Meta
Meta is an American multinational technology conglomerate. It owns Facebook, Instagram, and WhatsApp, among other products and services.
News Heat Trend
Related news
Manus is reportedly pursuing an equity transaction at a $2 billion valuation, with Tencent potentially becoming the largest shareholder
According to sources familiar with the matter, Tencent is in talks to become the largest shareholder of AI agent company Manus. Most existing investors, including Tencent, ZhenFund, and Sequoia Capital China, are discussing supporting a deal at a $2 billion valuation to block Meta's acquisition.
Meta denies AI computing power surplus, stating that renting out computing power can sometimes be more valuable than internal use
: Meta CEO Mark Zuckerberg stated that the company is considering renting out some of its AI computing power to external customers. He mentioned that some external offers are so high that Meta believes renting out computing power could be more valuable than keeping it for internal use.Zuckerberg emphasized that Meta does not have a surplus of AI computing power and that the company is still fully utilizing its existing computing resources. He noted that almost no company in the industry believes its own computing resources are sufficient.According to reports, Meta is exploring two types of cloud service models: one involves hosting proprietary or third-party AI models and charging based on API calls, while the other directly sells basic computing power, similar to the model of AI cloud computing company CoreWeave.Additionally, Meta has recently opened its Meta Model API and started charging for Muse Spark 1.1. Zuckerberg also said that SpaceX's model of temporarily renting out computing power to external companies is "very interesting," and Meta will also evaluate similar high-value transaction opportunities.
based
CoreWeave
Spark
Spark
Meta launches paid AI developer model for the first time
: Meta has launched a paid AI developer model for the first time. (Jinshi)
Spark
Spark
Meta's self-developed AI chip "Iris" is planned to start mass production in September, with a 2027 compute target of 14 gigawatts.
According to Reuters, Meta plans to mass-produce its self-developed data center AI chip "Iris" starting from September, as part of its fourth-generation Meta Training and Inference Accelerators project, to enhance the AI capabilities of platforms such as Facebook and Instagram and reduce reliance on external GPUs such as those from Nvidia and AMD. Internal memos show that Iris completed testing in just 6 weeks with no major defects; Meta plans to deploy 7 gigawatts of computing power this year and increase it to 14 gigawatts by 2027, with its AI infrastructure spending in 2024 potentially reaching up to $145 billion. To secure expansion, the company has signed long-term supply agreements with Samsung Electronics, Sandisk, and Sumitomo Electric to cope with "price increases" and shortages of memory and AI chips.
GPU.Net
Inference
Iris
Node AI
Citrini analyst jukan: It is highly irrational to extrapolate isolated channel conditions to the entire market
: Citrini analyst jukan posted on platform X, stating that some friends asked for his opinion on the Evercore channel survey. Evercore reported that several OEMs admitted they have simultaneously placed orders with both distributors and brokers for the same project due to extended lead times. Even if the three major DRAM manufacturers attempt to eliminate overbooking and duplicate orders, it cannot be completely prevented.He believes that extrapolating the situation occurring in a few isolated channels to the entire market is highly irrational. He also shared comments from Lynx on Meta's storage procurement on the same day: According to Lynx's survey, Meta is willing to purchase memory and flash storage from major suppliers at spot prices; unlike Apple, Lynx does not expect Meta to complain about high prices and anticipates that Meta has already factored in price increases in its capital expenditure plans.
Apple
Lynx
Meta 将在加拿大建设 91.7 亿美元人工智能数据中心
According to Cointelegraph, Meta will build an AI data center with an investment of approximately $9.17 billion and an installed capacity of 1 gigawatt in Alberta, Canada. This will be Meta's first data center in Canada and its 33rd data center globally.
Cointelegraph
Morgan Stanley: Meta's Self-Built GPU Services Signal Supply Shortage, DRAM Shortage Extended to Early 2028
According to TechFlow Research, the Morgan Stanley Weekly Report reinterprets the true meaning of the May SIA data. Meta's development of internal cloud services appears on the surface to be business competition with AWS/Azure, but in essence is a forced move driven by GPU shortage. When the enterprises with the strongest self-build capabilities in the market are all scrambling for capacity, the market is already in a state of extreme scarcity.
Agentwood Studios
Extended
GPU.Net
Node AI
Polymarket new: "Will the US government revoke public access to another major AI model?"
