LlamaIndex CEO Says AI Scaffolding Layer Collapse Is the Point, Not the Problem

Quick Facts

• LlamaIndex CEO Jerry Liu says the collapse of AI scaffolding frameworks validates the shift to natural language programming
• About 95% of LlamaIndex code is now generated by AI, with engineers typing in natural language instead of traditional code
• The company hit $10.9 million in revenue in June 2025 after raising $19 million in Series A funding

The scaffolding layer that developers once needed to ship LLM applications is collapsing. Jerry Liu, co-founder and CEO of LlamaIndex, says that’s not a problem. It’s the point.

Liu explains there’s less need for frameworks to help users compose deterministic workflows in a light and shallow manner. Even he acknowledges these types of frameworks are becoming less relevant.

“This means the layers between programmers and non-programmers is collapsing, because the new programming language is essentially English,” Liu said.

Natural Language Programming Takes Over

About 95% of LlamaIndex code is generated by AI. “Engineers are not actually writing real code,” Liu said. “They’re all typing in natural language.”

With every new model release, AI demonstrates better capabilities to reason over massive amounts of unstructured data. Tasks that would break AI agents three years ago now work with simple commands.

“This type of stuff was either extremely inefficient or just would break the agent three years ago,” Liu said. “It’s just way easier for people to build even relatively advanced retrieval with extremely simple primitives.”

Context Becomes Key Differentiator

When the technology stack collapses, Liu identifies context as the core differentiator. “Agents need to be able to decipher file formats to extract the right information,” he said.

LlamaIndex has built its business around this principle. The company developed data processing technology as the foundation for LlamaCloud, a knowledge management platform that handles everything from data ingestion through agent deployment.

Notable customers include Rakuten, Carlyle, Salesforce and KPMG. Phil Mui, senior vice president of Engineering for Agentforce at Salesforce, said LlamaIndex provides advanced async workflow abstractions that enable building scalable concurrent agents much faster.

Market Dynamics Shift

The collapse reflects broader market changes. Tech companies now offer 50% price reductions to retain clients considering building instead of buying software.

This “SaaS-pocalypse” is driven by what the industry calls vibe coding. Publishers and startups bypass traditional vendors, using natural language prompts to build custom tools in hours rather than years.

LlamaIndex’s business metrics support Liu’s thesis. The company hit $10.9 million in revenue in June 2025 and recently secured $19 million in Series A funding, bringing total funding to $27.5 million with a post-money valuation of approximately $93 million.

Infrastructure Investment Risks

The shift comes as massive AI infrastructure investments face scrutiny. Moody’s noted over $2 trillion in data center investment is planned between 2025 and 2028, but projects face execution risks from power constraints and construction delays.

AI adoption data shows companies with 250 or more employees began declining adoption rates in mid-2025, while smaller firms maintained steady growth. Only 54% of AI projects made the transition from pilot to production in 2025.

Liu’s position suggests LlamaIndex sees the scaffolding collapse not as a threat but as validation of their strategy to focus on context and data quality over complex development frameworks.

Read more: The AI scaffolding layer is collapsing. LlamaIndex’s CEO explains what survives.