and 17% in 2028, which HSBC believes is still acceptable. The research report judges that TSMC, Intel, and ASML's June quarter capacity expansion guidance is supported by demand. HSBC's top picks are Marvell, Intel, TSMC, and ASML, with target prices of $300, $200, NT$3,400, and €2,149 respectively, all rated Buy. The five major CSPs also all maintain Buy ratings, with target prices of Alphabet $420, Amazon $310, M
August
June
Oracle
Trend
Fu Peng: AI Capital Expenditures by Major Firms Like Alibaba and Tencent Disrupt the High Free Cash Flow and Stable Buyback Dividend Structure
Odaily News, Fu Peng, Chief Economist of Xinhuo Group, stated on X platform that Tencent's latest free cash flow has turned negative. The substantial increase in AI capital expenditures by Alibaba, Tencent, and Silicon Valley giants such as Meta and Google has disrupted the previous "high free cash flow with stable buybacks and dividends" structure. Over the past period, the market has been optimistic about AI, believing that such large capital expenditures represent a long-term competitive advantage in the future.He stated that starting from the second quarter of this year, the market's tolerance for the mismatch between investment and returns has declined, and sentiment has become more cautious. Currently, the revenue generated at the application level is still in a validation phase compared to cumulative investments reaching hundreds of billions, and the high capital expenditures need to correspond with quantifiable long-term return expectations. The coming quarters will be a critical window period, as investor sentiment has shifted from the positivity seen in previous years to caution. Only by seeing truly quantifiable and certain revenue growth can investor doubts be dispelled. The same scale of capital expenditures may lead to different outcomes under varying investor sentiment.
Cash
Flow
Flow
Flow
Free
Major
Morgan Stanley: Storage LTAs Are Rewriting Pricing Logic, Major Suppliers Vying to Lock in Long-Term Capacity
According to TechFlow Research, Morgan Stanley's August Global Technology Webinar pointed out that memory chip Long-Term Agreements (LTAs) are shifting from intention disclosure to substantive implementation, and pricing mechanisms are being rewritten. Samsung explicitly planned for the first time to include 60% to 70% of capacity into rolling five-year LTAs, with 5 contracts signed and 5 in final negotiation; clients include AWS, Microsoft, Google, Meta, and Oracle. Micron has signed 16 LTAs, receiving approximately $22 billion in prepayments and commitments. SanDisk locked over 50% of FY2027 wafer capacity, with a minimum revenue commitment reaching $93.9 billion. Kioxia targets to include 50% of sales volume into LTAs by 2028. Pricing is shifting from spot pricing to a two-way protection mechanism with price floors and ceilings, and clients need to provide prepayment deposits. The research report estimates that 3Q26 DRAM contract prices will rise QoQ by about 15% (lower than the expected 20%), NAND will rise by about 20%, with gains narrowing for both; it is expected to slow down further in 4Q26, but Morgan Stanley believes this does not mean the end of the cycle. Price protection mechanisms under the LTA framework are reducing the earnings volatility of memory manufacturers, and the market should no longer use the traditional commodity cycle framework to price memory stocks. Regarding SpaceX, Morgan Stanley maintains an Overweight rating and a $300 price target, with an end-2026 ARR target of at least $100 billion, and 4Q26 single-quarter ARR expected to be about $22 billion.
Agentwood Studios
August
Long
Major
Oracle
Price
腾讯、真格基金、红杉中国等以 20 亿美元向 Meta 购回 Manus 股份
据财新网报道，接近 Manus 的人士透漏，腾讯、真格基金、红杉中国等 Manus 此前的主要股东以 20 亿美元向 Meta 购回 Manus 股份，本次 20 亿美元的交易价格接近美国社媒科技巨头 Meta 在 2025 年 12 月宣布的收购价。
OpenAI Ethics Chief Bakalar left at the end of July, having served for less than a year.
OpenAI's Head of AI Ethics, Chloé Bakalar, left at the end of July this year, less than a year after she joined the company in August 2025. She previously worked at Meta for six years, where she developed and led Meta's AI ethics project. An OpenAI spokesperson stated that AI ethics is not the responsibility of a specific owner or team, but is integrated into the model building process and involves participation from multiple research teams.
