Meta Shares Climb 3.4% After Muse Wonder Debut; Can Subscribers Fund AI?

On Monday, Muse Wonder made its debut, and Meta's shares (META) were 3.4% higher — the model is the company's first top-tier frontier AI offering. That gain is roughly a third of the 11% one-day burst that Muse generated in September.

In Meta's view, its benchmarks put Wonder and the bigger Muse Wonder Full on a par with the top systems from OpenAI and Anthropic. The more muted response points to investors having already priced in a large part of the rebound during September's rally.

Why Did Shares Respond More Calmly to Wonder Than to Muse?

Meta gained ground even as S&P 500 technology stocks slid 0.9%, according to Bloomberg. Still, that gain was smaller than the 11% one-day jump that Muse, Meta's personal AI agent, produced in September. For the month as a whole, Meta shares added nearly 27%.

TechCrunch reported that the benchmark figures come from Meta's own testing. Muse Wonder Full, the more powerful version, is available only to subscribers of Muse Code, Meta's coding-focused AI tool.

Only the smaller Wonder is on offer to developers, at $4 for every million input tokens, the text units models read. Anthropic's Opus 5.5 is priced in that same vicinity, so the move reads as a subscription-focused play, not a bid to attract developers.

Can Subscriptions Justify Wonder's Spending?

Wonder will soon run the Muse agent, at first for paying subscribers. Jason Helfstein of Oppenheimer figures 115 million subscribers at $20 a month could generate roughly $28 billion in annual revenue.

Wells Fargo's Ken Gawrelski lifted his Meta price target to $1,000. He still believes Meta's commentary will keep near-term revenue expectations in check and calls fiscal 2027 the low point for earnings.

Distribution concerns also hang over the stock. Last month, Amazon barred Muse from its platform, saying the tool had browsed the web without disclosing its identity. Meta's stock trades near $742, nearly 5% below the September high of $779.82.

Strong frontier benchmarks may just be the entry ticket for big tech companies. The next repricing could depend on paid users more than on model scores.