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.

Overshoot makes it easy for developers to build and run real-time vision applications. AI can see and understand the physical world. This unlocks new applications in physical security, safety, gaming, robotics and general consumer products. Soon, video agents will watch your home and your pet when you’re away. However, existing platforms make it painful for developers to build these real-time applications. Slow inference. Limited model availability. Break at scale. Overshoot solves this. Today, 300+ developers connect live video feeds to the largest collection of Vision Language Models with 3 lines of code. And get responses in less than 200ms. 10x faster than any existing inference platform. Zero infra headache. Our moat is focus. Image and Video are fundamentally different modalities from text. By focusing on them, we are able to make strong technical leaps from codec, streaming protocols to inference engines. Younes and Zakaria are cousins. Zakaria graduated top of his class at LSE and MIT, and later built low-latency high-throughput pricing systems (surge) at Uber and inference engines at Meta. He previously built and sold a software product and won several prominent AI hackathons. Younes was a founding engineer at Cosmonio (later acquired by Intel). He built a training and serving Computer Vision platform from scratch. He witnessed firsthand customers abandoning traditional Computer Vision because it lacked the “general” intelligence LLMs have today. Together we’ve shipped large-scale systems and know where they break.

Agent-first development environment. Emdash enables engineering teams to develop software faster, by running multiple coding agents in parallel. We're an open-source and provider-agnostic desktop application.

We are building a "manipulate anything" model with the robot hardware that embodies it. We designed a hand-based data-collection device and a high DOF, direct drive robotic hand that match each other exactly, enabling us to eliminate embodiment gap and collect real-world data and deploy it directly. We want to scale real world manipulation data by deploying our devices "in the wild" like manufacturing factories, logistic centers to collect Tesla like data, use these data to train our model and then provide automation solutions to these industries.

Beacon Health (YC W26) is building AI agents for primary care - starting with value-based care. Primary care is the primary source of care for 200M Americans but physicians don’t have the bandwidth to manage their patient panels - preventative screenings, prior authorizations, referrals, and risk adjustment, all important work that steals time from patients. Our Al agents handle workflows for them directly in their EHR. We own them from end to end - putting their value based care on autopilot. We’re on a mission is to bring joy back to primary care.

Robby is an AI growth engine for home services businesses that uncovers revenue opportunities and collects additional data from every technician visit. Our AI combines customer and third party data to generate dozens of new leads every day, and then arms the technicians with talking points to increase close rates, giving GMs a single command center to manage and drive revenue. We have several paying customers that are each uncovering 6 figures in additional revenue opportunities every week. On our first day of launch, we helped close $40k in new business within an hour. Our team has spent the last few months riding along with dozens of technicians and shadowing every position in the office to best understand the industry. We're also studying for the EPA 608 exam to become certified technicians ourselves. Vineet and Feroze have both built critical products and infra at prominent startups (Ironclad, Porter, Affinity, Uber, Cloud Health Systems). Joe previously consulted PE firms on software acquisitions and led home services growth projects at Bain. We all met at Harvard Business School, and dropped out to work on Robby full-time.

Unisson is building trusted AI subject-matter experts for teams that are responsible for customer deployments (technical customer success, implementation, sales engineering). Within 20 minutes, our agents learn to use any product and can be used by teams in Slack, text, or email. Our agents gather requirements and context from existing tools enabling them to handle complex tasks like customer health audits, onboarding, product administration, and custom integrations. With Unisson, customer-facing teams have 24/7 access to an AI SME, allowing them to scale and save thousands of hours every month.

Your agent processed a refund without looking up the order ID, costing you thousands. You only found out three hours later from a support ticket. Evals, output scoring, and observability can reduce the likelihood of mistakes like these occurring - but there's no solution that inspects and prevents an action as it’s about to execute. Salus does that. We’ve built an API that wraps around your agent and checks its actions at run time, blocking incorrect ones and providing immediate feedback to guide retries. Kevin and Vedant were roommates at Stanford, where they both studied computer science.