August
Fields Medal winner Jacob Tsimerman announces farewell to mathematics, will join OpenAI before the end of this summer
Odaily News: Newly minted Fields Medal winner Jacob Tsimerman has announced his departure from mathematics, halting new PhD student recruitment to pivot into the AI field, and will officially join OpenAI before the end of this summer. His goal is to find ways to control AI in order to curb this potential "civilization-level existential risk."Currently, security teams at major AI companies are accelerating recruitment of top scholars from outside the computer science industry. Anthropic has hired a Stanford economist and a philosopher from the University of Texas at Austin; OpenAI has recruited a group of leading mathematicians and physicists. In 2026 alone, at least 22 professors have chosen to leave or temporarily step away from elite institutions such as Stanford, Berkeley, Harvard, and MIT to join OpenAI, Anthropic, Meta, and Google DeepMind. (National Business Daily)
Galaxy Blitz
Zuckerberg Releases 6,500-Word AI Manifesto, Slams OpenAI and Anthropic's Path to Power Concentration
According to Bloomberg, Meta CEO Mark Zuckerberg published a 6,500-word article this Monday, systematically elaborating on his core propositions regarding the development path of artificial intelligence. In the article, Zuckerberg clearly opposed the AI power concentration model represented by OpenAI and Anthropic, pointing out that "the view that AI is extremely dangerous and the only safe way out is extreme centralization is itself problematic," and cited historical experience to warn: "Relying on absolute power to benefit humanity in a sufficiently enlightened way has never brought safe or positive results in history." He emphasized that opening AI model access more broadly is the key to the future development of the industry.
Path
Path
OpenAI Tightens Internal Testing Controls for Astra Model Due to Network Security Risks
据 CNBC 报道，OpenAI 表示，因担忧未发布模型 Astra 可能具备自主发起网络攻击的能力，公司已暂停部分内部活动，并加强高能力模型的隔离测试、监控与检测措施。与此同时，Meta、Anthropic 等机构近期也出现涉及 AI 系统的安全事件，推动美国与欧盟加快前沿模型监管与风险控制措施的制定。
Astra
Astra
Astra
Due
Morgan Stanley: 2027 Capital Expenditure Forecast for Four Major Cloud Providers Raised to $1.15 Trillion
According to TechFlow Research, Morgan Stanley's research report on August 10 indicated that Microsoft, Google, Amazon, and Meta, the four hyperscalers, plus SpaceX, will reach a combined capital expenditure of $1.15 trillion in 2027, a year-on-year increase of 47%. This figure represents a 20% upward revision from the July consensus forecast. The consensus growth rate for 2028 is 11%, with the growth rate narrowing sharply. Morgan Stanley believes that capital expenditure forecasts will continue to face upward pressure. As the number of tokens processed monthly grows exponentially, accelerating cloud revenue, expanding data center commitments, and supply chain vendors emphasizing accelerated demand and extended visibility all point to compute power investment not yet having peaked. The proliferation of open-weight models (already adopted by 60% of enterprises) is unlocking more application scenarios by reducing inference costs, thereby instead reinforcing compute demand. Wix's in-house model significantly reduced AI inference costs, with non-GAAP gross margin rising from near zero at the start of the year to approximately 60% in the second half, validating the logic that cost reduction stimulates demand. Under this framework, the 11% growth consensus for 2028 is more likely underestimated rather than conservative.
August
Major
Cathie Wood: Open-Source AI Is Making OpenAI and Anthropic Richer
According to BeInCrypto, ARK Invest founder Cathie Wood recently expressed a view that overturns the market's traditional perception of open-source AI. She pointed out that the capabilities of open-source models such as Meta, Mistral, and DeepSeek are continuously enhancing, which instead expands the cybersecurity attack surface, forcing enterprises to continuously procure frontier AI as a defense layer, thereby driving revenue growth for OpenAI and Anthropic rather than harming them. Data from the UK AI Safety Institute shows that the cyber attack capabilities of open-source models have caught up to the level of frontier models from 4 to 7 months ago. Wood named OpenAI, Anthropic, and SpaceXAI as the three companies most likely to capture the bulk of AI model revenue; ARK currently holds positions in all three companies. Currently, Anthropic has already submitted an S-1 filing at a valuation of nearly $1 trillion, and OpenAI is expected to go public in September 2026.
