IIT Madras Alumni Offer $1 Million for 5 Unsolved Problems: Can AI Win It?

A startup founded by two IIT Madras graduates has put a $1 million prize pool behind five open questions in materials science. Here is what the problems are, what the rules say, and where the reporting does not quite agree.

The short version

  • Who: Discovered Materials, a San Francisco startup from Y Combinator's Spring 2026 batch, founded by IIT Madras alumni Advaith Sridhar and Akash Ramdas.
  • What: A $1 million prize pool for solving five open problems in materials science.
  • Announced: 5 November 2026.
  • Deadline: None fixed, according to Business Today.
  • The catch: Money is paid only when a solution meets the scientific criteria for that problem.

Prize headlines usually get a quick share and a quicker shrug. This one made me stop, mostly because of what the company did before announcing it. It pointed some of the strongest AI models in the world at these five problems, let them run for days, and then admitted they got nowhere much. Most startups lead with the win. This one led with the wall.

One small correction to the way this story is being shared online. The "IIT Madras" in the headlines is the founders' alma mater. The prize comes from a private company in San Francisco, not from the institute in Chennai.

Advertisement

What exactly is being offered?

Discovered Materials has put together a challenge with a total pool of $1 million. According to TechCrunch, which broke the story, the company recruited an advisory group of academics from leading universities to spell out five open problems. The company's own work is finding new materials for semiconductor chips using swarms of AI agents, so the questions sit close to its home turf.

The challenge is aimed at AI models first. But Business Today reports that it is open to eligible individuals, teams and organisations, so a university lab or a lone researcher is not shut out.

The five problems, in plain English

Materials science has no list like the famous Millennium Prize Problems in mathematics. That is partly why this exercise is interesting. Here is what the five challenges ask for, as TechCrunch describes them.

#ProblemWhat a solution looks likeWhy it matters
1Cuprate superconductivityA theory explaining how certain copper-based materials carry electricity with zero resistance at ultra-low temperaturesA long-standing physics puzzle; an explanation would sharpen our understanding of how electrons behave in these materials
2Silicon and silicon dioxide interfacesA numerical method to predict the structure where the two materials meetThis pairing is found in almost all semiconductors
3Grain structure and heatA numerical method that maps grain structure to thermal propertiesCould help design materials that shed heat better
4Light-driven changesA theory predicting why shining light on a material can change its behaviour, especially around superconductivityExplains an effect scientists can observe but not yet account for
5Density functional theoryA numerical method that improves this widely used technique for modelling atoms and moleculesA better method could make quantum chemistry calculations faster and more accurate

Felipe Jornada, a Stanford professor who helped shape the challenges, summed up the spirit of the list by saying that across the field there are phenomena we observe but cannot explain.

Advertisement

The two IIT Madras alumni behind it

Advaith Sridhar and Akash Ramdas met at IIT Madras more than ten years ago, per Business Today. Ramdas holds a PhD in materials science from Stanford and has spent years on semiconductor research. Sridhar came from the AI side, working on agent systems at Persona AI and Luma Labs, according to TechCrunch's earlier coverage of the company.

The startup closed a $9 million seed round in August 2026. Startup Fortune reports that Lightspeed India Partners led it, with Y Combinator, Peak XV Partners and angel investor Paul Graham joining in. So the $1 million is not a stunt funded by a bottomless balance sheet. It is a big slice of a young company's seed money, which tells you how seriously the founders take the idea.

How the $1 million is split: reports disagree

This is the part where I would slow down before repeating any number. Three outlets, three different pictures of the payout. The company's official challenge page was not available to me while reporting this, so I cannot tell you which figure is final.

OutletWhat it reports
TechCrunchA pool worth $1 million in total, with no per-problem split stated
Business Today$400,000 for the cuprate problem, $200,000 each for two others, $100,000 each for the remaining two
Startup Fortune$200,000 for each of the five problems

Both splits add up to $1 million, so neither is absurd on its face. If you plan to attempt one of these, read the official terms on the Discovered Materials grand challenges page rather than trusting any news summary, including this one.

