Meta's AI bets are boosting ad performance - and the bill is rising
Meta posted 26% year-over-year revenue growth to $51.24B in Q3, with AI carrying much of the load. The company's end-to-end AI ad stack - including Advantage+ - now runs at $60B+ in annualized revenue. Average price-per-ad rose 10% YoY, and Instagram Reels pushed past a $50B+ annual run rate.
Advertiser adoption keeps climbing. The number of advertisers using at least one of Meta's video-generation tools jumped 20% over Q2, signaling real trust in creative automation and performance gains.
What's actually working
- AI-driven ranking and personalization: Keeps people in feed longer and improves ad outcomes.
- Advantage+ automation: Smarter targeting and creative testing at scale for less manual effort.
- Short-form video momentum: Reels is now central to Instagram and monetizing better than expected.
- Sharper execution: "Meta has quietly turned AI into margin," said Jeremy Goldman of eMarketer. The tools are getting used - and they're producing.
The bill for scale
Staying competitive in AI is expensive. Meta lifted full-year capex to $70B-$72B to fund specialized talent and energy-heavy data centers. Shares fell on the spending update, and about 600 roles were cut in the AI division this month.
Reality Labs remains a drag, posting a $4.4B loss in Q3. That said, if AI keeps raising ad performance, those headwinds are easier to carry.
Competitive context
Amazon and other giants are building similar AI-supported ad engines. Meanwhile, TikTok continues to face ownership uncertainty in the U.S., and Meta has leaned into recommended creator content as user behavior shifts.
For marketers, that points to a simple truth: creative variety and signal quality will decide who wins the auction, not manual hacks.
What marketers should do now
- Test Advantage+ with guardrails: Run structured A/Bs against your current stack. Keep budgets, events, and attribution consistent to isolate lift. See Meta's overview of Advantage+ here: Advantage+.
- Feed the algorithm: Improve first-party data quality, connect conversions, and maintain clean event mapping. Strong signals beat manual tweaks.
- Go heavy on short-form video: Build multiple hooks, quick cuts, and creator-led variants. Aim for 6-12 iterative creatives per ad set to let the system find winners.
- Scale creative automation: Use video generators and dynamic formats for rapid iteration. Monitor fatigue weekly; refresh sooner, not later.
- Budget for higher CPMs: With price-per-ad up ~10% YoY, push for better ROAS through creative and conversion-rate gains, not pure media cuts.
- Measure incrementality: Layer geo holdouts or conversion lift where possible. Stop over-attributing to last click.
- Prepare for platform shifts: Keep a backup plan if TikTok's U.S. situation changes. Reels, Shorts, and creators give you options.
- Upskill your team: Treat AI-native media buying and creative ops as core skills, not side projects.
By the numbers
- $51.24B Q3 revenue, up 26% YoY
- $60B+ annual run rate for AI-powered ad offerings (including Advantage+)
- +20% QoQ in advertisers using at least one Meta video generator
- 10% YoY increase in average price-per-ad
- $50B+ annual run rate for Instagram Reels
- $70B-$72B updated 2025 capex outlook
- $4.4B Q3 loss at Reality Labs
Why this matters
The center of gravity is clear: automation sets the strategy, creative sets the ceiling. As Meta "industrializes" relevance, the brands that win will ship more variations, clean up their data, and let the system sort performance - while keeping tight measurement discipline.
If you want structured ways to level up AI-driven marketing, explore practical training built for marketers: AI Certification for Marketing Specialists and Courses by Job.
For the full context on Meta's results and guidance, see its investor page: Meta Investor Relations.
Your membership also unlocks:
