AI is once again the buzzword in this latest batch of YC Demo Day companies. Nearly 190 companies joined Y Combinator's Winter '26 cohort and announced their startups at Demo Day on Tuesday.
These companies work on products across industries such as legal, transportation, and healthcare.
Admittedly, given the sheer scale of attendance and the format of this year's Demo Day with media availability, I didn't hear every product pitch. YC posted each pitch video (rather than a livestream or in-person session invitation) about 20 minutes after the presentation.
Instead, I read about all 190 startups that presented, spent a day watching the pitches of startups I thought were interesting, and narrowed it down to the 16 that stood out as the most interesting startups in this crowded YC class.
ARC Award Foundation
Feature: Create benchmarks to help measure your progress towards AGI.
Why it's interesting: YC's nonprofit organization! But again, given that OpenAI, Anthropic, and GoogleMind already use organizational benchmarking in some form, I can understand why it would be included. The foundation aims to encourage more open source AGI research by hosting competitions and awarding research grants. One of the reasons for this AI revolution is to reach AGI (which Nvidia CEO Jensen Huang has said is already there), and it will be a historical record of how close we are to AI machines with general intelligence.
asimov
Description: Collect human movement data to train humanoids.
tech crunch event
San Francisco, CA | October 13-15, 2026
Why it's interesting: People around the world send videos of themselves moving or performing tasks to the company, which can be turned into data sets useful for training robots. It's part of a movement to turn humanoids into objects and find uses for them beyond supply chains and entertainment. The era of “Rosie the Robot” may still be a long way off, but I'm bullish on humanoid technology. Using data to teach humanoids the flow and, dare we say, grace, of human movement could help robots perform tasks less robotically.
avoice
What it does: Helps architecture firms automate tedious non-design tasks.
Why it's interesting: It's not every day you hear about new technology targeting the construction industry. The founder himself pointed out that this market is underserved (despite its rich potential). The tool uses AI to automate tasks that creative types like architects can find tedious, such as reviewing specifications, drawings, contracts, and proposals.
button computer
Feature: Wearable AI
Why it's interesting: The world is waiting for products from OpenAI, which acquired Jony Ive's company, and everyone is trying to make wearable AI a reality. Two former Applers (can we call them that?) have teamed up to start Button. The founders explained that Button is essentially a small computer, made for AI. Buttons connect to apps like email, Slack, and Salesforce, and use voice commands to interact with the apps and perform specific tasks. The next must-have hardware will likely be some form of AI wearable, so it will be interesting to see what comes along.
code wisp
What: Enabling anyone to build games using AI
Why it's interesting: The founders say all you have to do is teach an AI how to make games and it will make the games for you. It's fun, creative and exciting! When I was a kid, I tried to make games and always found them difficult and boring, but the excitement I had when designing them never went away. Vibecoding is all the rage in app building, and tools like this make imaginative execution much easier. Perhaps this will be the next generation of vibe building.
cross layer lab
Feature: Helps detect website spoofing
Why it's interesting: The rise of agent tools means it's easier to spoof websites, and bad actors aren't afraid to use this technology to trick people. Crosslayer Labs helps customers discover and monitor their online settings to help them stay protected from this new Internet threat.
Dwemerjon
Content: Teach language while doomscrolling
Why this is interesting: We spend so much time doomscrolling that it just fills up our brains and ends up making us feel irritated and atrophy. The startup is an app that shows users short videos in the language they're trying to learn, such as how it looks and scrolls on their TikTok feed. If there was any meaning to all the content we cram into our heads all day long, what a great way to find out. It combines something consumers can't stop doing (spending hours swiping through content on their phones) with something cool like learning a language. I'm very interested.
Lexius
Capability: Incorporate advanced AI into your security systems
Why it's interesting: It uses AI to enhance existing security camera systems, allowing them to detect and report incidents of theft and falls on video, replacing a fragmented and largely manual process. The company said it was targeting businesses with cameras without AI intelligence that could have captured the incident, but the company was not taking any action.
library research room
Features: AI-powered library management system tools
Why it’s interesting: This is AI entering one of the industries often overlooked by the tech industry: libraries. The startup has created an AI-powered library management system specifically to help schools with cataloging and cataloging. As the founders said in their pitch, there isn't much competition when it comes to automating or innovating on tools already in use in this space, so any new idea is a candidate for the “next big thing.”
Millirei
Feature: Radar system to help track small drones
Why it's interesting: Defense technology is one of the hottest categories in technology right now. As the founder of this company said, people are currently trying their best to track small drones in the field, but the human eye can miss many things or mistake small drones for birds, and vice versa. The startup uses sensors to identify what tiny drones actually are in the sky. Given the current geopolitical situation, new technologies are needed and emerging to help all countries stay ahead of the game and their adversaries.
mouse cat
What: Investigating fraud using AI
Why it's interesting: AI is the perfect tool for the job, and that's also true for people whose jobs include fraud and fraud. The company pulls a company's data from large cloud storages like Databricks and Snowflake, analyzes consumer data and activity for anything suspicious, and provides recommendations to take action. These AI-native tools are critical to combating the malicious AI that continues to emerge.
opalite health
What it is: Uses AI to help healthcare providers converse with non-English speakers
Why it's interesting: When two people can't understand each other, there's a lot left up to interpretation. In the medical world, it can be a matter of life or death. This AI medical translator breaks down language barriers and helps healthcare providers understand patients who speak different languages. In a globalized world (and a country as diverse as the United States), it's important that people have access to the health care they need, regardless of their language. Of course, the idea is not entirely original, as many other startups and health tech providers offer similar services.
sequence market
What it is: Allows you to trade different markets such as cryptocurrencies and predictions on one system.
Why it's interesting: As someone who likes to have everything in one place, I understand that you don't want a more fragmented process when executing trades in these markets. It's the same reason I still like going to big box stores to buy branded items (I like seeing all the options at once).
Shofo
Features: Basically all video libraries
Why it's interesting: The startup bills itself as “the world's video library.” This is pretty great if you ask me. I grew up on YouTube and Tumblr, and I remember how sensitive searches were back then when you were looking for something specific. It's like a custom video index that helps AI Labs efficiently find diverse datasets, but I love tools that make it easier to search and organize.
sonary
Function: Helps fix operational issues in the software itself
Why it's interesting: This startup is developing technology that sounds pretty cool. Connect to other monitoring systems, reduce alert noise (a distraction from finding alerts that actually matter), automatically identify the root cause of problems, and find ways to fix them or suggest further actions to engineers. Although the number of AI code review startups is growing (the functionality is also provided by model makers), there should also be room for independent startups once the code is deployed in production systems. This is another aspect of the workflow that founders are automating.
Terranox AI
What: Using AI to find uranium deposits in North America
Why it's interesting: The company's founders reminded us that uranium will be needed to power the next generation of nuclear energy. (Although nuclear power is generally considered safe these days, uranium itself is clearly toxic, and the founders didn't include in their pitch how it could be drilled safely.) These founders believe that nuclear power will be needed to power all the new data centers being built. The Earth will require large amounts of energy from a variety of sources to ensure that the human population can keep up with the ambitions of the AI revolution.