The PPP Prediction Market tool monitors that Polymarket has listed a new event: "Will the US government revoke public access to another major AI model?" The current probability is reported at 33%.The settlement rules state: If the US federal government, by the end of 2026, passes relevant legislation, issues an executive order, implements export controls, or takes any other action that substantially restricts US public access to a major AI model, the market outcome will be "Yes." Otherwise, the outcome will be "No." A "qualifying action" refers to a formal measure taken by the US government, the effect of which is equivalent to completely prohibiting the public from accessing a specific AI model within the United States. Furthermore, the settlement rules emphasize that regardless of the action's true purpose or nominal goal; if the action effectively results in the public being unable to access the model within the US—for example, prohibiting the model from being provided to foreign citizens or governments, as long as the general public cannot access the model through conventional channels within the US—it meets the qualification requirements. Merely excluding access to the model from a single channel is insufficient. Removals of public access that are not caused by any formal action of the US government are not eligible."Mainstream AI model" refers to the flagship, general-purpose large language model or multimodal foundation model developed by one of the following companies: OpenAI, Anthropic, Google (including Google DeepMind), Meta, xAI, Microsoft, Amazon, Mistral AI, DeepSeek, Alibaba, ByteDance, Moonshot AI, and Zhipu AI (Z.ai). Models designed for specific tasks, or those that are outdated, used solely for research or preview purposes, do not meet this criterion.The action can target a single model or a group of models, as long as at least one major AI model becomes inaccessible to the public within the US as a result. A temporary suspension of public access to a model meets this condition. However, if an action has been implemented or a related resolution has been issued, but the public can still access the model before the resolution takes effect, that action does not meet the condition.The information sources for this market are official announcements and information from the US government and the relevant AI companies. However, reliable media reports may also be referenced to form a consensus.The Odaily Seer Channel continues to monitor prediction markets, observing changes before prices are set.
Furthermore
Mainstream
Market
Moonshot
Moonshot
Moonshot
AI Frenzy Continues? US Tech Companies' Data Center Lease Commitments Hit a Record High of $850 Billion
The Kobeissi Letter, citing Bloomberg on X, stated, "US technology companies have committed a record $850 billion over the next few years for data center leases, an increase of $570 billion year-over-year, representing a 204% jump, and a sequential increase of $200 billion, or 31%."According to the report, Meta added $79 billion in new commitments in the first quarter of 2026, up 76% from the previous quarter, bringing its total commitments to approximately $183 billion. Microsoft added $41 billion in the same period, up 26% sequentially, bringing its total commitments to around $197 billion. Oracle leads with total commitments of approximately $250 billion, having locked in several key sites to fulfill its contract with OpenAI.
Center
Hit
Oracle
Record
AI Data Center Crusoe Plans to Raise $3 Billion, Valuation Could Rise to $30 Billion
AI data center Crusoe Energy Systems is in talks for a new round of financing worth approximately $3 billion. Upon completion of this round, the company's valuation is expected to roughly triple from the previous round, reaching the $30 billion range.Founded in 2018, Crusoe initially started with crypto-related businesses before transitioning into an AI infrastructure and data center service provider. It is currently categorized as one of the "neocloud" firms, a new type of cloud computing company focused on providing computational power support for generative AI. The company has signed compute supply contracts with tech giants including Meta and Oracle. As demand for AI infrastructure construction surges, the scale of its compute business continues to expand.According to previously public information, Crusoe completed a funding round of approximately $1.38 billion last year, with a valuation exceeding $10 billion. If this new round is successfully completed, it would become another large-scale capital move in the AI infrastructure sector.Market sources estimate that the final valuation for this round could be around the $30 billion level, but the deal is still in progress and has not yet been finalized. (Reuters)
Center
Crusoe Energy
Market
Oracle
Raise
Rise
Binance Adds 15 New bStocks Tokens as Margin Collateral
according to an official announcement, Binance CrossMargin, PortfolioMargin, and PortfolioMargin Pro have added 15 new bStocks tokens as qualified collateral assets, including Circle (CRCLB), Micron (MUB), NVIDIA (NVDAB), Sandisk (SNDKB), Tesla (TSLAB), SpaceX (SPCXB), Advanced Micro Devices (AMDB), iShares MSCI South Korea ETF (EWYB), Intel (INTCB), Strategy (MSTRB), Lumentum (LITEB), Meta (METAB), Microsoft (MSFTB), Palantir (PLTRB), and Invesco QQQ Trust (QQQB). The corresponding trading pairs are now available for margin trading. This feature is only available to VIP3 and above users in eligible regions, and loans are not currently supported.
