AI (Artificial Intelligence) Startups funded by Y Combinator (YC) 2026

Scale accelerates the development of AI within organizations of any size to deliver critical business insights and operational efficiency. Its data-centric infrastructure platform leverages RLHF (Reinforced Learning with Human Feedback) to help organizations build the strongest AI models that supercharge their business, with customers across industries including Meta, Microsoft, U.S. Army, DoD’s Defense Innovation Unit, Open AI, General Motors, Toyota Research Institute, Brex, Instacart and Flexport.

Podium is the all-in-one AI-powered lead management and communication platform used by more than 100,000 businesses to acquire and convert new customers. At the forefront of Podium’s innovation is its AI employee, who ensures businesses respond to leads instantly, anytime of the day or night—significantly increasing lead conversion rates and revenue. Podium helps businesses obtain more reviews, rank higher on Google, and consolidate lead channels. Businesses can call and text customers from the web and/or mobile app, send payment links, drive repeat business through bulk messages, and more—all managed and delivered from one easy-to-use dashboard. Additionally, with Podium’s AI Employee and automations, businesses can efficiently handle customer inquiries seamlessly across all communication channels, providing timely and accurate responses that drive sales. Podium’s innovative work has gained recognition on top industry lists, such as Forbes’ Next Billion Dollar Startups, Forbes’ Cloud 100, the Inc. 5000, and Fast Company’s World’s Most Innovative Companies. Podium was founded in 2014 and is headquartered in Lehi, Utah. It is currently backed by Accel, Summit Partners, GV (formerly Google Ventures), and Y Combinator.

Checkr builds people infrastructure for the future of work. And we believe everyone should have a fair chance to work. That's why we've designed a faster—and fairer—way to screen job seekers. For more information on our mission and products, visit https://checkr.com.

Top apps need weeks and multiple teams to test every release. Aucert does it in hours, and plugs into agentic coding workflows so issues get fixed in the same loop. Agents: Functional QA · Accessibility · Google Play Policy · App Store Policy Next: Performance · Security. DM if you need either.

Plyxl is the workspace that molds to your business. It builds apps, creates automations, and works across your entire company.

Tilo is a Linux-based operating system built for AI agents to work alongside people. On Tilo, computer use is up to 15x cheaper than on a normal OS. Today agents use computers the way people do. They take screenshots and click through apps, which is slow, expensive, and brittle. In Tilo, every app exports actions, so agents work over MCP while people use the UI. You still see what each agent does in your apps, and anything it changes can be undone. Any number of people and agents can work in one workspace in real time, across apps. Tilo runs as your OS, in our cloud, or headless on your machine, and you can open and share any workspace from the browser.

Altair is an AI-native medical billing service for independent practices. We take over a practice's billing end to end and get doctors paid by insurers. AI does the work; our billers handle the exceptions. That can only work if a practice can trust an agent alone with its revenue. So we formalized insurers' billing policies and put a theorem prover under the agent. Before the agent acts on a claim, the prover checks the action against that insurer's policies, and anything that fails goes to one of our billers. Every claim teaches Altair how each insurer behaves, so the longer we run a practice's billing, the better we predict what its insurers will pay, deny or delay. We're building the default way a new practice gets paid: a doctor connects their EHR and bills clean from the first patient, completely self-serve and on auto-pilot. Altair will be the financial platform every independent practice in America starts on and stays on.

At Praxis, we’ve built AI agents that run law firm operations, from first inquiry to paid invoice. Agents do the admin end to end, and the lawyer supervises and approves every step. Firms reply in minutes, onboard clients in days instead of weeks, and spend their time practicing law. Lawyers spend 62% of their day on admin nobody pays for: intake, conflict checks, chasing documents, engagement letters, follow-ups, billing. Every hour on admin is an hour a client waits, and clients go to the firm that answers first. One law firm owner shared: “Every month we lose 3-5 clients, $10k each, because we’re 3 paralegals to handle the process, it’s too slow.” We’re live with law firms across the US.

Graphide (YC F26) is a code editor that maps your codebase into a graph so you can see what your agents are building. Agents now write code faster than anyone can read it, and a line by line diff doesn't show what actually changed in the system. Graphide keeps a live graph of the modules, classes, functions, etc. how they connect, and what each agent is touching as it works. Graphide then checks their changes against your architecture, making sure your agents are implementing what you had in mind.

