May 13, 2026

Specs Become The Agent Control Plane

New Intel: OpenSpec is using specs to coordinate 30+ coding agents across teams, repos, IDEs, and planning tools. The enterprise wedge is making agentic coding usable in messy brownfield codebases.

OpenSpec

|

Developer Tools

|

YC W26

May 13, 2026

Medical Interpretation Becomes Clinical Documentation

New Intel: Opalite Health is pairing real-time medical interpretation in 150+ languages with EHR documentation. That turns interpreter cost savings into a clinical workflow product for hospital systems.

Opalite Health

|

Medical Interpretation

|

YC W26

May 13, 2026

World Models Become Robot Testbeds

New Intel: One Robot is building task-specific world models for training and testing robot policies in simulation. The wedge is realistic manipulation testbeds for VLA models before they hit factories or labs.

One Robot

|

Robotics Simulation

|

YC W26

May 13, 2026

Fish Inspection Moves Underwater

New Intel: OctaPulse is using robotics and computer vision to inspect fish in under 30 seconds with 90%+ accuracy. Underwater robotics could expand the product from trout grading into broader aquaculture health monitoring.

OctaPulse

|

Agriculture

|

YC W26

May 13, 2026

Hardware Compliance Gets An Agentic Workflow

New Intel: Noetic is automating hardware compliance research, documentation, and lab matching across standards like FCC, CE, FDA, UL, and ISO. The wedge is cutting certification timelines before hardware launches stall.

Noetic

|

Regulatory Compliance

|

YC W26

May 13, 2026

Program Synthesis Pushes Toward Embodied Intelligence

New Intel: Ndea is pairing program synthesis and ARC-style reasoning with world models, reinforcement learning, and robotics signals. That points toward embodied intelligence, with software benchmarks as the starting point.

Ndea

|

AGI Research

|

YC W26

May 13, 2026

Fraud Teams Get Autonomous Investigators

New Intel: MouseCat is building autonomous fraud investigators that review cases, generate rules, backtest them, and push into rule engines. That closes the loop between fraud ops and production risk systems.

MouseCat

|

Risk & Fraud Detection

|

YC W26

May 13, 2026

Conversation Replay Becomes Agent Debugging

New Intel: Moda is building monitoring for AI agents that catches behavioral failures, tool errors, prompt attacks, and replayable conversation issues. That makes agent reliability a category distinct from normal app observability.

Moda

|

MLOps

|

YC W26

May 13, 2026

Human-Backed AI Targets Care Reliability

New Intel: MochaCare is combining AI agents with human support for home care scheduling, caregiver workflows, and compliance. That hybrid model matters in a market where failed automation can mean missed patient care.

MochaCare

|

Home Care Tech

|

YC W26

May 13, 2026

Plain-English BI Gets An AI Analyst

New Intel: Menza is building a 24/7 AI analyst across 650+ data sources, with anomaly detection, forecasting, and plain-English analytics. Strong retention suggests the product may replace recurring analyst work and reduce dashboard dependence.

Menza

|

Data Analytics

|

YC W26