TLDR
- META stock climbed ~15% this week after a series of AI-focused announcements
- Meta will launch a custom AI chip in September, built with Broadcom and TSMC
- CEO Mark Zuckerberg said Meta is exploring renting its AI computing power to third parties
- New Muse Spark 1.1 model undercuts Anthropic’s pricing by charging $1.25 per million input tokens vs $5
- Meta’s stock turned positive year-to-date on Friday after the announcements
Meta Platforms (META) had a strong week on Wall Street, with its stock climbing roughly 15% after a string of AI announcements that gave investors a clearer picture of how the company plans to monetize its massive spending on infrastructure.
The stock rose more than 5% on Friday alone, pushing it back into positive territory for the year. It had been weighed down by concerns over rising capital expenditure and questions about returns on AI investment.
The week’s biggest talking points? Custom chips, a new AI model, and a pricing strategy that takes direct aim at the competition.
Meta’s Custom Chip Plans
Meta confirmed it will launch a custom-designed AI chip in September, developed in partnership with Broadcom and Taiwan Semiconductor Manufacturing. The goal is to reduce dependence on Nvidia and AMD, both of which have commanded premium prices as demand for AI chips has outpaced supply.
The broader ambition is to grow Meta’s computing capacity from a projected 7 gigawatts in 2026 to 14 gigawatts by 2027. That’s a doubling of capacity in a single year. Meta also plans to upgrade these chips on a six-month development cycle going forward.
On Wednesday, the company announced it will build a new data center in Canada, its 33rd facility globally.
CEO Mark Zuckerberg told Bloomberg on Thursday that Meta is exploring renting out its AI computing capacity to third parties. This could mean hosting competitor AI models, or simply selling chip and server access like a cloud provider. Either way, it would open a new revenue stream outside of Meta’s core advertising business.
Muse Spark 1.1 Debuts With Aggressive Pricing
Meta’s Superintelligence Labs unveiled Muse Spark 1.1 on Thursday. The model is designed for agentic tasks and coding, can coordinate multi-agent systems, and is faster than previous versions. Meta says it’s better at diagnosing and fixing bugs than earlier releases.
But the pricing is the headline. Meta is charging developers $1.25 per million input tokens and $4.25 per million output tokens. Compare that to Anthropic’s Opus 4.8, which runs $5 per million input tokens and $25 per million output tokens.
That’s a sharp undercut, and it positions Muse Spark 1.1 as a cost-effective option for developers who don’t need the most powerful models on the market.
What the Numbers Look Like
META closed Friday at $669.21, up $37.73 or roughly 5.97% on the day. The 52-week range sits between $520.26 and $796.25, and the stock carries a market cap of approximately $1.7 trillion.
The week’s moves came as Meta outlined a path to turning its AI infrastructure from a cost center into a potential revenue source — whether through model pricing, cloud rental, or chip capacity.
Meta’s next earnings report will be the real test of whether these announcements translate into numbers investors can hold onto.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
