With every new season comes another YC demo day. On Tuesday, the Spring 2026 cohort was revealed, filled with defense technology, robotics, AI infrastructure and developer tools, and of course AI agents.
TechCrunch spoke with eight investors to determine which were the hottest companies in this bunch—both the ones they’ve been watching and the startups they heard other VCs couldn’t get enough of. This list consists primarily of companies that were flagged by at least two investors as the most crowded of the lot.
That excitement translated into bigger rounds and huge prices for some of these companies. Just when we thought YC’s valuations couldn’t get any higher, they exceeded all expectations. This batch had at least two startups with valuations of $175 million or more. Investors were also clearly willing to pay a premium for proven, repeat founders.
We’ve organized the hits alphabetically below:
9 Mothers
What it builds: AI powered anti-drone systems
Why it’s a favorite: The Russia-Ukraine conflict shows just how deadly small drones are, with the systems now accounting for around 80% of casualties. Existing anti-drone solutions are expensive and often useless to block swarms of drones flying at low altitudes. 9 Mothers claims to have developed a more “economical” robot that can track and kill drones traveling at 60 miles per hour.
The startup, founded in 2024, has already closed $1.6 million in sales, with a single contract expected to expand to $35 million later this year. The company also promises investors that it can reach $1 billion in contracts. That potential combined with a clear need had VCs willing to invest, at a valuation of over $200 million, one VC told us. At this valuation, 9 Mothers is not only the most valuable startup of the lot, but potentially one of the most valuable in YC history. (9 Mothers did not respond to our request for comment.)
Arga Labs
What it builds: A tool that provides digital twin environments for testing AI agents
Because it’s a favorite: Artificial intelligence helps software engineers create more code at much faster speeds. This code needs to be tested, but traditional test environments, also known as sandboxes, cannot be built fast enough to keep up. Arga Labs solves this problem by instantly spinning up “digital twins” of a company’s software, allowing AI agents to safely test their code before it goes into production.
Adialande
What it builds: Mobile MRI clinics for early cancer detection
Because it’s a favorite: We know early detection saves lives, but the healthcare system doesn’t have enough MRI machines to screen everyone regularly. They are very expensive — they cost millions to buy and tens of thousands a year to maintain. Adialande tries to solve this problem. The startup says it has designed a compact MRI unit that can be transported in a small truck. Adialante’s business model is to bring these machines into clinics and charge $250 per scan. The company’s huge but worthy ambition is to change MRIs from being reserved for symptomatic patients to a routine annual check-up.
Compl
What it builds: Compliance management for physical products
Because it’s a favorite: It can be quite complicated to ship physical goods across borders. There are so many regulations, such as ensuring correct translations when selling in the EU or ensuring a beauty product label complies with each country’s labeling requirements. Compl hopes to make it easier by creating AI agents that help businesses manage and track compliance, risk and regulatory changes, and create the necessary documentation and product labels for products that will be shipped internationally. It’s no shock that investors have taken notice of a company like this – its product addresses a particular pain point for many businesses where AI could help.
Letter
What it builds: Satellites that can safely return products manufactured in space back to Earth
Because it’s a favorite: Space, due to its microgravity and near-perfect vacuum, is an excellent environment for manufacturing certain pharmaceuticals, semiconductors, and 3D printed human tissues and organs. Manufacturing in space could be the future, too Letter he wants to be a part of developing vehicles that can withstand the heat and return these products back to Earth. Unlike many space capsules that burn up or are discarded after a single use, Dispatch designs its vehicles to be refurbished and sent back into space again and again. VCs are betting that space manufacturing is closer than we think, and that there could be great value in handling the transportation of these products.
Lightsprint
What it builds: A tool that allows non-engineers to ship and build features without writing code
Why it’s a favorite: Lightsprint allows anyone, with or without coding experience, to build and ship production features for applications. In the past, if a change needed to be made to an application, a product manager had to wait for an engineer to fix it. But Lightsprint hopes to be the tool product managers use immediately. Users explain what they want to fix or adjust, then choose from visual options for how they want the changes to appear, and a Lightsprint AI agent will create it for them. After the product manager is satisfied, an engineer reviews, approves, and merges the new code into the existing code, and then the company can officially ship the new feature.
Game
What it builds: A tool that automates website creation and marketing development
Because it’s a favorite: Our sources weren’t kidding Game it’s hot Shortly after we added the company to our list, Ploy announced a healthy $27M round led by First Round and Y Combinator. The draw? Ploy was founded by Bryant Chou, co-founder and former CTO of Webflow, a drag-and-drop website builder with a recent valuation of $4 billion. Ploy takes website building even further by automatically creating landing pages, writing marketing copy and launching campaigns. The startup’s AI agents are constantly improving the site’s content with the goal of accelerating inbound growth. In doing so, the startup promises to eliminate the need for a large marketing team.
Sazabi
What it builds: An AI-powered platform for finding and creating fixes for software problems
Why it’s a favorite: Sazabi founded by Sherwood Callaway, a repeat founder of YC, whose resume also includes being an a16z scout, an employee at Brex and 11x. These kinds of experiences alone make him a lovable investor. But VCs are also interested in the product, a tool that helps find and flag problems in software production systems. Sazabi integrates with Slack to communicate with teams through this chat app. It performs log analysis (the script that the software writes while running) to determine why something broke in production and allows users to tell it to build and submit a fix for the problem with one click.
Silmaril
What it builds: AI security infrastructure
Because it’s a favorite: Autonomy is a hot space right now for investors, as they fund countless companies that put AI agents to work in important parts of the business. Silmaril works to ensure that agents cannot be compromised via “prompt injection” — meaning that the public parts of the agent cannot be manipulated by hackers via messages, emails, or other documents. Silmaril agents autonomously probe for new threats to agents and applications, and once they find a threat, autonomously train the firewall to develop immunity to it.
Superset
What it builds: A platform that allows developers to run at least 100 coding agents simultaneously
Why it’s a favorite: Coding agents have completely invaded the world of software development. As they multiply, there must be a way to manage them in one space. That’s what Superset seeks to do. Its tool allows developers to launch 100 or more coding agents simultaneously and manage them. Any command line interface (CLI) agent — such as Claude or Cursor — can be run on the platform, and each agent runs in its own little workspace, so there are no conflicts between them. The tool can also be opened in any integrated development environment (IDE), such as VS Code or Cursor.
Tasklet
What it builds: An artificial intelligence agent that performs tasks
Because it’s a favorite: The company’s premise is simple but has the potential to be incredibly effective. Tasklet is an AI agent that connects to the APIs of various work apps like Slack, Outlook, Google Drive, and more. Users can then ask it to perform tasks specific to those applications using natural language.
A user can ask it to automatically handle workflows and perform tasks such as sorting emails or downloading reports. It also runs continuously, meaning that even if the user closes a tab, the agent will continue to run. It can write and execute its own code as well as create interfaces.
Although some investors have noted that they are shying away from investing in horizontal tools like this, it seems that Tasklet is running against the grain. It’s positioned as a place where users go to issue commands rather than opening apps individually. There is a future, perhaps, in which this will become a default way for some to interact with software.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
