In the latest Earnings Reports released on Wednesday, Meat significantly raised its capital expenditure budget for this year, continuing its large-scale bet on AI.
However, in reality, $Meta Platforms (META.US)$ it is facing numerous difficulties in the AI field: delays in the release of AI technology, doubts about its "open-source" strategy, and the critical Llama 4 Behemoth model has yet to be launched... Investors are eager to know, what is the future of Meta?
The LlamaCon conference was "loud but lacking substance": Developers were disappointed, and Meta's role as a follower is hard to change.
Meta held its first LlamaCon AI developer conference on Tuesday this week, launching a series of new products, including a standalone Smart Phone app for the Meta AI assistant and a preview of the Llama large language model API for developers.
However, at the conference, Meta failed to release the highly anticipated inference version model Llama 4 Behemoth, described as "the most powerful mixed-expert AI model trained on 2 trillion parameters", which was originally scheduled for release weeks ago but has been delayed multiple times.
Brownstone Research released a report stating that Meta "did not present enough substantive information" at the conference, clearly indicating its lagging position in the AI field.
The bank emphasized that the highly anticipated Llama 4 Behemoth model failed to be released as scheduled, and the focus of this release seems more like an attempt to engage both Consumers and developers, but no breakthrough has been achieved in either area.
Meta's conference was a complete failure. This sentiment is justified.
In contrast, competitors like OpenAI, Anthropic, Google, xAI, and Mistral have already launched Consumer-grade chatbots and enterprise API interfaces, while Meta has clearly been left behind.
Meta's Global Strategy of 'open source' is being questioned, with concerns about its lack of innovation.
At the same time, Meta's 'open source' strategy has been seriously questioned.
The authority on open source software ecosystems, the Open Source Initiative, made it clear that 'The license for Llama LLM is clearly not an 'open source' license,' and further demanded that Meta 'correct its misstatements.'
According to analysis by Brownstone Research, Meta's Software is not truly open source, as it does provide the model's weights and allows others to manipulate those weights, but Meta does not allow Llama to be used for any purposes that violate human rights, does not permit the development of competing products, and does not allow training models with over 0.7 billion parameters.
The bank pointed out that these restrictions are contrary to the true spirit of open source.
Meta's capital expenditures have significantly increased as it continues to burn cash in a "high-stakes" bet on AI.
Faced with intense competition and technological delays, Meta's choice is to continue pouring money in.
In the earnings report released this Wednesday, Meta announced a significant upward revision of its 2025 capital expenditure expectations, increasing the upper limit by 7 billion dollars compared to what was announced in January this year, in order to acquire computing power more quickly.
If Meta spends according to its new projected upper limit, this year's capital expenditure will be 84% higher than last year, a number that nearly approaches Google's expenditure level this year, even though Google is a much larger company than Meta in terms of revenue scale.
During the earnings call, Zuckerberg expressed strong confidence in future opportunities in the AI field and detailed how Meta has already been using AI to improve content recommendations and ad placements for its social media users.
$Meta Platforms (META.US)$ The ongoing investment in AI presents both opportunities and challenges for investors. On one hand, Meta's strong distribution channels (Facebook, Instagram, and WhatsApp) give it a unique advantage. On the other hand, it seems that Meta's innovative capacity in the field of AI is lacking and is falling behind its competitors.
With the delayed release of the Llama 4 Behemoth model and other technological setbacks, investors have reason to question Meta's strategy and execution capabilities in the field of AI. Whether Meta can stand out in the fierce competition and bring returns on its massive AI investments will be a focal point of close attention from investors in the future.
Editor/Rocky
