For years, investors have been drawn to Y Combinator’s Demo Days to discover promising startups developing innovative technology. The accelerator has been responsible for some of the largest tech firms globally, including Airbnb, Reddit, Dropbox, Zapier, and Stripe.

This is why we prioritize observing the event to identify the most intriguing companies from each cohort. As I’ve done nearly every quarter since the accelerator switched to four batches a year, I consulted nearly a dozen investors to find out which startups were the most sought after at Y Combinator’s Winter 2026 Demo Day earlier this week.

To guarantee our list featured genuinely desirable standouts, a company had to be recognized as a ‘favorite’ by at least two distinct venture capital investors to qualify.

Regarding valuations, I’m hearing that at least a couple of startups have secured funding at a $100 million valuation, notably with those startups already generating run-rate revenues of $1 million or more. Even for the less prominent startups not included here, the “default” valuation this quarter appears to be about $30 million, which investors indicated is approximately double the current seed market average.

Without further hesitation, here’s the list:

Beyond Reach Labs

What it’s creating: Deployable solar arrays for satellites.

Why it’s a favorite: The startup claims to have engineered solar arrays that begin at the dimension of a dining table upon launch but expand to the size of a football field once they reach orbit. The founders assert their system can enhance available power ten-fold while reducing costs by 88%. Beyond Reach has already scheduled a flight for 2027 and asserts it has obtained $325 million in letters of intent from leading space agencies.

Byteport

What it’s creating: An incredibly fast file transfer protocol.

Why it’s a favorite: According to Byteport’s founder Jayram Palamadai, current file transfer protocols such as TCP are too slow for the AI era. Thus, he developed DART, which stands for Dynamic Accelerated Record Transfer, capable of transferring large files at an average speed of 10 times faster than TCP, and even as much as 1,500 times quicker on “reliable connections.”

Hex Security

What it’s creating: Continuous AI-powered security testing instruments.

Why it’s a favorite: To combat hackers leveraging AI for relentless cyberattacks, Hex is developing AI agents that can function as penetration testers, perpetually searching for vulnerabilities and security gaps in corporate infrastructures. By automating tasks that were previously manual and infrequent, Hex contends it can thwart attacks at a significantly lower cost. The startup claims to have surpassed run-rate revenue of over $1 million in merely eight weeks, which could explain why VC investors, as someone mentioned to me, “were vying” to fund the company.

Grazemate

What it’s creating: Autonomous drones for managing and monitoring cattle.

 Why it’s a favorite: Herding cattle on large ranches is an expensive and perilous task, typically requiring helicopters and motorbikes. GrazeMate’s founder, who grew up on a 6,000-head cattle ranch in Australia, envisioned a solution to simplify life for ranchers, prompting him to leave college where he was studying robotics.

GrazeMate’s drones can autonomously direct cattle to various parts of a ranch, assess animal weights, grass availability and growth, and can adhere to predetermined route plans.

GRU Space

What it’s creating: Permanent lunar infrastructure, beginning with a hotel on the Moon.

Why it’s a favorite: “Humanity will pioneer interplanetary existence. It’s not a question of if, but when, and that time is now,” asserts GRU Space founder Skyler Chan, a recent Berkeley graduate who previously developed software at Tesla and worked on NASA-sponsored space technology.

Chan contends his startup has devised a “moon factory” capable of converting lunar soil into structural bricks, which he aims to utilize for constructing a luxury hotel on the moon as a “catalyst” for broader lunar infrastructure. GRU’s lofty ambitions, including the goal to launch the first lunar hotel by 2032, have positioned it as one of the most discussed startups of this YC batch. The company has already garnered $500 million in letters of intent, an invitation to the White House, and even a reservation from the Trump family.

Luel

What it’s creating: A marketplace for human-sourced data to train multimodal AI.

Why it’s a favorite: Founded by two UC Berkeley dropouts, Luel is developing a data marketplace that links AI model creators with contributors who can supply “daily-life” activities, such as ironing or doctor-patient dialogues, to furnish audio, video, and image data. The company claims it is generating an ARR of nearly $2 million within just six weeks, driven by high demand from robotics and voice AI laboratories.

Pax Historia

What it’s creating: An alternative-history strategy game powered by AI.

Why it’s a favorite: Pax Historia enables users to alter history in ways conventional strategy games cannot. Utilizing generative AI, the game reacts to infinite, multifaceted geopolitical scenarios, from “What if Rome never fell?” to “What if the USA annexed Greenland?” The founders assert the game currently draws 35,000 daily users who have engaged in nearly 20 million rounds.

Stilta

What it’s creating: Agentic AI catering to intellectual property and patent attorneys.

Why it’s a favorite: Stilta’s founders claim that patent litigation can incur costs of up to $4 million per case, primarily due to manual document review expenses. The startup states its AI agent is capable of searching and scrutinizing patents across databases and scientific research, saving both time and legal costs.

The company’s agents are already being utilized by IP attorneys at pharmaceutical powerhouse Roche. For investors, another enticing aspect is that the founders originate from Sweden — recent Swedish successes like Lovable and Legora have fostered something of a “halo effect” surrounding companies from the region, as one VC investor observed.