Caveman is the optimization stack for AI agents. It observes real agent workloads, finds cost and performance inefficiencies, tests improvements, and safely deploys the ones that work.

Precision strike is rationed. Forces see more targets than they can afford to hit. Polemos builds AI-guided ballistic missiles and the compact launchers that fire them, designed for mass manufacturing. Our launchers form a strike network wired into the recon drones already in the sky: a drone geolocates a target, the nearest launcher fires, and a guided round lands in seconds. Short range, no wasted kilometers, no modified trucks. Built on production lines that already exist, so output scales when it matters most.

Officers join the force to protect their communities, yet over 40% of their day is eaten by paperwork that keeps them behind a desk. The burden reaches beyond lost time. Burnout is driving officers out of the force, leaving those who remain carrying an even heavier workload. Reconero is built to give that time back. By automating reports, forms, analytics, and more, we help officers finish their administrative paperwork faster, spend more time serving their communities, and protect the time they have with their families.

Most phones are built to keep you on them. We make an agentic phone that puts an army of agents to work for you. Tell it what you need done, and they handle it across the tools and apps you already use. You get more work done with less time on your screen.

Pomerado turns any website into an API / MCP in minutes. Empower your agents to use previously inaccessable consumer websites, legacy portals, and internal tools. You or your agent describe a task, Pomerado asks a few clarifying questions, and then delivers an integration. We manage hosting, permissioning, testing, and maintenance, so you can focus on building. As consultants at McKinsey and Scale AI, we saw that legacy systems were the true blocker to agentic workflows. Pomerado connects those systems to agents.

Rodeo builds a robotic technician that automates maintenance so fleets can get vehicles back on the road sooner. Initial deployments focus on oil changes at centralized depots, where routine maintenance competes for scarce technician time. Cofounders Michael Hazard and Colby Duke met at Applied Intuition, where they worked with OEMs like Toyota and General Motors to bring software and autonomy to vehicles. They saw firsthand how much manual work it takes to keep vehicle fleets running, and learned how to make autonomy reliable through structured testing, real-world data, and human intervention. They’re now applying their experience to the roughly $600B annual US vehicle maintenance market. The average US vehicle is older than ever, driving up maintenance demand at a time when shops already struggle to hire and retain technicians. That gap leaves fleets facing costly downtime and preventable breakdowns. Despite this need, vehicle maintenance has resisted traditional automation. Service information is fragmented and unreliable, and vehicle hardware varies significantly between makes and models. Rodeo builds on advances in robotic foundation models to learn maintenance skills that transfer across vehicles. Rodeo pairs robots with human technicians who ensure each job is completed correctly, handling work the robots can’t yet perform autonomously. These deployments generate training data for Rodeo’s models through both teleoperation and hand tools augmented to record how technicians work. This lets Rodeo earn revenue while collecting the data needed to automate more of each job and improve margins.

Native is building an autonomous marketing platform where business can do all of their marketing within one platform using our artificial intuition model called Norn, that predicts what people will respond to, researched and built in house.

Munito is an open-model inference provider for Europe, private by default. Its API lets users choose their models and keep inference and storage in the EU. Founded by the former inference lead at Mistral AI, Munito is built on one belief: open and independent intelligence, both models and inference, is key to economic growth and prosperity.

Check features, not code. Composal sends a swarm of agents to use your product like your users do. They test real flows and edge cases to show you what works, what breaks, and if it’s ready to ship.

Quivly is the agentic workspace for everything after the sale. Our agents turn kickoff calls into launch plans, chase blockers, catch risk and expansion signals, and have the full account story ready at renewal. With Quivly, companies are able to take their customers live in 7 days instead of 30, and then grow revenue from these customers by 4.3x on avg.

Starting with a modern interface for drones that users can build on and adapt. Works with planes, copters, rovers, boats and submarines.

Databuddy gives you your very own data analyst, always watching over your data, combining all your context to investigate and paint a full picture on every product issue that emerges in real-time; conversion drops, user frustration and churn. It makes your product better without needing to have you look at a dashboard and spend hours figuring out the why. one click to install or switch from any dated analytics provider.

Resin is an AI-native ECAD tool. It lets engineers: - research designs - generate schematics - generate any symbols/footprints - route simple to complex boards

Latent catches wrong or unsupported LLM answers before they reach your users. Instead of running a second model to check every output, it reads the model's own internal state as it generates, so every request is verified with no added latency, no extra inference cost, and nothing leaving your deployment. Teams running LLMs in production, in finance, legal, healthcare and customer support, get a review queue of the answers most likely to be wrong, calibrated to their own model and traffic, with a plain explanation of why each one was flagged.

