Generative AI Startups funded by Y Combinator (YC) 2026

The next constraint is not intelligence, it is the energy cost of running it. We deploy custom AI models inside energy companies that ingest SCADA data, work orders, and grid topology, operational data no public model has seen, grounded in physics-informed neural networks and differentiable grid 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.

ByteAsk is an AI coding agent built for C and C++, the code that runs trading systems, cars, chips, and operating systems. It does not stop at a diff that looks right. ByteAsk builds, debugs, and tests every change with the real toolchain until the change holds. If the compiler, the debugger, or the tests can catch it, then ByteAsk catches it too. It runs from the terminal or inside VS Code, Neovim, and Emacs, and is available fully on-prem.

Marker is a platform for rebuilding enterprises agent-first, delivered by forward-deployed engineers. It productizes every step from discovery to production, so every agent we ship makes the next one faster.

We’re an AI-native biotech, working on replacing needles with pills. Many drugs, like insulin or Ozempic, are wonderful except you have to inject them. We discover new molecules that you can take as a pill instead. Because the pills operate on biology which has been shown to be effective in people, rather than merely in animal models of disease, we avoid what has historically been the single biggest risk in drug discovery.  This should dramatically increase our probability of success. There’s $200B of treatable diseases and validated human biology without a pill because discovery programs lack a tractable chemical starting point. This is a zero-shot inference problem, because these targets have no training data, so LLMs and other historical interpolative AI approaches are insufficient. Instead, we are building our own extrapolative frontier models that reason beyond their training data. Most critically, we have repeatedly validated our platform on both our own and partner programs. On 4 biologically- and structurally-diverse targets where we had zero training data, we successfully picked binders that were validated experimentally, with confirmation ranging from primary screen hits to crystallographic structures.

Familiar is the Neolab for world translation. Dubbing into 30 languages. Now with Familiar, the same actor's performance is preserved instead of a mismatched voice over a mismatched face. One model does the voice, the lips, and the scene together, with identity permanence and scene & world understanding. Built by the researchers who pioneered joint audio-visual human models.

We deploy robots that learn new tasks from a single human demonstration in the factory. We're starting in contract manufacturing, where lines change every week and both traditional automation and today's VLAs fall over. Our founding team has 20+ years of experience in robotics. Navid was part of the core team that productized Spot at Boston Dynamics and led manipulation projects there. Before that, he was at Rethink Robotics. Mojtaba led whole-body control and manipulation efforts at Amazon Robotics and was a Research Assistant Professor at Northwestern University.

Almanac is an AI agent that deeply understands your work. Today people rely on a stack of productivity tools to hold the context of their work including todo lists, issue trackers, wikis, and CRMs. All of them have to be updated by hand, and none were built for AI. Almanac maintains its own self-updating productivity suite for you. This gives it an absurdly good understanding of your work, based on which it proactively gets things done for you. Almanac starts with individuals, but it's built for teams. As more of a team connects their work, the agent gains shared context across everyone, so it understands not just what you're doing but how the whole team's work fits together.

Illume is a 24/7 AI personal health & longevity companion that connects your wearables, blood panels, and genomic data to text you personalized and actionable insights. We pull in your health data and fuse it into one secure model. You get daily text updates on what changed, what it means, and what to do next; conversational health Q&A like "What should I do to improve my bloodwork?"; cross-data trend detection that catches patterns individual platforms miss; and a full platform to dig deeper, explore your health data, and text your Illume agent anytime. We're 3 MIT grads & friends, with experience at Windsurf, MIT CSAIL, Eli Lilly, & the Broad Institute.

Palette is a self-improving AI-native media platform built by MIT computer scientists and multimodal researchers that unifies the creative workflow into a single generation engine. We’re building a fully autonomous production engine that generates, distributes, and optimizes interactive media for every brand and audience.

Glen is the layer that unifies every agent your company runs. Glen pulls in agent sessions from Claude Code, Codex and Cursor, plus Slack messages, GitHub prs, tickets, docs and calls. When a new agent picks up a task, Glen finds relevant prior context and injects it in the agents context window to steer it, producing knowledge work that is significantly higher quality, faster and cheaper. Glen does more. You can query the original agent behind any line of code, or search every session your org has ever run. It distills what your team does into skills and surfaces the right one automatically. Agents work multiplayer, picking up the style and context of whoever else is on the file. Transcripts are unified across Claude Code, Codex and Cursor, so you can hand a session from one to another mid-task. And Glen can also double as a company brain, from Slack, in app, or by just installing Glen on any agent. We’re building toward a future where memory is no longer the missing piece of intelligence. A future where every agent maintains a lossless and continuous thread of institutional experience enabling the era of the compounding organization and fully unsupervised digital labour.

