On October 12, Meta made one of the company’s riskiest bets. Zuckerberg introduced Muse Wonder, the first state-of-the-art AI model, with the desire to finally compete head-to-head in the league of OpenAI, Anthropic, and Google. Because Meta moved towards being a 100% AI company months ago and spent a fortune signing a team of galacticos led by Alexander Wang who haven’t shown much so far.
Now, and as Meta’s financial manager Wang himself commented, it’s all or nothing. If it’s good, “everything will be fine, but if it’s wrong…”.
Muse. As we said, Muse Wonder and Muse Wonder Full are the first frontier models from Meta Superintelligence Labs. This is the result of several months of work by the team led by the aforementioned Alexander Wang, who reached Meta after founding Scale AI and “signing” Zuk’s company for $14.3 billion.
They’re bigger and more capable than the previous model Muse Spark, and according to Meta’s internal tests, they perform at comparable levels to the Fable 5.1 and GPT-6 Astra. According to independent results, things changed slightly, scoring 56 points on the synthetic analysis index compared to 58 on the Claude Opus 5.5.
There is no middle. It’s clear that the two new Muses are powerful, but they don’t just have to be: they must become an AI tool on the level of major rivals, and it seems the company’s financial health depends on it. In a recent podcast, Wang noted that Susan Lee, Mater’s CFO and most responsible for the company’s finances, told him several times during a board retreat in Hawaii that Muse must live up to its promises.
In this kind of statement, laughing and relaxed, you never know what the tone was, but according to Wang, Susan was very emphatic and repeated a phrase: if it’s good, everything will be fine, but if it’s wrong, things will be very, very bad. It makes sense that Susan is aware of this because she’s been on the meta since 2008 and has experienced radical changes like Facebook’s transition to meta due to the metaverse and we already know how that turned out.
His job is to make sure the company invests where it expects the greatest return, and that sometimes means curbing Mark Zuckerberg’s desire for infinite growth at any cost. Wang, being the company’s latest star signing, has a free hand from Zook, but Susan must square Axel, and to do so, the new muse must obey.
Matching stories. In fact, Susan is not alone. An article published in Bloomberg a day after Muse’s presentation showed how Zuckerberg admitted that his company had to commit hundreds of billions of dollars to AI infrastructure in 2025 without knowing whether the resulting products would be successful.
This narrative matches what we’ve seen in recent months, with Big Tech frantically buying hardware without the ability to connect all that equipment as other necessary investments such as land and energy take time to prepare. They have essentially committed billions to physical assets like chips and data centers without knowing if the model will work.
Goal in AI. It remains to be seen whether all these investments change the image of a company that, a year ago, was seen as a laggard in the AI race. At this point, the change in narrative hasn’t transferred to the stock market, where its shares are up just 6% in twelve months, compared to 20% for Amazon or 45% for Google. Investors were wary after seeing second-quarter results, where revenue rose 28%, but earnings per share fell short of expectations and free cash flow fell 91% as a result of all the money committed to AI.
There is another difference between Amazon and Google that is important in this conversation. In AI, Meta currently relies on its models, but Google and Amazon have another strength: They sell the cloud, and both Google Cloud and Amazon Web Services have skyrocketed in recent months because they provide AI infrastructure for those who don’t have the muscle to build their own.
In a few days, Meta will release its third-quarter results, and that’s when we’ll see if investor confidence muse wonder. Right now, hundreds of billions thrown at Meta AI are circling the triple hoop, and on October 21st we’ll know if the ball lands in or out of the basket.
Picture | Cheeky Pint (edited)
In Xataka | When Meta bought Manus, Chinese authorities interpreted a message: a conspiracy against him