Binance
Circle
NVIDIA
Strategy
VIP3
SanDisk’s stock price plummeted 10% due to industry impact, with Gate SNDK contract trading volume and open interest ranking among the highest in the market
SanDisk (SNDK) saw its stock price plunge over 10% today, affected by industry panic triggered by Meta's plan to sell excess computing power and market funds withdrawing profits from the AI sector. According to Gate platform data, SanDisk (SNDK) is currently trading at $1,743.00 (-14.35%), hitting a 24-hour low of $1,693.79. Despite the sharp short-term decline, SNDK’s long-term fundamentals remain optimistic, and related contract trading continues to be active. According to Coinglass data, Gate SNDK contracts recorded a 24-hour trading volume of $82.131 million and an open interest of $22.3788 million, both ranking among the top in the market.Gate's stock sector has established a 7×24 trading service system covering three core markets: US stocks, Hong Kong stocks, and South Korean stocks. It supports over 10,000 US stocks and ETFs, more than 1,500 Hong Kong stocks, and over 1,000 South Korean stocks, collectively covering more than 12,500 global stocks and ETF assets. Users can participate in global stock investments through the Gate unified account using USDT, with support for fractional share trading starting from as low as 0.01 shares and eligibility for stock dividend rights. The platform also supports cross-broker transfer of US and Hong Kong stocks, as well as corporate actions such as stock splits and reverse stock splits, further optimizing the stock investment service experience.
Gate
Tether
Meta Chief AI Officer Responds to Market Misunderstanding: Zuckerberg's Remarks on AI Agent Development Slowdown Not Specifically About Meta
针对有媒体披露 Meta 首席执行官马克·扎克伯格在内部会议上告诉员工，过去四个月的 AI 代理开发并未“以我们预期的速度加速”，Meta 首席 AI官 Alexander Wang 发文回应澄清表示，扎克伯格指的是整个行业在智能代理能力上的整体进展，而非 Meta 自身，他同时透露 Meta 即将发布新一轮 Muse Spark 更新，将显著提升模型的编程能力和智能代理能力，以更具竞争力地对标业内领先模型，并将陆续上线至 Meta AI 及全新 API 平台。
Market
Spark
Spark
Lenovo Executive: No Oversupply of Computing Power
Odaily Odaily News: Ablikim Ablimit, Vice President of Lenovo Group and Chief Strategy Officer of Lenovo China, stated that in the long term, whether in China or overseas markets, there is still significant room for AI computing power demand, and there is no situation of oversupply. He noted that currently, the number of mature, leading AI products on the consumer (C-end) side is limited, and the potential for commercial deployment on the enterprise (B-end) side has not yet been fully released. The growth dividends from historical productivity revolutions have always been driven by the industrial sector, indicating huge long-term demand for computing power.He also pointed out that currently, the AI industry chain faces issues of overly strong supply-side influence and an imbalanced supply-demand cycle. On the news front, Meta's recent opening of its computing power has sparked market concerns about a potential oversupply. (Yicai)
Strategy
Morgan Stanley Rates Meta: Cloud Focus on Compute Rental, Target Price $775
According to TechFlow Research, Morgan Stanley released a research report on July 1 stating that, regarding Bloomberg's report on Meta planning a cloud computing business, it judges that Meta is more likely to choose the lighter path of renting out idle computing power rather than building a full cloud service benchmarking AWS. The report calculates that renting out 250 MW of computing power at $40/watt could be accretive to 2028 earnings per share by approximately 8%, and when the scale reaches 1000 MW, the accretion could reach 33%, but this earnings accretion is viewed as a transitional buffer, not the core logic supporting the rating. Morgan Stanley also mentioned that Meta's self-held computing power will expand to 1.9 GW and 3.4 GW in 2026 and 2027 respectively, providing room for the rental calculations. Morgan Stanley maintains its Overweight rating on Meta with a target price of $775, representing approximately 37.6% upside compared to the closing price of $563.29, while setting the 2027 capital expenditure expectation at $175 billion; if the cloud computing business scales up, there is a possibility of an upward revision in capital expenditure.
Agentwood Studios
Focus
Price
White-haired stock guru Serenity agrees with Wells Fargo that Meta selling excess computing power is a strong signal of AI demand
Odaily Odaily: White-haired stock guru Serenity posted on X agreeing with Wells Fargo's view, stating that the market has completely misinterpreted Meta's stance on selling excess computing power. Wells Fargo stated that Meta's willingness to sell off surplus computing capacity is a positive signal that underlying AI demand and Unit Economics remain robust. Wells Fargo noted: "While Meta is moving forward with this plan, we do not see this as an indication that Meta will reduce capital expenditures (CapEx), nor do we believe that overall demand for computing power has declined."Regarding Neocloud companies, Wells Fargo believes this move further validates the significant opportunity in the AI infrastructure market, while also highlighting the potential for industry mergers and acquisitions, even though Neocloud providers may face certain competitive pressures in the future.
