At that time, we introduced the seven sisters of the US stock market, five of which were below the 200-day moving average. Looking back now, they have all recovered, with $Meta being the one with the most significant increase recently!🧐

Recently, the market has been extremely enthusiastic about Meta's Muse agent, which can be seen as the ultimate AI application. Today, let's discuss the logic behind it; fundamentally, it is a hardcore hardware computing power account and a commercial monetization loop.

1️⃣ First of all, how well is #Muse performing?

Muse has been launched for 13 days, with cumulative downloads surpassing 2.8 million within 12 days, and last Friday it topped the US iOS free app charts, surpassing ChatGPT, Claude, and Grok!

However, there is one issue: based on user test data, most users are not high-frequency "hot machines"; the actual daily active virtual machines are estimated to be around 1.5 million.

But how will this evolve in the future? We envision three scenarios:

Expanding from the US market to globally: Approximately 50 million users after one year.

WhatsApp channel fully ramping up: 25 million in six months and reaching 100 million in one year.

Combining WhatsApp + Instagram with aggressive promotion: Skyrocketing to 250 million.

Let's calculate the hardware account based on the middle scenario of 100 million users.

2️⃣ An extremely exaggerated "hardware account"

Muse relies on server CPUs to support the operation of virtual machines. Taking AMD's EPYC chip as an example, a single chip has 128 cores. If each user is assigned 2 vCPUs without sharing to ensure privacy, one chip can only accommodate about 60 VMs.

If pursuing 1:1 dedicated and 24/7 online:

For 1 million users: Requires tens of thousands of chips and 14 PB of memory.

For 25 million users: Requires 400,000 chips and 200 PB of memory.

For 100 million users: Requires 1.6 million chips, 800 PB of memory, and about 1.6 GW of power.

For 250 million users: Requires 4 million chips and 2 EB of memory.

For 1 billion users: Requires 16 million chips and 8 EB of memory.

This data 📊 is extremely exaggerated; what does that imply? This is undoubtedly a computing power behemoth. AMD's server chip shipments this year total only around 10 million, with the entire industry producing about 39 million chips. Just configuring VMs for 100 million Muse dedicated users would consume a substantial portion of AMD's annual production capacity! No wonder AMD's capacity ran out early, and the next generation of chips must continuously increase the number of cores.

Memory and power crisis. CPU computing power can be time-shared, but user files and resident data need to be retained; hence, DRAM must be firmly locked in. The entire DRAM capacity for 2026 has already been fully booked. 100 million users will directly lock in 1% of global DRAM bit capacity for a year; if it's 1 billion users, that's 10%.

Capital expenditure pressure. Meta's own power planning is 7 GW this year and 14 GW next year, which also includes AI training and advertising business. Just to handle the hardware for 100 million VMs, building the supply chain would require an investment of $30 to $50 billion. Even with a subscription fee of $20 per month, 10 million paying users would only generate $2.4 billion in revenue per year, which doesn’t suffice against Meta's annual capital expenditure of $130–145 billion; relying solely on subscription fees simply cannot cover the costs.

3️⃣ Since it's so expensive, what is Meta aiming for?

Relying solely on collecting "monthly fees" from C-end users would certainly lead to losses. Muse's real trump card lies in connecting the entire network marketing data, transforming into a corporate-level "full-time digital marketing officer," as shown in the video below.

The current Muse can directly connect to:

Advertising platforms: Google Ads, Meta Ads, and over 12 mainstream advertising channels, even including ChatGPT Ads.

SEO and data tools: GSC, GA4, Semrush, Ahrefs.

E-commerce and content ecosystems: Shopify, WordPress, Webflow, HubSpot, PostHog.

This means Muse can monitor your campaigns 24/7, answer your advertising return on investment, analyze SEO traffic, monitor anomalies on Shopify, and even actively make adjustments and optimizations in your website and ad accounts.

When Muse launches personal shopping, appointment scheduling, and calendar management, as well as B-end commercial agency services, the entire massive application ecosystem (FB + IG + WhatsApp) instantly transforms into a super launchpad for AI agents.

Currently, over 1 million businesses are using Muse's commercial agents weekly. Morgan Stanley's latest research report keenly points out that this represents the second growth curve beyond Meta's advertising, and the target price has been directly raised to $820.🧐

4️⃣ Pulling one thread affects the whole body; the entire AI hardware sector is reactivating.

The story of Muse is quickly propagating into the AI hardware sector. Market capital is starting to shift the logic of capital expenditure from purely "big model training demand" to specific "agent hardware demand." That's why even a slight stir of Muse in its early stages could stimulate substantial increases in the stock prices of $AMD and $Intel.

At the same time, the garden-wall war between platforms has begun. For example, Amazon has started to block Muse's scraping and access on its site. The struggle for data entry and exit between giants has only just begun.

In summary, from the perspective of capital markets, in the past, people were speculating on AI, focusing on "big tech companies with dreams"; now, the market has started to favor "giants that can turn AI dreams into real gold and silver on their super platforms."

From the deadlock in the hardware supply chain to the vast transformation foundation at the application end, and to the real paid loop at the B-end, this certainty of "dreams landing into profits" has formed a strong capital pursuit surge for Meta's stock price recently!🧐

For US stock speculation, I choose to use the #RWA US stock tokenization platform #MSX; let's invest together in the US stock market: http://msx.com/?code=Vu2v44

Early fans and partners of US stock investments can DM me. After filling out the form, you can enter the US stock exchange and discussion community for free (currently limited to 10 people per week, assistant review may take some time, thank you🙏)!

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。