- AngelsRound
- Posts
- Lapis ðŸ§
Lapis ðŸ§
Cross channel AI advertising platform.
Spotlight
What if the hardest part of AI advertising is knowing where to spend?
Quick Pitch: Lapis is building a cross channel ad platform that decides where to spend, what to run, and how to optimize campaigns across ChatGPT, Google, and Meta.
The Problem
Fragmented Stack: Paid ads still require separate tools for creative, media, landing pages, and analytics.
Walled Gardens: Google and Meta optimize spend within their own platforms, not across channels.
New Playbook: Conversational ads require targeting and creative built around natural language intent.
Why It Matters
Businesses spent roughly $410B on Google and Meta ads in 2025, much of it still managed through agencies and fragmented tools.
Now ChatGPT adds a new advertising channel with a different form of consumer intent.
As spend fragments across search, social, and conversation, deciding where and how to spend becomes more complex.
Snapshot
Industry: Advertising Technology
Headquarters: San Francisco, CA
Founded: 2025 (YC F2025)
Funding: Raising $6M (Seed+)
Backed By: Y Combinator, Founders Future
Traction: More than 1,500 marketing teams and dozens of enterprises across approximately 10 languages
Target Customers: Businesses spending $10K+ annually on paid ads, from growing brands to enterprises
Founder’s Edge
Varunram Ganesh, Co-Founder, CEO: Former Head of Growth at Warp, where he helped scale revenue from zero to $2.5M ARR.
Sai Surbehera, Co-Founder, CTO: Former Walmart search engineer with experience in AI and large scale retrieval.
Playing Field
Ad Platforms: Google and Meta automate ads within their own inventory.
Agencies: Manage across channels, but remain service heavy.
AI Ad Tools: Mostly automate creative and individual workflows.
Lapis's Edge: Learns the brand, targets the audience, predicts what will work, and continuously reallocates spend across channels.
Analysis
Bulls Case 📈
Campaign execution expands revenue potential per customer.
Cross channel optimization gives Lapis a structurally neutral position.
Campaign data could compound into better predictions.
Early ChatGPT ads give Lapis a head start.
Bears Case 📉
Managed services could constrain margins and scale.
ChatGPT ads may take years to reach meaningful spend.
Platforms control the inventory and data Lapis depends on.
The data moat depends on sustained customer spend.
Verdict
Consumer intent is expanding beyond search boxes and feeds into AI conversations, where users explain what they want before visiting a brand.
If that behavior scales, the valuable layer may be the one that interprets intent and directs ad spend across platforms. Lapis wins if it becomes that layer rather than another ad automation tool.
The Startup Pulse
Another happening week in startup funding. Three signals from this week:
Inference is becoming a major market of its own
AI applications can command huge valuations when revenue is real
Physical AI is expanding beyond humanoids into construction and mobility
Etched: Secured $700M at a $21B valuation. Lead investor Jane Street was already a customer, signaling real deployment beyond benchmarks.
Higgsfield AI: Landed $400M in Series B funding. With ~$700M in annualized revenue, it shows how quickly paid demand for AI applications can scale.
Groq: Pulled in $350M to expand its inference cloud, reflecting the growing opportunity in inference as AI usage scales.
Supported By
Strategic breakdowns of how the biggest AI companies actually work — their business models, their moats, their vulnerabilities. This week: Everyone said AI would kill Google. Here’s what actually happened.
Written by Ashher