Serenity
Unit
Morgan Stanley Bullish on GOOGL and META, True Valuation Undervalued by Over 30%
According to TechFlow Research, Morgan Stanley released an Internet Tracking Report, noting that Google and Meta's nominal EV/EBITDA multiples appear inexpensive (GOOGL 16.1x, META 8.9x), but after adjusting for stock-based compensation accounting treatment, the true multiple rises from 16.3x to 31.1x (+91%), still lower than the five-year average of 31.6x, implying that the true valuation of internet giants is undervalued by the market by more than 30%.
based
Bullish
Cometa
Meta Pool
MetaDAO
Metadium
Citrini analyst jukan: MediaTek reportedly locks in Meta as ASIC customer after TPU
: Citrini analyst jukan posted on X platform, stating that supply chain sources indicate MediaTek has essentially secured its second ASIC customer beyond Google. If nothing unexpected happens, this customer will be Meta, as previously rumored. Per its usual practice, MediaTek does not publicly comment on individual products, customer situations, or market rumors. Qualcomm currently seems to hold key major clients such as Meta, Microsoft, and ByteDance, but industry insiders believe that MediaTek is not at a disadvantage as it deepens its cooperation with Google and has the opportunity to lock in a second customer. According to semiconductor supply chain sources, based on the order prospects and generational transition pace of AI data centers and ASIC products from major cloud service providers, Google remains the most proactive and aggressive key client. MediaTek not only has two product codenames, Zebrafish and Humufish, but market information and industry insiders familiar with ASICs have confirmed that its participation in the v9 generation Triggerfish is almost certain. This implies that from late 2026 through 2028, and even extending to 2029, MediaTek can stably generate revenue from TPU mass production. Compared to Qualcomm's proposed 2029 cloud AI revenue target of $15 billion, MediaTek, holding multiple TPU ASIC orders, is only a matter of time before reaching the $10 billion level. The industry is also watching whether MediaTek can smoothly secure a second key cloud service provider as a major client. Combining earlier market information and recent supply chain confirmations, MediaTek is still actively cooperating with Meta on ASIC products, with the collaboration likely centered on AI accelerator chips. Meta has recently partnered with both Arm and Qualcomm, but those related products are aimed at CPUs. For its self-developed AI accelerator chips, there have been no clearer collaboration announcements. IC design industry insiders note that Meta’s recent cloud AI development strategy has indeed been relatively chaotic and unclear, with its internal chip development plans undergoing multiple adjustments. Even in the CPU domain, it has sought multiple partners and adopted different approaches. Regarding AI accelerator chips, although Meta has previously formally announced its cooperation plan with Broadcom, supply chain sources indicate this has not interrupted the ongoing cooperation plans between MediaTek and Meta. IC design industry insiders also emphasize that the cloud AI development roadmaps of MediaTek and Qualcomm remain distinctly different. MediaTek is concentrating all its resources on the ASIC business, while Qualcomm plans to advance both customization and standardization simultaneously, covering both AI accelerator chips and CPU product directions.
based
Serenity: Meta isn't cutting capital expenditure due to "overbuilding"; computing power remains constrained
Serenity, known as the "White-Haired Stock Guru," posted on X platform, stating that there is a lot of misleading information regarding Meta cutting capital expenditure due to "overbuilding." Meta would only sell computing power if there were excess capacity, but the current situation seems to be the opposite. Hyperscalers like Google cut their computing power allocation to Meta in March, as Meta's internal projects consumed too many resources, leaving Meta constrained by computing power. Meta was then immediately forced to sign massive contracts exceeding $48 billion with Neoclouds such as CoreWeave and Nebius.Meta only sells computing power when there is excess capacity. In particular, the large contracts it has signed with Neoclouds are largely take-or-pay agreements. Serenity expects Meta's guided capital expenditure to rise, as it is building more independent capacity.
CoreWeave
Serenity
META is building a cloud business to sell excess AI computing power
: According to MSX.COM data, Meta Platforms (META.O) saw its pre-market gains widen to over 8% at one point. The company is building a cloud business to sell its excess AI computing power.
Cometa
Comput3 AI
Meta Pool
MetaDAO
Metadium