Investors have been flocking to Y Combinator demo days for years to get their hands on promising startups building cool technology. After all, this accelerator has produced some of the world's biggest technology companies, from Airbnb and Reddit to Dropbox, Zapier, and Stripe.
That's why we keep an eye on events to find the most interesting companies from each batch. As I do almost every quarter now that the accelerator has moved to four cohorts a year, I asked nearly a dozen investors at Y Combinator's Winter 2026 Demo Day earlier this week which startups were most in demand.
To ensure that our list included truly popular and standout companies, a company had to be marked as a “favorite” by at least two different venture capital investors.
In terms of valuations, we've heard of at least a few startups raising funding with $100 million price tags, but what's notable is that those startups have already generated run-rate revenues of $1 million or more. Even the less talked about startups that aren't on this list appear to have a “default” valuation of about $30 million this quarter, which investors tell me is about double the current seed market average.
Without further ado, here is the list.
beyond reach lab
What we're building: Deployable solar arrays for satellites.
Why it's popular: The startup claims to have developed a solar array that is the size of a dining table at launch, but expands to the size of a football field once it reaches orbit. The founders say their system can increase available power tenfold while reducing costs by 88%. Beyond Reach is already planning a 2027 flight and says it has secured $325 million in letters of intent from major space companies.
tech crunch event
San Francisco, CA | October 13-15, 2026
bite port
What we're building: An incredibly fast file transfer protocol.
Why it's popular: According to Byteport founder Jayram Palamadai, existing file transfer protocols like TCP are too slow for the AI era. That's why he built DART, which stands for Dynamic Accelerated Record Transfer. DART appears to be able to transfer large files on average 10 times faster than TCP, and up to 1,500 times faster on “reliable connections.”
hex security
What we're building: An AI-powered continuous security testing tool.
Why we love it: To combat hackers who use AI to launch non-stop cyberattacks, Hex is building AI agents that can act as penetration testers and constantly probe a company's infrastructure for vulnerabilities and security gaps. Hex claims it can prevent attacks at a fraction of the cost by automating manual processes that were once performed infrequently. The startup claims run-rate revenue of more than $1 million in just eight weeks, which may be why VC investors were “vying” to invest in the company, one person told me.
Glazemate
What we're building: Autonomous drones to herd and monitor cattle.
Why it's popular: Moving cattle on large ranches is an expensive and dangerous process, often using helicopters or motorcycles. GrazeMate's founder, who grew up on a 6,000-cow cattle farm in Australia, dropped out of university to pursue a degree in robotics in search of ways to make ranchers' lives easier.
GrazeMate’s drones can automatically guide cattle to different areas of the farm, estimate animal weight, grass availability and growth, and follow pre-specified route plans.
GRU space
What we're building: Permanent lunar infrastructure, starting with a lunar hotel.
Why it's popular: “Humanity is going to have interplanetary travel. It's not a matter of when, it's now,” says GRU Space founder Skyler Chan, a recent graduate of Berkeley. He previously developed software at Tesla and worked in space technology funded by NASA.
Chan claims his startup has developed a “lunar factory” that can turn lunar soil into structural bricks, and plans to use it to build luxury hotels on the moon as a “wedge” for broader lunar infrastructure. GRU's astronomical ambitions, including the goal of opening the first lunar hotel by 2032, have made it one of the YC group's hottest startups. The company has already secured a $500 million letter of intent, an invitation to the White House and even a booking from the Trump family.
Ruel
What we're building: A marketplace for human-captured data to train multimodal AI.
Why it's popular: Founded by two University of California, Berkeley dropouts, Luel is building a data marketplace that connects AI modelers with contributors who can submit “everyday life” activities like ironing or patient-doctor conversations, and provides audio, video, and image data. The company claims to have generated nearly $2 million in ARR within six weeks, driven by high demand from its robotics and voice AI labs.
pax historia
What we're building: An AI-powered alternate history strategy game.
Why we love it: Pax Historia allows users to rewrite history in a way that is not possible in traditional strategy games. The game uses generative AI to address an infinite number of complex geopolitical scenarios, ranging from “What if Rome never fell?” “What would happen if America occupied Greenland?'' The founders claim that the game currently attracts 35,000 daily users, who have played nearly 20 million rounds to date.
Stilta
What we're building: Agentic AI for intellectual property and patent attorneys.
Why it's popular: Stilta's founders claim that patent disputes can cost up to $4 million each, largely due to the cost of manual document review. The company says its AI agents can search and analyze patents across databases and scientific literature, saving both time and legal fees.
The company's representatives are already being used by pharmaceutical giant Roche's IP lawyers. Another attractive aspect for investors is that the founders are from Sweden. Recent Swedish successes such as Labable and Regola have created something of a “halo effect” around companies in the region, one VC investor said.