Ashr Manifold is a self-contained model training, hosting, observability, and continual learning platform for purpose-built open-weight models. Startups can use Manifold to post-train and manage fleets of continuously improving open-weight models without having a dedicated research team or infrastructure. Ultimately, this reduces operating costs and allows a greater degree of control over proprietary company knowledge. Enterprises can use Manifold to encode institutional judgements and priorities into open-weight models, reduce token costs, and manage every level of the AI life cycle, from data curation, to model serving on weights and infrastructure they own.

Ritivel is an AI-native workspace for regulatory and medical writing teams at life sciences companies. Our AI agents draft CTDs, CSRs, INDs, and BLAs in minutes instead of weeks, helping drugs reach patients faster. Every month a drug is delayed from reaching the market costs pharma companies ~$45M in lost revenue and more importantly, delays treatment for thousands of patients who need it. Yet regulatory teams are stuck in a painful, months-long process: synthesizing thousands of pages of clinical data, hunting through SharePoint folders for the right documents, chasing colleagues for missing inputs, and manually drafting submissions. Ritivel deploys AI agents that transform how regulatory teams work. Our agents generate comprehensive first drafts directly in Microsoft Word, following FDA formatting requirements and pulling from your existing clinical data and prior submissions. Every data point includes citations linked back to the original source, so reviewers can verify with one click. Connected to SharePoint and Veeva, our agents automatically source the right documents. Integrated with Outlook, they handle reminder emails and follow-ups automatically. We're engineers who previously built AI copilots at Microsoft Research. After 50+ conversations with pharma professionals, we discovered the critical bottleneck in regulatory submissions and built Ritivel to solve it.

Anto is building a foundation model for microbial communities, making the gut microbiome computable for the first time. We predict drug toxicity and efficacy across diverse populations and fix drugs so they work for everyone — solving the hidden root cause of most drug failures. Founded by Arvid (Broad Institute of MIT and Harvard; Nature-published researcher who pioneered quality-aware, goal-directed sparsification) and David (Harvard Medical School Gastroenterology, J&J), second-time founders who published the breakthrough at leading machine learning AI conferences and Nature.

World model datasets and research

Expected Parrot helps companies simulate their customers with AI agents to explore pricing, product, marketing, communications and other scenarios at scale. Our open-source library and no-code web app let you design custom agent personas and conduct interviews and surveys with them using LLMs of your choice to generate results and reports. Everything is cached and reproducible, and you can send the same surveys to human respondents to validate results in the same interface.

Absurd is an online whiteboard for filmmakers working with AI. We use it for our own work: our Kalshi spot "Election Day" did 1M+ views, and our videos average 400K+ organic views. Try it at absurd.com

LunaBill is building AI Voice callers for healthcare billing teams. Insurance claim follow-up calls account for the 80% of the daily workload of a healthcare billing team, with each call averaging 30 minutes. Since our launch in July we are now at $764K in contracted ARR, of which $428K is live revenue and the rest will go live by January 2026. Our customers see a 10x increase in claims followed up per biller in the first week and we have converted 100% of pilots to paying customers. We have automated 50,000+ calls and signed partnerships with top healthcare systems in the country. The founders have known each other for over a year. Suhail is a 2x founder, with 3+ years of experience building in healthcare including operating a medical billing company with 200+ healthcare clients. David founded Skiesoft, Taiwan’s biggest healthcare AI scribe company, used by thousands of healthcare professionals. He also worked as a data scientist at UCSF, building models to audit medical-coding accuracy.

