Any other app does not have moat - anyone can do the same app if it basically wrap the LLM.
If anything, LLM just destroy thier current moat, I.e. if everything is getting behind a chat interface, no one would would see ads.
On "replacing expensive humans" agree that's part of it, but the bigger play is augmenting existing products. Meta's Q3 2025 guidance shows ad revenue still growing 21.6% YoY. They're using AI to make existing ads more effective (better targeting, higher conversion), not replacing the ad model entirely.
On the moat question this is where the infrastructure spending makes sense. You're right that wrapping an LLM has no moat, but owning the infrastructure to train and serve your own models does. Meta has three advantages: (1) 3B+ daily users generating training data competitors can't access, (2) owning 2GW of infrastructure means $0 marginal cost for inference vs paying OpenAI/Anthropic, and (3) AI embedded in Instagram/WhatsApp/Facebook is stickier than standalone chat.
On ads behind chat interface this is the real risk. But Meta's bet seems to be: short-term AI improves existing ad products (already working), mid-term AI creates new surfaces for ads (AI-generated content, business tools), and long-term if chat wins, Meta wants to own the chat interface (Meta AI), not lose to ChatGPT.
The $75B question is whether they're building a moat or just burning cash on commodity infrastructure. Time will tell, but the data advantage plus vertical integration gives them a shot.
What's your take do you think the data moat is real, or can competitors train equally good models on synthetic/public data?
0xrelogic•3mo ago
Key findings: • CoreWeave: $14.2B (6+ years, Nvidia GB300 GPUs) • Oracle: ~$20B (multi-year cloud deal) • Blue Owl/Hyperion: $27B (joint venture, private credit financing) • Scale AI: $14.3B (49% stake)
What's interesting isn't just the size—it's the structure. Meta is using private credit and joint ventures instead of traditional CapEx. The Hyperion deal: Meta owns 20%, Blue Owl owns 80%, but Meta is on the hook for 16 years.
The math is uncomfortable: $75B in infrastructure spending vs ~$7B in AI revenue. That's a 10:1 cost-to-revenue ratio.
Happy to discuss the economics, the private credit angle, or whether this is sustainable.
almosthere•3mo ago
When they start advertising in GPT (subtle hints to eat cheetos for example, or for rich people to buy a Maserati) the entire ratio will invert. It will be one of the largest advertising wins in the world.
0xrelogic•3mo ago
Meta's Q3 2025 ad revenue is expected at 48.5B(21.666-72B, so still a significant cost-to-revenue gap.
Regarding advertising in AI assistants specifically OpenAI's Sam Altman actually addressed this in June 2025, calling ads in ChatGPT a "trust-destroying moment" (per The Decoder). There's been internal pushback at OpenAI over "engagement farming" tactics, and users already assume product suggestions are sponsored, which creates trust issues.
The more realistic monetization path for Meta (based on their earnings guidance):
Better ad targeting through AI (already happening - 11% increase in ad impressions Q2 2025) AI business tools (Meta AI for businesses) Infrastructure-as-a-service (selling excess capacity) Direct advertising in AI responses faces major regulatory and user trust hurdles. ChatGPT reached 800M weekly users by Sept 2025, but monetization through subtle product placement has sparked backlash even internally at OpenAI.
Sources: Meta Q3 2025 earnings preview (LSEG, Nasdaq), The Decoder (June 2025), Yahoo Finance marketing analysis
zekrioca•3mo ago
0xrelogic•3mo ago
For the Hyperion deal specifically (confirmed by Meta's official announcement Oct 21, 2025):
80% owned by Blue Owl Capital (private credit firm) Financed through $27B+ debt arranged by Morgan Stanley PIMCO as anchor lender (144A bonds, maturing 2049) Meta has 20% equity + 16-year residual value guarantee So Meta's "guarantee" is essentially a long-term lease commitment with a 4-year initial term + extension options. If it goes bust, Meta is still on the hook for 16 years of payments, but Blue Owl/PIMCO absorb most of the asset risk.
The other deals (CoreWeave 14.2B,Oracle 20B) are traditional service contracts Meta pays for capacity, vendors own the infrastructure.
This is actually the largest private capital deal on record according to Bloomberg (Oct 16, 2025).
Sources: Meta official announcement, Bloomberg, Data Center Dynamics