BeeSafe AI is a fraud-prevention platform for enterprises that helps them protect their customers against trust-based attacks like "pig butchering" and impersonation. By engaging fraudsters at the source in real-time, the AI system identifies and shuts down channels by extracting deterministic data on mule accounts and attacker infrastructure. Our intel has already enabled financial services and government agencies to intercept scammers mid-transaction and prevent victim losses.

Autumn is building the first real-time signal intelligence platform for GTM teams. We monitor posts, commits, blogs, and announcements, surfacing buying signals the moment they appear. Define your ICP and the signals that matter, and we deliver a condensed, real-time feed filtered by intents.

Lucent automatically watches your user sessions and alerts you when users hit bugs or UX issues. You're recording thousands of sessions but nobody's watching them. Lucent does - catching silent issues breaking your product that you'd otherwise never know about.

Maywood automates deal execution for investment banks, from decks to diligence The Maywood platform automatically generates finished, on-brand CIMs / presentations, models, and diligence responses that prioritizes editability and auditability, with best-in-class security ensuring information stays protected. Maywood reinvents the deal process with a single, living workspace for every deal.

Vela is the AI Scheduling Assistant built to handle the ambiguity that breaks scheduling. It has taste: it knows to prioritize a client over an internal catch-up, and it understands "Let's do early next week" without setting specific parameters. You just add Vela to the conversation, and it handles the white-glove negotiation as a member of your team, even at scale. Today, Vela is already used by professionals with the most complicated scheduling workflows. Enterprise recruiters are using Vela to schedule thousands of multi-party interviews and large sales teams rely on Vela to coordinate cross-functional calls.

Samora runs multilingual voice operations for organizations with high-volume calling without requiring engineering teams, call centers, or AI expertise. Many important tasks still depend on phone calls, including customer support, outreach, follow-ups, screening, and intake. But voice AI often fails in production because real callers interrupt, go off-script, switch languages mid-sentence, and speak in local dialects. Teams also worry about safety and compliance, especially when an agent can accidentally make the wrong promise. Samora solves this with voice agents that sound local and follow strict rules. Samora agents handle inbound and outbound calls, collect information, and update internal systems like CRMs, ticketing tools, and ATS platforms. When a call becomes sensitive or uncertain, the agent hands off to a human. It can also follow up over WhatsApp, SMS, or email when needed. With Samora, teams get: 1. Voice agents that work in real conversations, including dialects and code-switching 2. Strict safety controls so agents don’t make unauthorized offers, discounts, or promises 3. Calls that result in completed actions, system updates, and safe human handoff

Sentrial is the observability platform for AI agents in production. When your agent loops indefinitely, hallucinates, or frustrates users, Sentrial detects it in real-time, diagnoses the root cause, and recommends specific fixes. We provide the monitoring, evaluation, and debugging infrastructure that traditional tools weren't built for, so teams can ship agents with confidence and iterate based on real production behavior, not just evals/benchmarks.

Stilta is the AI-native software for patent practitioners, built on agentic workflows to amplify how IP work is done end-to-end. Instead of spending weeks manually researching and analyzing patents, practitioners can reduce manual work by up to ~60% while unlocking new revenue opportunities. Innovation is key to solving the world’s most important challenges, but the tools to protect innovation haven’t kept up. Patent teams are stuck doing high-stakes work with a process built around manual document review and disconnected point solutions. Existing tools are rigid and fail by not integrating AI with clear traceability (every output must be source-backed, referenced, and auditable). General LLM solutions generate text, but they don’t solve trust and integration in legal-grade workflows. We’re four engineers from McKinsey/QuantumBlack who’ve spent the last few years building and deploying AI for large organizations. Now we are building Stilta because the practitioners protecting the world’s most valuable ideas deserve tools that match the stakes. We’re already live with leading firms in both Europe and the US.

