AI trades moved back to the center of the U.S. stock market on Sept. 21. The Nasdaq Composite rose 2.26% to 27,122.09, setting a record closing high, while the S&P 500 gained 1.49% to 7,764.70.
Among individual names, Meta climbed about 11.4%. AMD rose about 10%, pushing its market capitalization above $1 trillion for the first time. Intel added about 12%, and Arm advanced about 17%. The broad strength in semiconductor stocks came as Muse, Meta’s personal AI agent, rapidly moved to the top of U.S. app store rankings.
Muse reached No. 1 on both major U.S. mobile app stores in less than two weeks
Muse was not released on Sept. 21. Meta launched the product on Sept. 8 and described it as a personal AI agent that can carry out tasks on a user’s behalf.
Unlike chat tools built mainly to answer questions, Meta said Muse can use a browser, fill out forms, handle email and book travel. It can also continue some tasks after a user closes the app. Sensitive actions such as sending emails or making payments still require user confirmation.
Less than two weeks after launch, Muse had climbed to the top of the free app charts on Apple’s U.S. App Store and Google Play. App data firm Appfigures estimated that the app had reached about 1.1 million cumulative installs as of Sept. 18. The report noted that the figure was not officially disclosed by Meta.
Meta said Muse is available through a standalone app and through WhatsApp. The product is centered on free features, with a subscription option for higher usage.
Investors are reassessing how AI applications could drive compute demand
The rise of Muse has prompted the market to revisit the demand profile for AI infrastructure. If personal agents keep browsing the web, calling tools and completing multi-step tasks for users, compute demand may no longer be driven mainly by model training. A larger share could come from inference and task orchestration in everyday use.
The gains in AMD, Intel and Arm reflected that line of thinking. A personal AI agent does more than generate a single reply after a prompt. To complete a goal, it may need to search, compare, fill in information, call apps and wait for confirmation across several steps. If users adopt that behavior at high frequency, the compute load per task could exceed that of a standard chat interaction.
AMD and Meta already have a direct business relationship
AMD and Meta are not linked only by market narrative. In February, the two companies said they would expand a long-running partnership, with plans covering up to 6 gigawatts of AMD Instinct GPUs and deeper cooperation around EPYC server CPUs.
AMD said products supporting the first 1-gigawatt deployment are expected to begin shipping in the second half of 2026.
On the operating side, AMD reported $6.7 billion in data center revenue for the second quarter, up 107% from a year earlier. The company said the increase was driven by demand for both EPYC processors and Instinct GPUs.
Several links still need to be tested before app momentum turns into chip revenue
The significance of Muse is not only that Meta has launched another AI app. The market is looking at the possibility that personal AI agents could move into mainstream use. Even so, there are still several steps between app popularity and chip revenue.
The first is retention. Reaching No. 1 on the free charts shows users are willing to try the product, but it does not show whether they will still use it every day a few weeks later. Meta has a large distribution network across its apps, which can amplify early downloads quickly. A stronger signal of product value would be whether users keep handing real tasks such as shopping, scheduling and email to Muse.
The second is monetization. Muse is currently built around free features, and Meta plans to offer higher usage through subscriptions. Whether user growth turns into revenue will depend on paid conversion and usage intensity per user. If task volume rises quickly but the cost of running each task is also higher, a larger user base may not improve profit right away.
The last step is chip demand itself. AI agents need GPUs for model inference, and they also need CPUs for system operations and task scheduling. Actual procurement volumes, though, will depend on how Meta optimizes models, allocates workloads and uses existing equipment. Muse’s surge in popularity and AMD’s stock move happened in the same market run, but that does not mean Muse has already created a confirmed AMD order of any specific size.
Record highs in the Nasdaq reflected both the AI theme and the broader market backdrop
AMD’s investment case in this move had both a current foundation and a forward-looking element. The current foundation is visible in its 107% year-over-year growth in data center revenue last quarter and in its existing GPU, CPU and system-level cooperation with Meta. The forward-looking element is the possibility that if Muse and similar AI agents gain durable, high-frequency usage, Meta may need to deploy more compute capacity.
The first part is already visible in earnings and partnership disclosures. The second still needs to be confirmed through shipments, revenue and usage data.
The Nasdaq’s record close also showed that the market is again willing to price in wider adoption of AI applications and the upstream demand for compute. The report added that lower oil prices and lower U.S. Treasury yields on the same day also created a supportive backdrop for technology stocks. Framing the entire rally as a case of one app lifting every chip stock would overstate the immediate market impact of a single product.
The next points for investors to watch are not limited to whether Muse can stay at the top of download rankings. The market is also looking for any disclosure from Meta on active users and paid conversion, whether the cost of running AI agents can come down, and whether the existing Meta-AMD partnership can translate into product shipments and data center revenue on schedule.
About MSX
The report described MSX as an RWA trading platform focused on access to global financial markets. It said the platform is among the earlier operators in on-chain U.S. stock trading and offers spot and derivatives trading tied to nearly 400 tokenized stocks and pre-IPO assets.
The article also said MSX has built a broader digital finance offering around U.S. equities spot and perpetuals, crypto-to-crypto trading, pre-IPO products and research services. It ended with a risk reminder that macroeconomic conditions and the U.S. stock market can be highly volatile, and that the material is for research observation only and does not constitute investment advice.