YC-backed HyLight scales its autonomous inspection fleet as Spacesheep launches an AI-native workspace

Saturday, October 24, 2026

The latest from Y Combinator

YC-backed HyLight is scaling its autonomous inspection fleet. HyLight says it raised €13 million to expand from four hydrogen-powered airships to 15. It is hiring both a Software / Cloud / Data Engineer and an Embedded Software Engineer to make its data platform, payload systems and field upload tools robust enough for that larger fleet. The company says its airships already fly missions for major French energy and rail operators, making the hiring push part of an effort to move from working systems to fleet-scale operations.

Why it matters: HyLight’s expansion depends not only on more airships, but on reliably moving and processing the LiDAR, imagery and methane data they collect. The roles cover that full chain, from onboard capture and field uploads to customer-facing insights.

Y Combinator

Y Combinator

YC F26 startup Spacesheep launches an AI-native collaborative workspace. Y Combinator describes Spacesheep as a new surface where people and AI agents work together. Agents can turn their output into interactive documents, dashboards, design mockups or internal apps; teams can comment, watch updates, request changes and search across versions.

Why it matters: The product’s pitch is to make growing volumes of agent-generated work easier to review, share and manage in a collaborative format rather than leaving it as a stream of messages.

Y Combinator on LinkedIn

YC’s Winter 2027 application deadline is November 2. A Y Combinator application notice carried on Hacker News says applications for the Winter 2027 batch are open until November 2.

Why it matters: For founders considering YC, the notice establishes a specific upcoming deadline rather than a general invitation to apply.

Hacker News

Advice for early-stage founders

Treat this as general operating guidance, not a report of a new YC development.

• Start with customer discovery. Speak with people who have the problem before committing to a large build. Ask what they do today, what is frustrating about that workflow, and what they have already tried. Look for repeated, specific pain—not broad enthusiasm.

• Build the smallest useful product. Turn the clearest problem into one narrow workflow that delivers value quickly. Keep manual work behind the scenes if it helps you test whether the outcome matters before investing in infrastructure.

• Prioritize by learning value. Choose work that either improves the user’s result or answers an important uncertainty. Defer polish, edge cases and expansion until users repeatedly succeed with the core product.

• Use feedback as evidence, not a verdict. Watch what users do, where they stop, and whether they return. When feedback conflicts, investigate the underlying job and context rather than adding every requested feature.

• Treat distribution as part of the product. Decide where the people with this problem already look for solutions, then make a clear demonstration that shows the problem, the product and the result. YC’s Startup School discussion on making launch videos recommends treating every demo as a launch and choosing channels based on conversion rather than reach alone. Track qualified conversations, sign-ups and retained users—not views by themselves.