Pixley turns kids' drawings and ideas into fully animated cartoons – in minutes. Kids are natural storytellers. With Pixley, any kid can draw a character, describe a world, and watch their story come to life as a personalized animated cartoon. Most kids' screen time is broken: passive, algorithmic content designed to keep kids watching, not growing. Pixley is the antidote. Every episode is safe, ad-free, and built around Social-Emotional Learning (SEL). Parents can choose any lesson they want to teach their child, and Pixley weaves it naturally into the story their kids create. We're building the creative platform for the next generation – where kids are the directors, not just the audience. Additionally, the same technology powers Pixley Studios, our B2B content service for brands and content creators who want to produce high-quality AI video without the cost or time of traditional production. We’re 19 year old CS dropouts who’ve known each other since elementary school and won 4 hackathons at 15. After seeing AI flood the internet with low-quality content, we created Pixley – a safe, personalized way for kids to turn their ideas into animated stories.

Koyal is an Agentic AI Filmmaking Platform that allows you to seamlessly convert script or audio into personalized, visually compelling stories. We’re a team from Carnegie Mellon, MIT and Meta, with experience in AI video generation. https://koyal.ai builds directly upon our published research at the AI conference NeurIPS in December. We’ve been covered in the international press for our patented secure personalization protocol CHARCHA: the genAI captcha. Announcement: https://x.com/meh_agarwal/status/1965866965425348687 Demo: https://x.com/meh_agarwal/status/1965137062119752099 We currently have paid pilots with the biggest global music labels, podcasts and production houses, and just launched our public beta at beta.koyal.ai We recently concluded a pilot with Universal Music and T-Series (world’s largest music label) alongside Bollywood production houses like Maddock Entertainment, and Collective Artists Network! As part of that, we launched 5 music videos for Grammy and Oscar winning artists like A R Rahman, Ricky Kej and Shankar Mahadevan all crossing 1.5 million views! We've recently built SOTA capabilities in (i) multilingual script to audio to video, (ii) Agentic lip-sync with dynamic movement & (iii) Multi-character consistency.

Specific builds enterprise environments and datasets representative of work inside real-world businesses. The next frontier of AGI is bottlenecked by data that is hard to access, and we make that easier.

Bezel helps fashion brands create virtual photo/video shoots with AI. Upload images of clothes, select the human you want to model it, and Bezel generates pictures and videos that rival a full studio production. Every detail of the clothing is rendered flawlessly. Try it for yourself.

Automatic insights on all your company data. Speed up business workflows to visualizations with a 0 learning curve.

Knowlify makes it easy for companies to create animated explainer videos using AI. Upload any document and/or prompt deck to get a professional explainer video in minutes - no designers, animators, or video editors needed. Teams use Knowlify to produce training videos, customer literacy videos, product explainers, onboarding content, and internal communications at a fraction of the traditional cost and timeline. Available as a self-serve platform or a fully managed service.

Doe is where AI agents actually do your work. Connect your tools, tell Doe what you need, and agents handle it. When you want to do something yourself, Doe makes you faster and smarter. The boring stuff disappears. Everything else gets better.

Kashikoi is a simulation engine to benchmark AI agents. We generate CPU friendly world models that autonomously interview agents and generate deep behavioral assessments. We built a similar technology at Moveworks which was used to ship 250+ enterprise agents to customers daily.

Paper Instruments post-trains LLMs to accelerate knowledge work.

We build autonomous research agents that accelerate frontier world models.

Cohesive is the agentic CRM for blue collar businesses like janitorial, pressure washing, landscaping, HVAC, and pest control businesses. Cohesive fully automates sales and marketing outreach to local businesses and property owners in the service area, helping local service businesses grow even while servicing clients in the field.

Auctor brings alignment and intelligence to software implementation, enabling every project to deliver its intended value with clarity and precision.

.Wave builds a continuous inference engine for streaming AI models, including full-duplex voice and streaming speech recognition. Customers bring their models; we integrate and run them, managing execution, session state, and scheduling to reduce end-to-end latency and serving cost per live session. We use cadenced inference to support more concurrent sessions per GPU while keeping latency low and consistent.

AI creative tools became powerful quickly. But they also got complicated. Most of them require creatives to do the engineering. Prompts. Parameters. Node graphs. Creative work doesn’t start there. It starts with thinking, exploring directions, and iterating on ideas. With Awen, we’re trying a different approach: A creative partner that turns intent into exceptional visuals through conversation.

Tergle builds AI to replace Auditors.

Dexto makes it easy for developers to integrate AI agents into any software application. We do this by exposing agents as simple APIs that can be called from anywhere. Developers have used Dexto to add customer support agents into WhatsApp and analytics agents into Slack.

