At Felicis, our philosophy has always been to back exceptional founders right before their moment of breakout.
Today, we’re seeing the start of AI-driven trends that will unlock the next wave of market-defining companies. If you're starting something in one of these areas, we’d love to hear from you.
That said, some of our favorite investments have come from founders who imagine something completely new. If you’re pursuing an idea outside these themes, we’re just as eager to hear from you.
Here are some of the promising areas that stand out to us:
The Merging of Knowledge Work and Software
AI-Native Tooling for Every Industry
Next-gen AI enterprise tools won’t just automate, they’ll inspire with their user experience. Whether in design, finance, marketing, recruiting, sales enablement, or more, we want to back intuitive, AI-first software that helps humans do higher-order work.
Agents Enter the Workforce
Agentic engineers and support reps are just the beginning. Back office agents and more are coming. We want to fund the new “employees” shifting jobs that are typically offshored (support, IT, data entry, HR) and changing how companies operate.
The AI-Powered CFO
From helping CFOs with accounting, reconciliation, and closing their books, there are a number of ways AI can help finance teams. We’re excited about a future where AI can be even more impactful and do complex work like forecasting.
AI Reinventions of Leading Software Platforms
AI is moving from platforms into workflows. Tools like your CRM or HRIS will be AI-enhanced and work proactively. These will be revolutionary applications that improve dramatically the more they are used.
Resilience and Defense for a More Secure Global Future
Autonomous Systems
We can no longer wait for reaction times of humans - advancements in perception, sensors, hardware, and models will introduce a greater level of autonomy to both leverage and defend against.
Energy Resilience
Energy is a key bottleneck for AI and will be for the foreseeable future. It’s time to reinvent how we capture, manage, and distribute energy and make it as efficient as possible.
Security That Keeps Up With Increasing Complexity
AI-native Observability, Governance, and Compliance
Companies need clear insight into what their AI is doing and why, for both security and compliance. An agentic web means new primitives are needed for access, authentication, and more.
Identity to Manage AI Scale
Agents need guardrails, including scoped access, verifiable actions, and runtime controls and monitoring. In the same way that OAuth standardized user access, MCP and A2A will demand new agent-native access control layers.
Proactive Security Services
The velocity of change requires security that anticipates and prevents issues, not just reacts.
Infra That Enables AI-Native Apps at Scale
Infrastructure for Agents
As agents become the primary users of software and the web, infra must evolve to meet their needs with how they operate and handle data. We want to fund the layer that makes agentic work seamless, safe, and builds on new standards like A2A and MCP enabling agents to work across systems.
SDLC for the AI Developer
Within the decade, AI agents will write 90% of code. As a result, the software development lifecycle (SDLC), which requires gathering, design, implementation, testing, deployment, and maintenance, must be radically reshaped.
The AI Middle Layer
There is so much more that needs to be built out to help AI apps, from evaluation and orchestration to testing and model routing. We believe the AI middle layer will ultimately prove more valuable than the cloud’s middle layer.
Voice as the Default Interaction Layer
Voice AI gets more capable by the week, and will become foundational infrastructure for AI applications. There is room for a sizeable middle layer to build value for developers of voice AI apps.
Healthcare Systems Become Massively More Efficient with AI
AI as the Admin Layer for Healthcare
From prior authorizations to claims processing, the admin layer of healthcare is ripe for automation. These workflows are structured, painful, and perfectly suited for LLMs.
AI for Life Sciences
The drug development process takes 10-15 years, requiring constant paperwork and documentation. The AI products in this category will have high ACVs and be incredibly sticky.
Intelligence Unleashed
Self-Improving AI
Self-improving AI and agents that take on increasingly complex and consequential decisions. Agents that are truly autonomous, working nonstop and 10x faster.
The Robot Revolution
Just like smartphones birthed a trillion-dollar ecosystem, general-purpose robots will redefine labor, logistics, and daily life. We think highly of robotics teleops, data generation and simulation, and vertical applications, but are open to being surprised.
Vertical Agents
RL environments paired with expert human feedback are bridging the gap between models and economically valuable work. From consulting to financial analysis, a Cambrian explosion of AI apps is underway, and RL is the catalyst.
