Sensemaker

Startups tame the chaos of multimodal data

Enterprises are overloaded with unstructured, multimodal inputs. The opportunity is continuous systems that clean, structure, validate, and govern multimodal data so downstream AI workloads actually work.

— Jennifer Li

AI revives cybersecurity hiring

AI will automate repetitive security work enough to break the hiring crunch cycle. AI-native tools help teams identify what to automate, freeing people to focus on real adversaries.

— Joel de la Garza

Agent-native infrastructure becomes table stakes

Infrastructure is shifting from human-speed to recursive, bursty "agent-speed" workloads. Legacy backends weren't built for one agentic goal to fan out into thousands of sub-tasks.

— Malika Aubakirova

Creative tools go multimodal

2026 is when tools accept whatever reference media you have and help you extend, edit, reshoot, and match motion with far more control—from meme-makers to Hollywood.

— Justine Moore

The AI-native data stack continues to evolve

Truly AI-native architecture is still early. The next phase is tighter linkage between data and AI infrastructure, including how agents pull context across systems of record.

— Jason Cui

The year we step inside video

Video shifts from something we watch to something we inhabit. Models sustain characters, objects, and physics long enough for actions to matter—a buildable medium for simulation and games.

— Yoko Li

Systems of record lose ground

AI turns CRM/ITSM from passive databases into workflow engines. The classic system of record becomes commodity persistence while leverage shifts to intelligent execution environments.

— Sarah Wang

Vertical AI evolves to multiplayer

The next unlock is "multiplayer" coordination across parties with distinct permissions. Multiplayer systems coordinate stakeholders, sync changes, and let counterparty AIs negotiate.

— Alex Immerman

Creating for agents, not humans

Optimization shifts from human attention patterns toward machine legibility. The new design problem is structuring information for agent retrieval, not just human clicks.

— Stephenie Zhang

The end of screen time KPI

Outcome-based pricing pushes apps away from time-spent metrics. Users capture huge value with minimal screen time as AI automates work—the challenge becomes measuring ROI differently.

— Santiago Rodriguez

Healthy MAUs emerge

Consumers who aren't sick but want recurring health monitoring become a major segment. AI lowers costs and prevention-oriented insurance products make "healthy MAUs" viable.

— Julie Yoo

World models take the spotlight

AI world models create explorable 3D environments from text, pushing storytelling toward interactive worlds—a "generative Minecraft" and simulation grounds for training agents.

— Jonathan Lai

"The year of me"

Products stop optimizing for the average consumer and start optimizing for the individual. Personalization becomes the defining product advantage across education, health, and media.

— Joshua Lu

The first AI-native university

An institution built around intelligent systems that adapt continuously—courses, advising, research updating via feedback loops. Evaluation shifts to grading how students use AI.

— Emily Bennett