VOYGR helps AI apps and agents understand and act in the real world with comprehensive place data. Unlike static mapping APIs, VOYGR combines accurate place intelligence with fresh web context – news, articles, and events. Real-world interactions rely on maps: 10M+ apps/websites, up to 40% of search queries, and 20% of LLM prompts need local context. Yet a restaurant recommendation can fail because the place is closed, menus aren’t searchable, or delivery goes to the wrong entrance. AI pushes expectations further – answering semantic prompts like “specialty coffee shops in SF with Wi-Fi and YC founders", and taking actions like making reservations or placing orders. We’re experienced builders in maps, search, and ML to solve this. Vlad knows this space inside and out – he worked on the Google Maps APIs GTM, plus firsthand customer experience from building in ridesharing and travel. Yarik has spent over a decade leading ML/search teams at Apple, Google, and Meta. Several customers already run VOYGR to continuously validate places data accuracy at scale, and we outperform on accuracy and coverage. Maps were built for humans browsing pins; agents need continuously updated place ground truth they can act on – from copilots and in-car assistants to delivery robots and AR. As agents become the interface to commerce and services, VOYGR aims to be the default ground-truth layer.

Tsenta is a desktop app that matches you with jobs, optimizes your profile & applies on your behalf. It runs on your device, so you have granular control over every application being submitted in real-time. We built Tsenta after applying to over 3,000 jobs manually, so that nobody else would have to. Check us out at https://tsenta.com :)

FullSeam is an AI employee for finance and accounting teams. It logs into a company's accounting, billing, banking, and CRM tools and completes routine tasks typically done manually — like accounts receivable, accounts payable, billing, and updating financial records.

Ruma Care automates prior authorizations - we’re starting with clinics that administer the most expensive specialty medications. Biologics are expensive medications that cost $80-$150K+ per year per patient, and often need to be administered in specific settings called infusion clinics. Medical providers buy these drugs up front to secure pricing deals and safer inventory - they are reimbursed for the cost of the medications through the patients’ insurance. Because they’re so expensive, these specialty medications are denied by insurers at a rate of 37% each year. When they are denied, providers don’t get reimbursed for the medication and are left covering the costs. Ruma Care solves this so clinics get paid, patients get treated faster, and care teams spend less time fighting insurance. We do this in three ways: 1. For each medication, diagnosis, and insurance combination, Ruma extracts the submission criteria required to get the drug approved by insurers. 2. We pull out the correct forms, then automate the form-filling for prior authorizations. 3. Lastly, we enroll patients in copay assistance programs, all in one workflow and platform. We turn a process that traditionally spans 70+ online portals, disparate paper forms, and manual phone calling into a single, streamlined platform. As we process prior authorizations, our models learn from denial patterns to understand exactly what insurers need to see to get a prior authorization approved.

Oximy helps enterprises understand and manage AI usage across their workforce. Used by security, finance, and transformation teams to track adoption, control spend, and manage governance as AI scales.

APIs weren’t built for agents. Orthogonal fixes that. Developers and agents get instant access to hundreds of APIs through our MCP or SDK. No API key management, no billing headaches, pay as you go. API providers list once and become instantly discoverable.

We help highly regulated financial institutions and fintechs prevent financial crimes. With our AI agents, you can process 10X compliance workload per analyst, with bullet-proof audit logs for regulators!

Strand AI builds foundation models that predict missing biological data from measured readouts.

You’re wasting millions of dollars on custom industry robots that take months to implement. Today, robotic automation means redesigning your entire facility for the robot to work. You need to change the floor plan, adjust conveyor belts, and re-educate your staff. This is inefficient and unnecessary. With Servo7, you don’t have to do any of that. Automating with Servo7 is simple and fast, because our robots: 1) Deploy in existing operations 2) Learn from simple demonstrations 3) Keep improving on the job We have previously worked on autonomous defense systems, LLMs, autonomous driving, and hyperloops. Not just in the lab, but actually deploying them in the field. We’re already working with warehouses to automate their fulfillment. If you work in assembly, manufacturing, or logistics, email founders@servo7.com. We're onboarding our next partners now.