Why the AI test matters more than the prize

Before going public, Discovered Materials gave frontier AI models internet access and other resources and let them work autonomously on each problem for several days. TechCrunch reports they made limited headway. The same report adds a fair caveat: a newer model or a lot more computing power could change that picture quickly.

Sridhar's own framing is that the company does not mind who solves them, whether that is his team, a professor or a big AI lab. The challenges are meant to be goalposts, a public way to see whether AI is actually climbing in this field. That is a more honest pitch than "we will cure everything with AI," and I think it is the more useful one.

Advertisement

My take: what I would watch for

Weak prize announcements tend to share a habit: the problem is vague, so the winner can always be argued about later. Materials science is genuinely harder to pin down than maths. As Sridhar himself pointed out, there is rarely a clean yes or no. So the real work here was not the money. It was getting academics to write the problems precisely enough that a judge, or a model, can check an answer.

Three things I would keep an eye on:

1. The verification step. A theory that sounds right on paper still has to match experiments, and experiments are slow. Expect any claimed solution to take months of scrutiny, not days.

2. The size of the problem against the size of the prize. Cuprate superconductivity has resisted physicists since the materials were discovered in 1986. A few hundred thousand dollars will not change who is working on it. It might change who is willing to publicly try.

3. Whether a lab claims one. If a model solves even the smallest of the five, that would be a real marker for AI in science. If nothing happens for a year, that is a useful data point too, and the company has said as much.

For Indian readers there is a quieter point. Another pair of IIT graduates is building at the frontier of a hard science problem, and doing it from outside India. Whether that bothers you or makes you proud is a personal call, but it is worth noticing.

Can you enter, and how?

Based on the reporting so far, there is no fixed deadline, and entries from individuals, teams and organisations are allowed. A payout happens only if the proposed solution satisfies the scientific criteria laid down for that specific challenge. I have not seen a public description of the submission process, so that is another reason to go straight to the official page.

Advertisement

Frequently asked questions

Who is offering the $1 million prize?

Discovered Materials, a San Francisco startup from Y Combinator's Spring 2026 batch. It was founded by IIT Madras alumni Advaith Sridhar and Akash Ramdas.

What are the five problems in the challenge?

A theory of cuprate superconductivity; a numerical method for silicon and silicon dioxide interface structures; a numerical method linking grain structure to thermal properties; a theory of how light changes material behaviour, especially around superconductivity; and an improvement to density functional theory.

How much does each problem pay?

Reports differ. TechCrunch gives only the $1 million total. Business Today reports $400,000, $200,000, $200,000, $100,000 and $100,000. Startup Fortune reports $200,000 each. Check the official challenge page for the final numbers.

Is there a deadline to enter?

Business Today reports no fixed deadline. Prizes are paid only when a solution meets the scientific criteria for that problem.

Can humans enter, or is it only for AI?

It is framed around AI models, but it is reported as open to eligible individuals, teams and organisations. The founder has said it does not matter whether a model, a professor or a frontier lab solves it.

Has anyone solved any of the five problems yet?

I could not find any report of a solution as of 9 November 2026. In the company's own tests, frontier AI models worked on each problem for days and made limited progress.

How much funding has Discovered Materials raised?

A $9 million seed round in August 2026, led by Lightspeed India Partners with Y Combinator, Peak XV Partners and angel investors including Paul Graham.

The bottom line

Strip away the headline and this is a bet that AI can do more than predict: that it can explain. The odds of someone collecting a prize soon look low, and the company seems to know it. What makes the announcement worth your attention is the honesty of the setup. Five written-down problems, a public record of AI failing at them, and an open invitation to do better. If a solution turns up, you will hear about it. Until then, treat any figure you see for the individual payouts, including the ones above, as provisional.

HO

Harsh Ojha

Correspondent, Newszest.in  |  1 Year in News & Media
The newest byline in the newsroom, Harsh covers technology, politics and the startup world with an eye for what's actually new versus what's just repackaged. A year in, he's already found his beat.
Covers: Technology • Politics • Startups
Advertisement

Post a Comment

Previous Post Next Post