Emergent builds autonomous coding agents that replace traditional software development by generating, testing, and deploying production applications directly from plain-language intent. Our systems run in production at global scale and are used to build millions of real applications. Since our public launch, Emergent has reached $50M ARR in 7 months making us one of the fastest-growing startups in the world. 5M+ users across 190+ countries have built 6M+ applications on Emergent. We've raised $100M from Khosla Ventures, SoftBank, Google, Lightspeed India, Prosus Ventures, Together Fund, and Y Combinator.

assistant-ui helps frontend developers add AI chat to their apps. Our product is used by several dozen companies, like LangChain, Stack AI, Browser Use, and Athena Intelligence. We're building an open source Typescript/React library and a backend as a service, with first class integrations into popular agent frameworks like LangGraph and Vercel AI SDK.

TrainLoop makes it effortless for developers to supercharge LLM performance through reinforcement learning.

Wildcard is the full-stack AEO & GEO platform for ecommerce and retail brands. We audit and action on how you show up across across AI search and AI shopping platforms like ChatGPT, Google AI Overviews, Gemini, Amazon Rufus/Alexa, and Walmart Sparky. We build the roadmap to close the gaps, and create the content that gets you recommended. Most tools stop at a visibility score. We do the work, and are built for ecommerce brands and retailers, not SaaS. Brands using Wildcard now show up first in AI answers where they used to be invisible. Wildcard is building the command center for agentic commerce, everything brands need to serve a new type of shopper: AI agents; including support for emerging standards like the Agentic Commerce Protocol (ACP) and Universal Commerce Protocol (UCP).

Calltree makes enterprise-grade AI support reps for call centers. Our AI reps take voice calls and chats like other AI support agents but can also use internal enterprise tools like a homegrown CRM or internal dashboards, like the highly trained humans working there today.

Trava is building AI agents for global trade compliance. Importers lose billions each year to duty overpayments and filing errors. Trava audits customs entries and automates classification to help importers protect margins and stay compliant at scale.

Sennu AI builds AI systems that fundamentally transform how your business operates

City building departments are understaffed and overworked. Meanwhile, builders lobby to remove power from building departments. Because building departments are usually the only profitable/self-sustaining branch of a local government, this is problematic. Permitify allows city build-plan reviewers to 10x their productivity covering the labor shortage, reducing the time required to issue a building permit, and allowing the city government's only profitable branch to stay intact.

a0.dev is a platform that helps mobile developers build high quality apps in days instead of weeks by generating custom react native projects. We've launched top ranking apps that have been used by hundreds of thousands, and now we're creating the tools we wish we had to build faster.

Alice turns generic course materials into personalized learning and exam prep using AI. We’re Duolingo for exams, and more than 10k college students have already used Alice. Kim, the CEO, is a domain expert in education with more than 10+ years of university teaching experience. He is also technical by background. MRR is $20k, growing +10% week-over-week the last 8 weeks Alice's core market opportunity is +$10b, with the global education market being almost $10trillion. As AI starts to disrupt education, Alice will use in-built network effects and virality in its product to create a global winner. Go Alice. Thank you!

Confident AI provides enterprises an all-in-one platform to evaluate, observe, red team, and govern AI applications. Built by the creators of DeepEval (16k stars, 8m monthly downloads), Confident AI is able to offer battle-tested, open-source evaluation algorithms while providing the infrastructure needed for teams to stay confident their AI systems.

Docs are integral to a great customer experience, but managing them is a challenging, multi-team effort. The most successful companies recognize that fantastic docs drive adoption and bad ones cause customers to churn. Promptless is used by fast-growing startups and Fortune 500 enterprises alike to automatically update docs based on feature releases, support tickets, internal Slack conversations, and more. Promptless is designed to be flexible—it can integrate with any docs platform, be triggered automatically from many workflows, and can pull in context from any internal knowledge source. With Promptless, documentation becomes a strategic advantage for growth, not a bottleneck or a liability.

x1 is an AI app builder that helps anyone create real, App Store-ready iOS apps. Users start with a structured product plan, design polished iPhone screens in a visual canvas, build features step by step, and prepare launch assets like App Store listings, screenshots, monetization, and analytics. Instead of generating one-shot AI demos, x1 gives builders a guided system for turning an idea into a real app.

Pearson builds AI agents to automate corporate transactions. Large law firms use us to run gigantic books of businesses with AI, reducing their cost of delivery 40-60%.