AI for inventory starting with libraries. Librar speed up the manual & boring work, so librarians can build reading culture and community. A complete library system for any collection.

Supply chains still run on spreadsheets and inboxes because ERPs are rigid and the data is messy. We're solving this problem by building an ontology where agents can operate and provide outcomes for our customers. We start with procurement, by automating three-way match: extracting supplier invoices from email and matching them to the correct PO and receipt across ERPs. Learn more at pollinate.tech

Mendral is an AI DevOps Engineer that observes, diagnoses, acts, and learns continuously. It autonomously handles CI failures, flaky tests, slow builds, broken releases, and code reviews, improving with every build and run. Over time, Mendral evolves into a self-improving delivery system that manages performance, quality, security, and compliance, all without manual intervention.

Moda alerts you when things go wrong with your AI agents.

There are 2 problems: The first, CAD software is slow and the second, designing with it is a tedious process with many steps. Aurorin solves both of these problems by being a CAD software built from the ground up on a modern tech stack with AI at its heart. Building the CAD kernel lets us fully leverage AI while modernizing execution speed. Almost all existing CAD software (Solidworks, NX, Creo, Catia, Fusion 360, Onshape etc) use the same couple of underlying CAD kernels built in the 80's. It's common for these programs to take 4 hours to open a file. Aurorin uses a custom parametric and B-Rep driven CAD Kernel built to take full advantage of modern CPUs and GPUs and be AI native.

MouseCat enables companies to automatically detect, investigate, and mitigate emerging fraud trends

ARC Prize builds AI benchmarks that measure general intelligence. Our benchmark, ARC-AGI, has been used by OpenAI, Anthropic, Google DeepMind, and xAI. Founded by Mike Knoop and Francois Chollet, we inspire open source artificial general intelligence (AGI) research through benchmarks (the ARC-AGI series), global competitions, research grants, community, and content, we exist to guide researchers, industry, and regulators on the path to AGI. We believe that AGI requires more than just scaling up existing AI models. It demands a fundamental shift towards systems capable of genuine fluid intelligence, the ability to adapt to novel challenges and solve problems efficiently, much like humans do.

Polymath is an applied research lab focused on increasing the reliability and autonomy of AI agents, in order to better serve people. We believe that the key to getting models to perform reliably across long horizons is the quality, complexity, and abundance of RL environments. We're building automated environment factories to produce high-fidelity training and evaluation environments at scale, providing frontier labs with the building blocks they need to make the next breakthrough.

CodeWisp lets anyone build and publish web games using plain-English prompts. Traditional game development requires learning complex engines, managing assets and code, and following rigid workflows, making it slow and inaccessible for most people. With CodeWisp, users describe the game they want. CodeWisp generates the game’s code, structure, and assets, and users can refine it with follow-up prompts. This makes it possible to go from idea to a playable web game in minutes, without prior game-development experience.

Balance is an AI accountancy helping SMBs close their monthly books. Our agent work as an in-house finance hire, chasing unpaid invoices, preparing monthly reports, and flagging issues before they worsen.

Crow lets users control your app through chat. Connect Crow's AI agent to your product, and users can type what they want instead of clicking through menus. Crow connects to your APIs and data sources, navigates your UI, and executes real actions. Set guardrails, define workflows, track every interaction, and deploy in under a week—no rebuilding required.

Compresr provides an API that compresses LLM context without losing what matters. It’s a drop-in for agents and RAG that cuts token costs and improves accuracy.

EigenPal.com automates your document processing workflows with AI. KYC docs, invoices, claims, shipping forms, contracts, receipts, loan docs - we automate that. Works on messy scans, handwritten forms, third-party docs. Powerful workflow builder and team functionality. On-prem ready or use our cloud.

The best grid connection is no grid connection. Voxel data centers capture, store, and consume solar energy onsite without the grid, completely bypassing utility connection delays. By leveraging the vast supply of second-life EV batteries with our novel DC microgrid architecture, we drastically reduce engineering overhead and time-to-power.