Ploid is the AI-native people data infrastructure for developers and companies building products, workflows, and agents around people. We've built an index of 2B+ people across professional and social sources, giving developers one API to search for people, enrich profiles, find emails and phone numbers, resolve identities, and retrieve up-to-date information from across the web. Unlike traditional people databases built around static records, Ploid combines its massive people index with real-time retrieval and identity resolution, delivering fresh, reliable information whenever it's needed. From sales and recruiting to customer intelligence, trust and safety, and AI agents, Ploid replaces fragmented data providers with a single infrastructure layer. 2B+ people. One API. Built for AI.

Agents will do most of the work in the future, but only if teams can trust them unsupervised. Stacktrace is the control layer for every agent: it sees every action, detects risky behavior, enforces the bounds your team sets, and steps in when an agent drifts from intent.

We’re building a brain-computer interface anyone can wear, so you can use a computer at the speed of thought. To get there, we’re recording brain data at model company scale.

We build infrastructure to power agentic systems such as software factories, multiplayer agent frameworks and agent coordination.

CONUS structures fragmented medical device information for surgeons. We combine manufacturer documentation, hospital protocols and medical research in one place, giving surgeons and MedTech teams reliable answers across products and procedures. We already work with leading manufacturers including Johnson & Johnson, Stryker and Zimmer Biomet, and since launching six months ago we’ve gone live in 20 hospitals. Long term, hospitals will use CONUS to select and order implants across manufacturers.

Port your development environment to the cloud. Give every coding agent on its own cloud computer, bring your Codex or Claude subscription, start agents from Slack and Linear, and manage them from your desktop, phone, or terminal.

UFO is the operating platform for your business built by the creators of Github Copilot and Perplexity Computer. UFO learns your company, goals, tools, and workflows. UFO builds what your team needs to keep work moving.

OpenHack is an always on AI Security Engineer. OpenHack understands your business, system design and developer intent like a real security engineer. It ingests all this context to find the hardest vulnerabilities through AI codebase scanning, AI pentesting and Supply Chain Security. It verifies and prioritizes vulnerabilities automatically, and filters out false positives like a real security engineer drastically reducing overhead.

OpEra Technologies is the AI decision engine for factories, warehouses and fleets. Companies with physical operations make thousands of decisions every day, and we help their teams make the best one at every step from production to delivery: (1) Planning: how much to make, buy and stock. (2) Scheduling: what runs when, and who works on what. (3) Allocation: which warehouse or store gets each order. (4) Routing: boat, plane or truck, and which route each one takes. Founded out of MIT's Operations Research Center.

https://musicarena.io/

Hue turns production usage of agents into repeatable tests that preserve the tools, data, and state behind the real run. As AI agents take on real work, their behavior depends increasingly on the nuances of the world around them. With Hue, teams can test and improve their agents under conditions that mimic real usage.

Mantle is a full-stack field service company that deploys, fixes and maintains machines for the companies that build them. We're not a marketplace or software tool, we own uptime directly to provide better, more reliable service without burdening your engineers on support calls. $500B+ a year is spent to keep machines running in the US, we're coming for all of it.

Ageospatial helps organizations value the risk on their physical assets (properties, land, infrastructure) remotely using satellite imagery. 97% of commercial properties are never physically inspected because it's too slow and expensive. So insurers price risk based on what the client or broker tells them, and about half the time at least one of those records is wrong. The result is mispricing, fraud, and an estimated $36B in premium leakage every year in the US alone, which gets exposed the moment a disaster hits. Upload a list of addresses and their GeoAI reads satellite and street-level imagery, maps, real-time data, public records and the insurer's own data to determine what's really there, which natural and human hazards surround it, what changed over time, and which properties are worth inspecting. Each property becomes a living record that underwriting, claims and pricing teams can trace from past to present. Ageospatial deploys inside carriers and MGAs as an AI-native service, so insurers get outcomes delivered.

Cyclon is rebuilding the phone around AI. Instead of opening apps and tapping through screens, you tell your phone what you want and it does it for you. With one touch, you can spin up AI agents to control your phone or handle tasks in the background. AI isn’t just inside an app - it’s the phone itself.