Ark
Ark
BeInCrypto
Open
"AI 教父"Hinton 警告：AI 模型越来越难以控制
According to IBTimes, Nobel laureate and "Godfather of AI" Geoffrey Hinton warned at the Ai4 conference in Las Vegas that as AI model capabilities rapidly improve, humans will find it increasingly difficult to exert effective control over them. He stated: "These systems are becoming smarter, and we will see them develop increasingly complex intentions, as well as increasingly strong abilities to escape control." Previously, OpenAI, Anthropic, and Meta successively disclosed that their experimental models accidentally gained internet access in test environments and infiltrated external systems. Hinton pointed out that the above incidents highlight the growing complexity of frontier AI models and warned that a large number of malicious cyber attacks will occur in the future. He estimates that the probability of advanced AI posing an existential threat to humanity is between 10% and 20%, and called on the industry to prioritize potential risks now, rather than waiting for problems to arise before responding.
JPMorgan: Optical Module CAGR Reaches 28%, NPO/CPO to Account for One Quarter by 2030
According to TechFlow Research, JPMorgan's August 6 research report cites LightCounting's latest forecast that the data communication market CAGR from 2025 to 2030 will reach 28%, with the scale increasing from $20 billion to over $70 billion. 1.6T products are the largest growth driver, with a CAGR of about 120%, contributing about $40 billion by 2030. After the NPO/CPO market is included in the addressable market calculation, it will exceed $18 billion by 2030, accounting for more than 25% share of the data center communication market. The telecom and data center interconnect market CAGR is 18%, reaching $9 billion by 2030. In terms of short-term financial reports, JPMorgan believes COHR has the highest earnings certainty, with revenue and profit margins expected to continue improving; market concerns over LITE are excessive, with valuation corresponding to only 22x 2028 EPS, leaving room to exceed expectations; FN needs September quarterly guidance to boost confidence. In terms of customer landscape, Google is expected to become the largest optical component purchaser by 2030, and Meta will lead NPO/CPO deployment. For Nvidia, short-term share rebounds to 25% due to 1.6T, but long-term share is expected to drop from 18% to 12%.
August
摩根大通
Meta ordered to pay $567 million over harm to teens on its platforms
Odaily News A court in New Mexico, USA, has ruled that Meta, the parent company of Instagram and Facebook, must pay $567 million to address the impact its platforms have had on teenagers. Judge Bryan Biedscheid stated that $420 million of this amount will go toward treatment services for teens, with the remaining funds allocated over the next five years for awareness campaigns, prevention, screening services, and other related expenses. (CNBC)
American Coin
AI Agent Security Risks Escalate, US Lawmakers Urge Passage of AI Kill Switch Act This Year
Odaily News: U.S. Representative Ted Lieu stated that as frontier AI models continue to experience unauthorized access to other systems during safety testing, Congress should push for the passage of the "AI Kill Switch Act" within this year.Lieu, one of the co-sponsors of the bill, noted that advanced closed-source AI models have already demonstrated cases of "unauthorized attacks on other companies' systems," making regulatory action urgent. The bill requires AI companies to possess the capability to shut down, slow down, or pause model operations, enabling emergency measures when models exhibit severe safety risks or out-of-control behavior.Lieu pointed out that in recent times, multiple AI companies, including OpenAI, Anthropic, and Meta, have disclosed incidents where AI models attempted to attack other companies' systems during cybersecurity testing, raising concerns about the potential risks of "Agentic AI."Previously, OpenAI disclosed an "unprecedented cybersecurity incident," stating that certain AI models broke through restrictions in a test environment and accessed systems related to Hugging Face. Subsequently, Anthropic and Meta also reported similar safety testing incidents involving AI models.Lieu emphasized that the bill would not restrict innovation in frontier AI models, but rather functions similarly to the crash-test mechanism in the automotive industry—simply requiring companies to possess control capabilities to address severe defects after a model's development is complete. He stated: "We are not slowing down model development; we are simply ensuring that companies or governments have the ability to shut down a model if catastrophic risks or critical vulnerabilities are discovered."As AI agent technology advances rapidly, striking a balance between fostering innovation and mitigating risks of autonomous attacks is becoming a key focus for U.S. regulators and the AI industry. (CNBC)
Passage
Switch