Perfectly is the first ever end-to-end AI recruiting agency. We learn your deep hiring preferences via our intake, then we source and outreach to thousands of candidates that fit exactly what you're looking for. Then candidates conduct an AI screen that can be reused for future jobs, and are delivered straight to the hiring manager's portal for interviews. Recruiting is broken because information between candidates, hiring managers, and recruiters is lost at every step. We realized the opposite could be true: The more you tell us, the more accurately our algorithm can match you. That's how we build trust.

Shofo builds large-scale social media training datasets by collecting, labeling, and enriching public content for pre-training and fine-tuning AI models. Our indexes continuously collect and update every video across hard to access social media platforms. Companies use Shofo to avoid building their own data collection and processing systems and spending months turning unstructured social content into model-ready data. We started with TikTok and are expanding across every major social platform.

Q2Q is the AI-native outbound engine for deal sourcing. Tell us the types of companies you want to buy, and we'll get qualified meetings booked for you. You can think of Q2Q as your personal team of AI interns. You give us your buying criteria, and we'll go out and find companies, write personalized emails, and get first calls booked with business owners for you. As the human in the loop, you'll approve messages before we send. Q2Q handles the sourcing work so you can focus on evaluating great deals.

Scout Out helps residential contractors estimate jobs faster and submit clearer, more competitive proposals. Today, every contractor bidding a residential remodel or rebuild has to manually interpret plans, calculate material and labor quantities, and assemble a proposal from scratch. The process is slow, inconsistent, and expensive — and small differences in scope often decide who wins the job. ScoutOut automates this workflow. We use computer vision trained on residential building plans to generate structured material and labor takeoffs, which contractors can review, customize, price, and turn into professional proposals. Instead of starting from zero on every bid, contractors start with a standardized, high-quality estimate they can trust. The result: contractors bid faster, price more consistently, and submit clearer proposals — helping them win more residential remodeling and contracting jobs with less overhead. We’re starting in Los Angeles, working closely with residential contractors on remodels and post-fire rebuilds following the Palisades and Eaton fires.

Opalite makes it easy for healthcare providers to communicate with non-English speaking patients. We’re built to work inside existing clinical workflows, so providers can communicate in real time without waiting for an interpreter or delaying care. Opalite is available instantly, 24/7, and replaces traditional interpretation services with software that’s more reliable and saves millions of dollars per year. We've built a voice AI system that is the safest, most compliant, and most accurate medical interpreter in the world.

Our mission at 10x Science is to enable biopharmaceutical development to keep pace with AI-powered drug discovery. Thanks to AI, the front end of drug discovery is flooding the pipeline with new drug candidates. However, current solutions for drug development are unable to resolve drug complexity and cannot keep up with modern demand. Our platform is the unlock for this bottleneck. By applying modern AI to the development stage, specifically to protein characterization and quality assessment, we provide an end-to-end solution to dramatically speed up and scale up the preparation of candidates for the clinic. Our team is composed of Stanford University researchers from a Nobel Laureate's lab with 18+ years of collective domain expertise, 47+ scientific publications, and a 2x YC Founder. We move at the highest velocity and are positioned to lead the next-generation of modern AI drug development.

Current multimodal models can see and hear. But they talk when they shouldn't. They can't tell if you're speaking to them or someone else. We are building an AI that knows when to stay silent, yet still understanding what's going on in your conversation, so it can best assist when you do need it in real time. Our first version, fern-0.1: provides real-time expert opinions on demand, instant task delegation, zero interruptions. All as fast as ChatGPT voice and Gemini live.

Bizzy handles your calls, captures and qualifies leads, and books appointments directly into your CRM. It understands your services better than a receptionist would, routes and escalates calls, and syncs with your existing tools to incorporate complex workflows, all so your team can focus on servicing leads, not qualifying them.

Verdex uses satellite imagery and weather data to create a persistent, time-anchored record of what happens on every acre. This turns land into a measurable asset whose outcomes can be audited objectively at scale. Verdex is built to replace manual field inspections in insurance verification and risk management.