Meta stock rallies 15% in best week since 2024 as AI strategy wins over skeptical investors

Meta Stock Posts Best Week Since Early 2024 on AI Revenue Plans

Meta stock jumped approximately 6% on Friday and gained nearly 15% for the week, marking its strongest weekly performance since early 2024 and erasing the company's year-to-date losses 1

1

5

5

. The $1.7 trillion market-cap company had been flat on the year while the tech-heavy Nasdaq-100 climbed 18%, making it one of Big Tech's notable laggards 4

4

. The dramatic reversal in investor enthusiasm for AI came after Meta detailed concrete plans to monetize AI computing capacity and unveiled new AI model development initiatives that signal CEO Mark Zuckerberg's commitment to competing directly with OpenAI, Anthropic, and Google.

Meta Compute Drives Investor Enthusiasm for AI Infrastructure Monetization

The primary catalyst behind the rally is Meta Compute, the company's plan to sell AI computing capacity and models to external customers, directly addressing Wall Street's concerns about massive capital spending with no clear return path 2

2

. According to an internal memo reported by Reuters, Meta plans to double its cloud computing capacity to 14 gigawatts next year 4

4

. Wolfe Research estimates that for every gigawatt Meta monetizes at roughly a $25bn rate, earnings per share could rise around 20% 2

2

. This represents a fundamental shift in Meta's AI strategy, transforming what had been viewed as a colossal cost center into a potential revenue stream that could diversify the company beyond its advertising-dependent business model.

Options Traders Signal Strong Bullish Sentiment

Options traders piled into Meta on Friday, with volume running at more than three times the 30-day average and 78% of the stock's $1.8 billion in options premium tied to calls 1

1

2

2

. More than twice as many calls were bought compared to puts, with eight of the top 10 contracts by volume being calls as of midday 1

1

. Analyst targets have clustered in the low-to-mid $800s over twelve months, with the published scenario range running from about $720 at the bearish end to roughly $869 at the bullish one 2

2

. Bank of America analyst Justin Post argued that Wall Street is undervaluing Meta's AI infrastructure, estimating it's worth $12 billion per gigawatt with potential for significant upside considering the specialized AI capacity Meta is building 4

4

.

New AI Models Signal Aggressive Push Into Competitive Market

Three months after introducing Muse Spark, Meta made two significant AI model development announcements this week 3

3

. On Tuesday, Meta released Muse Image, a new AI model for creating images designed to attract creators and advertisers to its offerings 3

3

. On Thursday, the company launched Muse Spark 1.1, an AI coding product aimed at running agentic and coding workloads that will compete directly with Anthropic and OpenAI 1

1

. Developers can now pay to use Muse Spark 1.1 through a new API platform 4

4

. These releases underscore progress at Meta Superintelligence Labs, which is being led by Alexandr Wang, and demonstrate Meta's aggressive efforts to make a splash in AI models despite competitors having significant head starts 3

3

.

Reducing Dependence on Nvidia Through Custom Chip Development

Meta is pushing its own MTIA AI chips into production to cut its dependence on Nvidia, with plans to begin producing chips with designer Broadcom in September 2

2

4

4

. This vertical integration strategy aims to give Meta more control over its AI infrastructure costs and capabilities, potentially improving margins on both internal AI operations and external cloud services. The move aligns with broader industry trends as tech giants seek to reduce reliance on third-party chip suppliers and optimize hardware specifically for their AI workloads.

Skepticism Remains About Competing With Established Hyperscalers

Despite optimism around AI strategy, Meta Compute faces significant challenges. The company has not sold anything yet and has never run a cloud business for external customers, while AWS, Azure, and Google Cloud have a decade of head start on operations, sales, and trust 2

2

. These established hyperscalers dominate the market Meta is attempting to enter. There's also an uncomfortable interpretation of renting unused computing capacity: it could signal shrewd monetization or suggest the company bought more compute than it can use 2

2

. What Wall Street bought this week was a narrative, and whether Meta can actually sell compute against three entrenched competitors remains an open question. Investors had been anxious for months about the scale of Meta's AI investments with no visible route to a return, and while the company has now offered a plan, execution will determine whether the rally was justified.