Grep AI is the fastest way to create, optimize, and deploy agents for tedious, complex operational work that slows down your business. Our self-improving agents learn from every run, making them smarter and cheaper as you scale. IG.com a global retail brokerage used our agents to review compliance alerts and now 80% of their customers onboad without manual intervention, doubling first trades and propelling them to a FTSE 100 company. Customers including Airwallex, Worldpay, Citizens Bank, FV Bank and Shopmonkey have used our agents to automate over 250,000 of operational work across compliance, onboarding, underwriting, corporate due diligence, document verification, merchant categorization, market and competitive research, rev ops and much more. We founded Grep AI after experiencing firsthand how manual workflows can prevent a business from growing when we ran the platform team at Brex. We're now on a mission to rid businesses of tedious, repetitive so their employees can focus on the most high leverage tasks that matter.

Decent is a health insurance brokerage for small businesses. We compare every way to cover your team, show what each one costs, and then keep improving your plan all year. We're paid by insurance companies, so we're free to you.

Viraj Aero builds gas turbines that deliver both power and cooling water for data centers. It needs no grid connection and no water pipeline. Burning natural gas already produces water. Our patented system captures that water from the exhaust. Part of it cools the data center. The rest is turned into steam and injected back into the turbine. That gives up to 40% more power from the same fuel (cheaper compute) and cuts NOx emissions by up to 99%, which cuts permit approvals from years to weeks. We're building a genset around our first-of-its-kind gas turbine. It fits in a 10-foot container and runs anywhere, even in the desert. That matters because two-thirds of new data centers are being built in water-stressed regions. We first built this steam-injected turbine for aircraft, where carrying water tanks isn't an option and efficiency is critical. We're currently integrating our full-scale prototype into an aircraft. The team comes from GE, Rolls-Royce and Formula 1, and we're already working with Safran, one of the world's largest aircraft engine manufacturers.

Faraday builds industrial AI that learns how enterprises operate and executes long-horizon, multi-step tasks from the back office to the shop floor. We sit on top of fragmented systems of record, coordinating work across existing software and teams. We combine operating procedures, enterprise data, and expert feedback to capture the rules and judgment behind complex processes. In customer-specific RL environments, we optimize the agent harness around each company’s systems, constraints, approvals, and exceptions. We’re connecting to industrial hardware to bring live production data into the same feedback loop. Human corrections, machine data, and operating results help improve how agents plan and execute future work.

Turning ideas into trades. Every great investment starts as an idea about the world. Institutions have teams of analysts turning those ideas into positions. Volaren is that team, processing data through fundamentals and quantitative models that can be personalized with AI. Turn your views into trades, portfolios, and back-testable frameworks and strategies, and Volaren stays with you from there, showing you what's actually driving each position and what it means for your next move.

Agent Relay lets software teams' AI coding agents message each other directly, so people stop copy/pasting between Claude Code, Codex, Cursor, and their teammates' agents. A frontend agent can ask the backend agent what an API returns, two agents can sort out a merge conflict between themselves, and your agent can answer a colleague's agent while you sleep. Agents connect wherever they run: terminals, desktop apps, phones, laptops, cloud boxes, or a Mac mini across the world. Every session goes into shared, searchable history, so decisions don't get stuck in one person's chat. You can even let a customer's agent debug with yours. Setup takes about 5 minutes: give your agent the link and it installs itself. Use our hosted version free, or self-host the open-source version

The next constraint is not intelligence, it is the energy and operational cost of running it. We deploy custom AI models inside critical infrastructure companies that ingest SCADA data, work orders, human interventions, operational data no public model has seen, grounded in physics-informed neural networks and differentiable simulators. This creates a flywheel: more operator data → better process understanding → more efficient operations → cheaper energy → cheaper AI → more deployment → more data. https://www.youtube.com/watch?v=0uDH6pMn5OE AI compute doubles every 2.2 years. Data center demand grows from 415 to 945 TWh by 2030. Without near-zero energy cost, AI remains limited by human infrastructure. We are the layer that runs AI at zero energy cost. Energy is a $10 trillion market. The infrastructure layer doesn't exist yet. We are building it.

Applied Kinetics builds and deploys specialized AI employees to handle the coordination work behind America's energy and data center buildout. Our agents chase suppliers, track equipment approvals and delivery dates, and flag delays so project teams can keep critical equipment moving without hunting for updates across emails and spreadsheets.

Autonomous accounting platform for global companies. We close your books and give you financial clarity while you run your business.