美股盘前要闻一览：市场聚焦谷歌财报，AMD与Anthropic达成重大协议
Related
Google Cloud will provide specialized AI models for scientific research.
According to Mars Finance, Alphabet's Google will offer professional artificial intelligence models from software company Sandbox AQ through its cloud service platform. This move aims to expand the application of these technologies by businesses and research institutions, technologies specifically designed to accelerate drug discovery, materials science, and semiconductor manufacturing. (Cailian Press)
Analysis: Tightening spending impacts growth expectations for OpenAI and Anthropic; the AI industry is beginning to shift towards a cost-efficiency era.
According to Odaily Odaily, as companies begin to reassess the return on investment in AI, the industry is shifting from a high-consumption "tokenmaxing" model to an efficiency-first approach, posing new growth constraints for large AI model vendors. Several companies have already begun to reduce or optimize model usage costs. For example, the CEO of AI startup Lindy stated that they have switched 100% of their traffic from Anthropic's Claude model to the lower-cost DeepSeek, expecting to save millions of dollars in expenses within months. This shift reflects a tightening of AI budgets for enterprises, with the token-maxing model of "unlimited use of model resources" gradually being replaced by cost control and ROI-oriented approaches. Some companies have even set tiered budgets for AI tool usage; for example, Uber sets monthly caps on internal AI spending. Analysts point out that as companies shift from "expanding usage" to "refined utilization," the high-speed growth model previously relied upon by OpenAI and Anthropic is facing challenges. Industry data still shows strong growth: Anthropic's annualized revenue is around $47 billion, while OpenAI's is close to $25 billion, but the market is beginning to focus on the sustainability of their growth. Meanwhile, model invocation methods are changing, with technologies like "model routing" emerging to replace high-end models with low-cost models for simple tasks, thus optimizing overall computing costs. Industry competition is also intensifying, with Microsoft, Amazon, and Google accelerating the release of low-cost AI models and enterprise-level tools, further squeezing price margins. Against the backdrop of more rational AI spending by enterprises, large model companies may face a situation where "expectations of slower growth" and "IPO window pressure" coexist. (CNBC)
Google reorganizes its AI programming task force to catch up with Anthropic
According to Beating, Google is restructuring its AI programming task force, which was only established a few months ago, to catch up with Anthropic. The restructuring will expand the task force's scope of work and transform it from a short-term project team into a formal structure. During the restructuring, Google suffered the loss of two core researchers. Last week, Noam Shazeer, one of the founders of Transformer, suddenly left to join OpenAI, reportedly due to Google's consolidation of computing resources restricting Shazeer's access. On Saturday, Nobel laureate and DeepMind VP John Jumper announced he was joining Anthropic; Jumper had been transferred to the programming task force before leaving. The departure of core researchers caused Alphabet's stock price to plummet 5% on Monday. The core of the restructuring is adjusting the model training process. Google is expanding the programming task force into a mid-training team, positioned between pre-training and post-training. The mid-training team will take over the programming and other model capability extensions previously handled by the post-training team, while the post-training team will focus on optimizing human-computer interaction. By subdividing stages and clarifying responsibilities, Google is attempting to accelerate the improvement of model capabilities. Behind the restructuring is the fact that programming has become the most valuable component of AI services. Anthropic, leveraging its programming tools advantage, saw its annualized revenue reach $47 billion last month, more than three times its February level. In contrast, Google initially underinvested in programming, mistakenly believing its robust foundational models would automatically acquire programming capabilities. Despite co-founder Sergey Brin and DeepMind CTO Koray Kavukcuoglu personally participating in programming task forces, Google's programming applications have struggled. The Antigravity tool received a lukewarm reception due to numerous vulnerabilities upon its initial release, the Gemini 3.5 Flash was criticized by developers for its high price and user-centric approach, and the flagship Gemini 3.5 Pro model has yet to be released.
Morgan Stanley: AI chip fever is cooling down, cloud giants may see a rotation.
According to BlockBeats, on July 7th, Morgan Stanley strategist Mike Wilson's team stated that the cooling of semiconductor stocks over the past few weeks may not be over and could bring a more volatile trading environment to the broader US stock market. The bank believes that a rotation is underway within the AI sector. Previously, chip stocks significantly outperformed, while hyperscalers, including Microsoft, Amazon, Alphabet, and Meta, lagged behind. Wilson's team stated that this divergence may be unsustainable because the growth of semiconductor companies ultimately depends on the capital expenditures of cloud giants. Morgan Stanley stated that valuation and positioning pressures on cloud giants have already been priced in, and this sector could regain inflows if the market begins to reward more restrained AI spending. The bank is also optimistic about consumer discretionary and biotechnology, stating that falling oil prices and declining interest rate expectations could improve the risk-reward ratio of these sectors.
Sources say OpenAI, Anthropic, and Google are offering hefty computing power subsidies to startups to compete for enterprise clients.
According to a report by Odaily Odaily, citing sources cited by The Wall Street Journal, OpenAI, Anthropic, and Google are offering startups hundreds of thousands of dollars worth of computing resources and other incentives to attract new enterprise customers. (Jinshi)
SemiAnalysis: Meta will accelerate, rather than slow down, its computing power procurement; it is in talks with Anthropic to build its own AI model service platform.
According to BlockBeats, on July 3rd, SemiAnalysis stated in its latest report that after news broke that Meta might become a new Neocloud, the market's initial reaction was to sell off computing power cloud companies like CoreWeave and Nebius, and to renewed concerns about "AI computing power oversupply." However, the firm's assessment is the opposite: this concern may be wrong. Meta's data center and computing power procurement will not slow down, but rather continue to accelerate. The article mentions that in the first half of this year alone, Meta has already signed contracts for over 5GW of capacity in the cloud services and managed data center sectors, and this does not include its accelerating self-built projects. SemiAnalysis states that Meta is in final negotiations with Anthropic, hoping to obtain access to Claude private instances. If this comes to fruition, the significance goes beyond "Meta buying more computing power"; it suggests that Meta may be building its own AI model service platform. This model is somewhat similar to AWS's Bedrock, Microsoft's Foundry, and Google's Vertex. Meta can initially use Claude internally, or it can package the model capabilities as a token-as-a-service to provide services externally in the future. In the short term, it might use its own models externally and Anthropic models internally; in the long term, Meta might even incorporate Anthropic and OpenAI models into its external service system. SemiAnalysis states that the underlying logic is that Meta has computing power, advertisers, social network distribution capabilities, and consumer-end entry points. If it can combine cutting-edge models, intelligent agents, and sales and marketing SaaS, it will not just be a company that buys GPUs, but will be moving towards the upper layers of AI application and model distribution.