The AI Layer Cake: Where the Real Opportunity Lies in Enterprise AI
Over the last 18 months, AI has transitioned from hype to implementation. However, not all AI opportunities are created equally. Founders looking to build enduring businesses need to understand where real value is being created and where it isn’t.
At Sierra Ventures, we use a framework we call the AI Layer Cake—a five-layer model that helps us evaluate the enterprise AI stack and identify the most compelling opportunities.
This layered view offers founders a strategic lens for building in today’s rapidly evolving AI landscape.
The Five Layers of AI Startup Opportunity
Core AI Infrastructure
This bottom layer includes the foundational hardware and systems powering AI: GPUs, data centers, and the hyperscaler backbone used by AI labs and model providers. These businesses often require deep technical specialization and massive capital investment. Sierra does not invest at this layer.
Applied AI Software Infrastructure
This is where general-purpose models are tailored to specific technical use cases—enhancing performance, accuracy, or user experience in targeted domains. Sierra invests at this layer. One example is Cimulate AI, which improves e-commerce conversions by delivering AI-personalized search experiences using edge LLMs that are context-aware of a customer's e-commerce infrastructure. Another example in the AI Data Orchestration infrastructure company Astronomer which builds, runs, and maintains data pipelines using context to uniquely generate and maintain them.
Horizontal Applications
These are industry-agnostic AI tools designed to augment common business functions like legal, HR, or sales. Sierra is active in this layer as well. Eudia, for example, reduces legal department costs for large enterprises by automating and accelerating contract workflows.
Vertical Applications
This layer focuses on domain-specific AI built for industries like insurance, healthcare, or logistics. Sierra invested in Weav.ai, which uses AI to streamline insurance underwriting by automating policy review and risk analysis.
Novel Innovations
The top of the stack involves entirely new business models and data ecosystems enabled by AI. These companies often don’t resemble traditional SaaS. Sierra invested in Revalia Bio, a biomedical company creating organ testing infrastructure that generates new datasets to speed up drug development.
Why the Middle and Top of the Stack Matter Most
While infrastructure is essential, it’s increasingly dominated by hyperscalers and capital-intensive players. The more attractive opportunities for startups exist further up the stack—where applied intelligence intersects with real workflows, and where new data unlocks new business models.
In enterprise, the value isn’t in generalized intelligence—it’s in contextual execution. Startups that deeply understand the problem space and embed AI into specific workflows will have the edge.
The Rise of Specialized AI Agents
We're seeing a shift from general-purpose tools to specialized AI agents. These agents operate within well-defined functions or industries and are already delivering results:
- Insurance: Weav.ai is automating underwriting through real-time AI-driven analysis.
- Legal: Eudia is streamlining enterprise contract work with AI copilots.
- E-commerce: Cimulate AI is boosting conversions with intelligent product discovery.
These agents are not just support tools—they are intelligent systems executing work that previously required significant manual effort.
What Founders Should Be Thinking About
As you build, ask yourself:
- What unique workflow, dataset, or domain advantage are we unlocking with AI?
- Are we building a tool—or a system that delivers outcomes autonomously?
- How defensible is our position against horizontal platforms?
The most successful AI startups won’t just add intelligence to existing software—they’ll reimagine the entire product experience around it.
Looking Ahead
AI adoption is accelerating in the enterprise, and the companies best positioned to lead are those building in the middle and upper layers of the stack. Sierra Ventures is investing in founders who are bringing real intelligence to real problems—at scale.
If you're building in applied infrastructure, horizontal or vertical applications, or creating novel AI-native business models, we’d love to connect.