By 
Meta stock (NASDAQ: META) jumped sharply after the company released Muse Spark 1.3, giving investors evidence that its AI spending may eventually produce returns.
The model is designed for coding and longer-running agentic tasks. Meta says it uses roughly 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2 in comparable engineering work.
Bank of America analyst Justin Post remains bullish. He has a Buy rating and an $810 price target, implying about 32% upside from Meta’s September 3 close of $610.68.
The debate is whether those advances can justify the infrastructure bill behind them.
Muse Spark gives investors visible AI progress
Meta has spent much of 2026 facing questions over whether AI investment is outrunning commercially useful progress.
Muse Spark 1.3 gives bulls something more concrete.
Meta says the model handles longer-horizon tasks better, manages multiple workflows in one thread and improves coding efficiency. It is available through Muse Code and the Meta Model API.
Bernstein reiterated an Outperform rating and an $800 target, arguing that Meta’s AI-enhanced advertising engine remains a major advantage. The firm believes Meta is on track to rival or surpass Google Search in advertising revenue.
Meta does not need Muse Spark to become a standalone business on OpenAI’s scale.
Better models can improve recommendations, ad targeting, engagement and developer tools across Meta’s apps. The payoff can appear inside its existing profit engine rather than only through model sales.
BofA sees more AI upside than the valuation reflects
Bank of America’s case rests on the gap between Meta’s execution and valuation.
Post highlighted Meta’s rapid model-release cadence and said the agentic improvements matter as the company develops a consumer AI agent internally known as Hatch.
There is also a hardware angle.
BofA estimates Meta’s planned MTIA custom-chip deployments could eventually represent 15% to 20% of its total AI capacity. Greater reliance on in-house silicon could lower computing costs as workloads expand.
At around $617 when BofA made its case, Meta traded near 18 times projected 2027 GAAP earnings, below its historical multiple of roughly 21 times and the broader market.
BofA’s $810 target is based on 24 times projected 2027 earnings.
The $810 case still has an expensive hurdle
The risk is that Meta’s AI bill remains enormous.
Its infrastructure push is raising fixed costs and pressuring margins and free cash flow, meaning new model releases must eventually translate into measurable economic returns.
KeyBanc remains constructive but more conservative, as the firm cut its target to $760 from $855 while keeping an Overweight rating.
KeyBanc said Meta Superintelligence Labs had made “meaningful progress” with Muse Spark and argued investors were “under-appreciating platform stickiness” among consumers and advertisers.
But the burden of proof keeps rising.
Investors will want evidence that Muse Spark gains adoption, that agents such as Hatch become useful products, and that MTIA chips reduce computing costs.
The post Meta stock jumps 4% after Muse Spark 1.3: why BofA still sees 32% upside appeared first on Invezz