VCs pick 11 standout startups from YC Demo Day

Each Y Combinator Demo Day brings a new wave of ambition, and the Spring 2026 cohort did not disappoint. Unveiled on Tuesday, this batch was dominated by defense tech, robotics, AI infrastructure, and, as expected, a flood of AI agents.

TechCrunch surveyed eight venture capitalists to identify the hottest companies from this cycle, focusing on startups that generated the most buzz among investors. The excitement was so palpable that it translated into record-breaking valuations, with at least two companies commanding valuations of $175 million or more. Investors showed a clear preference for proven, repeat founders and were willing to pay a premium for them. Here are the 11 standout startups, listed alphabetically.

9 Mothers

This startup is building AI-powered counter-drone systems. The war in Ukraine has exposed the devastating effectiveness of small drones, which now account for roughly 80% of casualties. Existing counter-drone solutions are often too expensive and ineffective against swarms flying at low altitudes. 9 Mothers claims to have developed a more affordable robot that can track and kill drones traveling at 60 miles per hour. Founded in 2024, the company has already booked $1.6 million in sales, with one contract expected to expand to $35 million later this year. Investors are betting on a pipeline that could reach $1 billion, pushing the startup’s valuation past $200 million. One VC described it as the most highly valued startup of the batch, potentially one of the most valuable in YC history.

Agra Labs

Agra Labs addresses a critical bottleneck in modern software development. As AI agents generate code at unprecedented speeds, traditional testing environments cannot keep up. Agra Labs solves this by instantly creating digital twins of a company’s software, allowing AI agents to safely test code before it reaches production. This tool is essential for engineering teams struggling to maintain quality while accelerating output.

Adialante

Early detection saves lives, but the healthcare system lacks enough MRI machines for regular screening. The machines are prohibitively expensive, costing millions to buy and tens of thousands annually to maintain. Adialante is tackling this with a compact MRI unit that fits in a small truck. Their business model brings these mobile clinics to doctors’ offices and charges $250 per scan. The ambitious goal is to transform MRI scans from a diagnostic tool for symptomatic patients into a routine annual screening.

Complir

Shipping physical goods across borders is a regulatory nightmare. Companies must navigate a maze of rules, from proper translations for the EU to country-specific labeling for beauty products. Complir uses AI agents to help businesses manage compliance, risk, and regulatory changes, automatically generating the necessary documentation and product labels. It addresses a painful, universal problem where AI can provide immediate value.

Dispatch

Space manufacturing is gaining traction, thanks to microgravity and a near-perfect vacuum ideal for producing pharmaceuticals, semiconductors, and 3D-printed human tissues. Dispatch wants to be the logistics provider for this emerging industry, building reusable vehicles that can safely return manufactured products to Earth. Unlike most space capsules that are discarded after one use, Dispatch’s vehicles are designed to be refurbished and sent back up repeatedly. VCs are betting that space manufacturing is closer than we think, and that handling the transportation will be a lucrative niche.

Lightsprint

Lightsprint empowers non-engineers to build and ship production features without writing a single line of code. Instead of waiting for an engineer to make changes, a product manager can describe what they want fixed or adjusted, select visual options for the changes, and let a Lightsprint AI agent build it. After the product manager is satisfied, an engineer reviews, approves, and merges the code. This tool promises to dramatically accelerate feature development and reduce bottlenecks.

Ploy

Ploy was already generating significant buzz before Demo Day, and it validated that excitement by announcing a $27 million seed round led by First Round and Y Combinator. The startup is founded by Bryant Chou, co-founder and former CTO of Webflow, a drag-and-drop website builder last valued at $4 billion. Ploy takes website creation further by automatically generating landing pages, writing marketing copy, and launching campaigns. Its AI agents continuously refine website content to accelerate inbound growth, promising to reduce the need for a large marketing team.

Sazabi

Sazabi is founded by Sherwood Callaway, a repeat YC founder with a resume that includes stints at a16z, Brex, and 11x. That pedigree alone makes him an investor favorite. But the product is equally compelling. Sazabi is an AI-powered platform that finds and fixes software problems in production. It integrates with Slack, performs log analysis to determine why something broke, and lets users generate and submit a fix with one click.

Silmaril

As companies deploy AI agents to handle critical business functions, security becomes paramount. Silmaril focuses on preventing prompt injection attacks, where hackers manipulate an agent’s public-facing parts via prompts, emails, or documents. Silmaril’s agents autonomously probe for new threats and, upon discovery, retrain the firewall to develop immunity. This is essential infrastructure for the autonomous future.

Superset

With coding agents proliferating in software development, managing them in one space is becoming a necessity. Superset allows developers to launch 100 or more coding agents simultaneously and manage them all from a single platform. Any CLI agent, like Claude or Cursor, can run on the platform, each in its own isolated workspace to prevent conflicts. It also works with any IDE, such as VS Code or Cursor.

Tasklet

Tasklet is a task-performing AI agent that connects to work applications like Slack, Outlook, and Google Drive via their APIs. Users can tell it to perform tasks using natural language, such as sorting emails or pulling reports. It runs continuously, even if the user closes a tab, and can write its own code and build interfaces. While some investors are moving away from horizontal tools, Tasklet is betting that a command-based interface could become the default way people interact with software.