A Startup Posted $1M in Science Prizes. Run Your Own Heat Test.
A Y Combinator startup called Discovered Materials announced about $1 million in prizes for five unsolved science problems on November 5, 2026. Here is how kids can test heat and materials at home like real scientists.

Have you ever left a hot cup of cocoa on the table and noticed it cooled faster on a metal tray than on a wooden board?
That tiny mystery is the same kind of puzzle that a startup is now offering real money to solve. A young company just announced about one million dollars in prizes for five science problems nobody has cracked yet.
You can't win those prizes this week. But you can think like the people who might, and you can run a heat test at your kitchen table. Let's go.
What did the startup announce?
A Y Combinator startup called Discovered Materials announced a challenge with a prize pool worth one million dollars in total. TechCrunch reported it on November 5, 2026, and says it offers $200,000 for each of five solutions.
Y Combinator is a famous program that helps new companies get started. Discovered Materials was in its spring 2026 batch and is based in San Francisco.
The company uses AI models to search for new materials, which are the stuff things are made of, like metal, glass or plastic. Its goal is to find better materials for advanced computer chips.
What are the five problems?
They're grown-up science puzzles, so here are kid-sized versions based on TechCrunch's list.
- Superconductors: one problem asks for a theory to explain why some materials carry electricity with zero resistance at very cold temperatures.
- Silicon edges: another wants a way to predict how silicon and silicon dioxide, two chip materials, fit together where they meet.
- Heat and grains: a third asks how the tiny crystal pieces inside a material affect how well it moves heat away.
- Light and behavior: the fourth is about why shining light on certain materials can change how they act.
- Better modeling: the last wants a better method for computer models of atoms and molecules.
Don't worry if that sounds like another language. Even adults call these hard. The point is that real companies pay for answers to real questions.
Why does heat matter to computers?
Computer chips make heat, and too much heat slows them down or breaks them. Pulse 2.0 reports that the startup says some AI chips can reach about 140 watts per square centimeter. That's a huge amount of heat in a tiny space.
Think of a laptop on your legs after a long video call. It feels warm because the chip is working hard and the heat has to go somewhere.
Better materials can carry heat away faster. If a company finds one, it can help chips stack closer together and run cooler. That's the business idea: solve a hard science problem, then sell the result to the people who build chips.
Who started the company?
The founders are Akash Ramdas and Advaith Sridhar. Pulse 2.0 says Ramdas has a materials science PhD from Stanford and about 11 years of research in semiconductor materials, while Sridhar studied AI at Carnegie Mellon. They met at IIT Madras more than ten years ago.
They raised a $9 million seed round, with Lightspeed India Partners leading and Y Combinator and Peak XV Partners joining.
Notice what they did. One founder knew the science. The other knew the computers. Together they covered a gap neither could cover alone. That's a great tip for any team project, including yours.
Why would a company give away prizes?
Prizes act like goalposts. TechCrunch quotes the CEO saying that tracking which problems get solved shows whether AI models are making progress in this area.
The company also tested leading AI models first. According to TechCrunch, it let them work for days on each problem, and they made little progress. So the prizes aren't easy. They're a way of saying "here's where the hard edge is."
A prize also gets smart people talking about your problem. That's marketing and research at the same time. Lots of businesses, from small shops to big labs, use contests to find ideas they wouldn't think of themselves.
Run the Cocoa Cooling Test
Time for your own mini lab. You'll need a grown-up for the hot water.
Gather: four identical mugs or jars, a kitchen thermometer, a timer, paper and a pencil. Wrap each container in a different material: aluminum foil, a wool sock, bubble wrap and nothing at all.
Steps:
- Ask an adult to pour the same amount of warm water (not boiling) into each container.
- Write down the starting temperature of each.
- Check every 5 minutes for 30 minutes.
- Record all the numbers in a table.
- Draw a simple line graph for each container.
Predict first. Which material will keep the water warmest? Which will let heat out fastest? Write your guess before you start.
Then compare. Were you right? If not, that's perfect. Surprises are how scientists learn.
Turn the test into a business idea
Here's where the startup brain comes in. Suppose the sock won. Who needs to keep things warm? Lunch boxes, coffee cups, pet water bowls, hand warmers.
Pick one and ask five people what they'd pay for a better version. Write the answers down. If three say "I'd buy that for $5," you've got a lead worth following.
You could sketch your idea on paper or use a notebook, a spreadsheet or a planning tool like Foundra that helps beginners turn an idea into simple steps. Ask a parent to help you choose.
Keep the experiment honest. Don't claim your sock cup works better than anyone else's until you've tested it several times.
How can you think like a scientist founder?
Start with a question you can actually measure. "Which keeps water warm longest?" is measurable. "Which is best?" isn't.
Change one thing at a time. If you use different cups and different amounts of water, you won't know what caused the result.
Write everything down, even failures. Nobody collected a prize without lots of wrong answers first.
And ask for help. The founders of Discovered Materials worked as a team, and they asked experts at universities to help define the five problems. According to TechCrunch, a Stanford professor, Felipe Jornada, helped develop them. Great scientists borrow brains.
What if your experiment gives a boring answer?
Then you've learned something. Maybe all four containers cooled at nearly the same speed. That's a result, and it raises a better question: was your test fair, or were the differences too small to see with a kitchen thermometer?
Real researchers hit this all the time. TechCrunch notes that the AI models in the startup's own test made little progress on the five problems, which tells the company where the limits are. A "no progress" result still draws a map.
Try the test again with one change. Use thicker wool, add a lid or switch to a bigger cup. Run it three times and average the numbers. If the pattern holds, you can trust it more.
Then share what you found with a friend or teacher. Explaining your result out loud is one of the best ways to check whether you understand it.
Key takeaways
- Discovered Materials announced a prize pool worth about $1 million for five unsolved materials science problems on November 5, 2026.
- The startup is from Y Combinator's spring 2026 batch and raised a $9 million seed round.
- Chips get very hot, and better materials can help move heat away.
- Prizes can work like goalposts that show where the hard edge of a field is.
- You can test materials at home by wrapping containers and tracking temperature.
- Change one thing at a time, write everything down and ask for help.
Frequently asked questions
What is Discovered Materials? It's a Y Combinator startup that uses AI to look for new materials for computer chips.
How big are the prizes? The total pool is worth about $1 million, with $200,000 offered per solution, according to TechCrunch.
Can kids enter the contest? The article doesn't say who can enter, so check the company's website for rules.
What is materials science? It's the study of what things are made of and how those materials behave, like how well they carry heat or electricity.
Is the cooling test safe? It is with an adult pouring the water. Use warm water, not boiling, and keep hands clear of hot containers.
What makes a good science experiment? A question you can measure, one change at a time, and careful notes.
Sources
- A YC startup is putting $1M behind five unsolved materials science problems for AI, TechCrunch, November 5, 2026
- Discovered Materials Raises $9 Million Seed Funding To Accelerate AI-Powered Semiconductor Materials Discovery, Pulse 2.0
- Discovered Materials Raises $9 Mn For AI-Led Semiconductor Material Discovery, Inc42
- Discovered Materials Raises USD9 Mn Seed Round, Entrepreneur India
- Discovered Materials launches AI agents to find chip materials labs can make, Runtime Wire
Ready to help a young entrepreneur get started?
Foundra Kids gives young founders a simple, fun way to plan their first business.
Try Foundra Kids

