Muse ignites tech stock rally! The Philadelphia Semiconductor Index has risen five days in a row! AMD joins the "trillion dollar club"! (In-depth summary)

Muse tops the App Store → AI agent consumer demand validated → Demand for computing power spreads from GPU to CPU → AMD, Intel, Arm stocks soar → Philadelphia Semiconductor Index up for five consecutive days

I. Overview of Market Performance

On Monday (September 21), U.S. semiconductor stocks rallied collectively, creating a technology stock rally centered on “AI agent computing demand.”

Core Stock Performance:

Supporting Factors: Besides optimism regarding AI chip demand, falling oil prices, declining U.S. Treasury yields, and expectations for progress in ending the Iran war also provided additional support for the stock market.

II. Core Catalyst: Muse Tops the App Store

2.1 Fundamental Product Positioning Upgrade

The immediate catalyst for this round of the rally comes from Muse, the personal AI agent launched by Meta earlier this month. Unlike traditional chatbots primarily responsible for generating text or answering questions, Muse is positioned as a personal AI agent that candirectly execute tasks on behalf of users, assisting users with online shopping, buying movie tickets, scheduling services, and other real-life tasks.

This difference in product positioning is the key to understanding the current rally: traditional chatbots are "Q&A tools," while Muse is a "task-performing agent." The former consumes computational power per inference, while the latter requires continuous operation, task planning, multi-tool invocation, and background execution, resulting in a completely different demand structure for computing infrastructure.

2.2 Early User Data

Muse quickly climbed to the top of the free app chart on Apple's App Store, indicating strong initial consumer demand. This performance has reignited market imagination around the mainstream adoption of AI agents.

III. Structural Change in Computing Demand Logic: Spreading from GPU to CPU

3.1 Limitations of the Traditional AI Investment Logic

The previous wave of generative AI investment largely revolved aroundGPU suppliers such as Nvidia, concentrating on the massive parallel computing power required for model training.

3.2 The Increment Brought by AI Agents

With Muse’s consumer popularity, the market is focusing more on theother type of computing power demandpotentially brought by AI agents. Compared with traditional chatbots, AI agents not only need to generate answers, but also:

• Understand user goals

• Plan tasks

• Invoke various applications and tools

• Continuously execute a series of operations

This means that if AI agents are ultimately adopted at large scale, theworkload for inference, task orchestration, and server infrastructurecould all increase accordingly.

3.3 Wall Street’s Core Judgment

Jefferies analyst Jacky He stated that, as AI agents are adopted more broadly by consumers, higher AI inference, task orchestration, and infrastructure workloads are expected to boostserver CPU demand.

Wedbush analyst Matthew Bryson pointed out that AI agents depend on a large number of compute-driven applications and, in this market, the major compute suppliers includeIntel and AMD. This explains why AMD and Intel stood out among chip companies in this round of AI-driven gains.

Core Logical Chain:

Large-scale popularization of AI agents → continuous inference and task orchestration workloads → server CPU demand growth → AMD, Intel, and Arm benefit

IV. AMD’s Milestone: Joining the “Trillion Dollar Club”

4.1 Stock Price Reversal

AMD surged nearly 10% on Monday, with its market cap breaking the $1 trillion threshold for the first time. Previously, concerns about the sustainability of the AI investment boom and the returns on massive AI capital expenditures had weighed on AMD’s share price. From the June high to the July low, the stock fell nearly 26%. However, recent weeks have seen a clear reversal in market sentiment. As of Monday, AMD’s cumulative gain in September is about 30%, on track for its best single-month performance since May.

4.2 Meta and AMD in the Supply Chain

Data show thatMeta is AMD’s second largest customer, contributing about 5.5% of AMD’s revenue. Therefore, if Meta continues to expand investment in AI agent-related infrastructure, AMD could become a direct beneficiary. This supply chain relationship is the key to understanding why AMD’s stock price is so sensitive to Muse-related news.

4.3 Expansion of the “Trillion Dollar Club”

With this market cap breakthrough, AMD has officially joined the global chip industry’s “trillion-dollar club.” At present, several semiconductor companies, includingNvidia, Micron Technology, Broadcom, and TSMC, have all reached or surpassed the $1 trillion mark in market cap. Besides the above chip firms, there are now 15 other listed companies worldwide with market caps of at least $1 trillion.

V. Resistance and Risk Signals

5.1 Platform Blocking: Amazon Blocks Muse

The rapid expansion of Muse has already encountered resistance from other internet platforms.An Amazon spokesperson said the company had already blocked the Muse AI agent from accessing its retail website as of Sunday night. This incident reveals a fundamental contradiction in the AI agent business model: the agent executes tasks for the user, potentially bypassing the platform’s advertising and recommendation systems to complete transactions directly, threatening the platform’s own business interests. If more platforms follow Amazon in blocking access, Muse’s functionality and user experience may be materially affected.

5.2 Uncertainty in Computing Demand Transmission

While the logic from GPU to CPU is self-consistent, there are two checkpoints that require verification:

• Can AI agents develop from early hype into large-scale daily applications: The current download boom of Muse partly benefits from the traffic effect of the Meta ecosystem. Its organic user acquisition and retention still need to be observed.

• Can CPU suppliers achieve substantive revenue growth: Whether growth in inference and orchestration workload can convert into incremental orders for AMD and Intel depends on the capex structure and procurement decisions of companies like Meta.

VI. Overall Conclusion

The essence of this round of market moves is the market's re-pricing of the next phase in AI industry computing demand structure. In past years, AI investments mainly revolved around large model training and GPUs. However, as AI shifts from “answering questions” to “executing tasks for users,”the importance of inference, task orchestration, and server CPUs may further increase.

Core Transmission Chain:

Muse tops the App Store → AI agent consumer demand validated → Computing power demand spreads from GPU to CPU → AMD, Intel, Arm stocks soar → Philadelphia Semiconductor Index up for five consecutive days

Key Judgments:

• Short term: The early success of Muse provides the first consumer product validation for the AI agent narrative, with market sentiment notably boosted, and AMD crossing the trillion-dollar mark as a landmark event.

• Medium term: If AI agents such as Muse further develop from early consumer hype to large-scale daily usage, the primary driver of AI computing demand might expand from model training to continuously running agent workloads. Whether CPU suppliers like AMD and Intel can achieve substantive revenue growth from this will be a key market focus in the next phase.

• Risks: The incident of Amazon blocking Muse’s access to its retail site suggests that conflicts of interest between AI agents and existing internet platforms could become a major hurdle for scale adoption; meanwhile, the transmission of computing demand from GPU to CPU still needs to be verified with real order data.

Most critical observation points: Product roadmap disclosures at Meta Connect (September 23-24), Muse user retention and paid conversion data, and changes in CPU-related procurement in the capex guidance of hyperscale cloud vendors.