Back to the AI frontline? Meta surges 15% in a single week
```
Meta's stock recorded its strongest weekly performance since February 2024 this week, driven by a series of AI strategy advances. Market confidence in its AI capabilities and commercialization path is rising significantly.
On Friday, Meta's share price surged 6% in a single day, with a total weekly gain of 14.8%, marking its best weekly performance since at least February 2024 and turning its year-to-date growth positive to about 1.4%.
The last time Meta posted comparable strong weekly performance was in February 2024, when investors responded positively to early results from the company’s “Year of Efficiency” cost-cutting plan, which aimed to recover its reputation tarnished by heavy bets on the metaverse through financial discipline.
This stock price breakout signals that Meta may be gradually shedding its market label as an “AI laggard,” opening up space for further expansion of its AI strategy.
As mentioned by Wallstreetcn, on July 9, Meta launched its flagship model Muse Spark 1.1, which has already outperformed Google’s Gemini model in several tests, including agent capability, programming, and multimodality.
Meanwhile, according to Reuters, Meta is advancing plans for in-house chip mass production and is rapidly expanding its computing power infrastructure. Deutsche Bank analyst Benjamin Black has raised Meta’s potential incremental third-party cloud services revenue estimate from $17 billion to $24 billion accordingly.
Research firm SemiAnalysis released a report predicting that Meta’s Meta Superintelligence (MSL) could surpass Google in frontier AI capability rankings within the next six months, shifting the AI competitive landscape from Google and OpenAI’s duopoly to a tripartite balance among Meta, OpenAI, and Anthropic.
Muse Spark 1.1’s low pricing targets a price war
Muse Spark 1.1, launched by Meta this week, is its first commercial model with near state-of-the-art agent programming capability, complete with a paid API interface.
CEO Mark Zuckerberg posted on X Thursday, emphasizing that the model is priced “very cheaply,” sparking speculation that Meta intends to proactively initiate an AI inference price war, pressuring competitors.
Meta Model API offers each account $20 in free credits and charges on a pay-as-you-go basis—input is priced at $1.25 per million tokens, and output at $4.25 per million tokens.
Richard Windsor, founder of Radio Free Mobile, noted in a Friday research note that the release of Muse Spark 1.1 confirms recent reports about Meta’s plans to launch a new computing power sales business. Windsor wrote:
There’s growing evidence that, given the attractive rate of return, Meta will launch a new business selling computing power to third parties.
He further pointed out, Muse Spark’s AI programming capabilities are close to top-tier models—“but priced at only 25% of their cost,” making it highly attractive for the mass market.
In-house chip R&D and compute scaling significantly boosts cloud revenue potential
As cited by Wallstreetcn, Meta plans to start mass production of its internally-designed AI chip, codenamed “Iris,” in September this year. The chip is co-designed with Broadcom and manufactured by TSMC, completed testing in just six weeks, and has already secured multi-year supply agreements with Samsung, SanDisk, and Sumitomo Electric.
In terms of computing scale, Meta plans to deploy 7 GW of compute this year, doubling to 14 GW by 2027.
To support these goals, Meta is building five gigawatt-scale “titan” hyperscale data center clusters and has developed its own “AI-Backbone” network architecture, which enables Meta to asynchronously scale complex training tasks across geographic distances of thousands of kilometers.
Deutsche Bank analyst Benjamin Black stated in a Thursday research note that these plans to scale computing power could mean about $24 billion of potential incremental third-party cloud service revenue for Meta, well above the previous estimate of $17 billion.
He also noted that Meta’s in-house chip development could open a practical path to cost reduction and efficiency improvement for the company.
SemiAnalysis: Meta AI could surpass Google within six months
According to Wallstreetcn, research firm SemiAnalysis believes that after a year of aggressive capital investment and restructuring, MSL is expected to surpass Google in frontier AI capability rankings within the next six months.
The report notes that the current duopoly of Google and OpenAI will be reshaped into a tripartite balance among Meta, OpenAI, and Anthropic.
The core judgment of the SemiAnalysis report is the speed of compute expansion. Meta’s trajectory in scaling AI compute is expected to surpass the combined compute power of OpenAI and Anthropic by the end of the year.
Reuters, citing an internal memo, reported that Meta’s capital expenditure cap for AI infrastructure is as high as $145 billion this year.
On the talent front, Meta has reassigned 3,000 engineers to its internal reinforcement learning environment factory to build proprietary data that commercial data suppliers cannot easily replicate, and has invested $14.3 billion in Scale AI, thereby bringing in top research talent from OpenAI, Anthropic, and other organizations on a large scale.
SemiAnalysis believes that assessing MSL solely by current benchmark performance is “missing the forest for the trees.” The real key is future growth potential rather than just the initial starting point.
The report points out that if Zuckerberg maintains the current pace of capital investment, Google may be permanently ousted from the top tier of global hyperscale AI players.
Risk Warning and DisclaimerThe market has risks, and investment requires caution. This article does not constitute individual investment advice, nor does it take into account the particular investment objectives, financial situation, or needs of any specific user. Users should consider whether any opinions, views, or conclusions in this article are appropriate to their individual situation. Investing based on this information is at your own risk. ```