{"id":3491597,"date":"2026-08-10T12:00:00","date_gmt":"2026-08-10T12:00:00","guid":{"rendered":"https:\/\/techingeek.com\/index.php\/2026\/08\/10\/discovered-materials-is-engaging-in-ai-whack-a-mole-to-seek-out-better-chips\/"},"modified":"2026-08-10T12:00:00","modified_gmt":"2026-08-10T12:00:00","slug":"discovered-materials-is-engaging-in-ai-whack-a-mole-to-seek-out-better-chips","status":"publish","type":"post","link":"https:\/\/techingeek.com\/index.php\/2026\/08\/10\/discovered-materials-is-engaging-in-ai-whack-a-mole-to-seek-out-better-chips\/","title":{"rendered":"Discovered Materials is engaging in AI whack-a-mole to seek out better chips."},"content":{"rendered":"
<\/div>\n
\n

Chips executing AI tasks tend to overheat: This is one factor contributing to the high electricity usage in data centers and the need for cooling solutions. Consequently, innovators are leveraging AI to address the issue it has caused.<\/p>\n

Discovered Materials is the newest entrant, aiming to employ swarms of AI agents to uncover new materials that enhance the creation of more efficient integrated circuits. The startup recently announced it has secured a $9 million seed funding round from Lightspeed India Partners, having recently graduated from Y Combinator, with investments from Peak XV Partners and angel investors such as Paul Graham, Gokul Rajaram, and Thariq Shihipar.<\/p>\n

Founders Advaith Sridhar and Akash Ramdas joined forces to establish the company, leveraging Ramdas\u2019 expertise from earning a PhD in materials science at Stanford and Sridhar\u2019s experience with agents at Persona AI and Luma Labs.<\/p>\n

The duo has developed a software pipeline that utilizes Anthropic models within a custom setup to generate leads for materials, then applies foundational physics models they have trained to conduct simulations that confirm the potential of candidate materials.<\/p>\n

\u201c[Ramdas] was making perhaps 20 hypotheses a day during his PhD,\u201d Sridhar shared with TechCrunch. \u201cNow, we can generate thousands of hypotheses daily by having these agents operate round the clock in the cloud, exploring the research directions he provides.\u201d<\/p>\n

Discovered Materials unveiled examples of hundreds of new materials today, along with their \u201cMaterial Discovery Bench,\u201d designed to monitor how cutting-edge models tackle this challenge.<\/p>\n

Similar initiatives have emerged from companies like MatNex, SandboxAQ, and CuspAI, but Discovered Materials is focusing specifically on the thermal challenges faced by semiconductor materials as a pathway to success. The startup claims it has already identified several materials with properties comparable to those used by leading chipmakers but is unable to disclose further specifics.<\/p>\n

One of the hurdles is the engineering trade-off: If a new material could lower heat generation or enhance dissipation, it may still be impractical to manufacture a chip using it, or its electrical characteristics might be diminished.<\/p>\n

\u201cIt\u2019s somewhat like playing whack-a-mole with atomic structures,\u201d remarked Hemant Mohapatra, the Lightspeed partner who spearheaded this funding round, in conversation with TechCrunch. \u201cA material is only practical in the real world if all properties align simultaneously, which adds intrigue to this search problem.\u201d<\/p>\n

Mohapatra anticipates that the market for predicting new substances will become commoditized as models advance. The distinction with Discovered Materials lies in Ramdas\u2019 extensive experience in the domain and the capability to operate a lab capable of swiftly experimenting and validating potential candidates \u2014 a feat he asserts both founders have already accomplished with multiple new materials.<\/p>\n

Once valuable candidates are identified, Sridhar mentions the company will seek to patent the application of these materials in GPUs, or the method of creating chips from the substances, licensing them to chip manufacturers. He aspires to have patent-worthy new materials within the year.<\/p>\n

However, amidst all the enthusiasm, there has yet to be any drugs or materials discovered by AI that have made a significant commercial impact. The closest instance might be Insilico Medicine\u2019s Renterosib, the inaugural drug found with generative AI that has progressed to a Phase II clinical trial. In terms of materials, promising contenders like MatNex\u2019s rare-earth-free permanent magnets and new semiconductor materials developed by Panasonic and Citrine Informatics have been identified, but they have not yet been deployed commercially at scale.<\/p>\n

These approaches may be maturing as AI technology progresses, yet Mohapatra contends that finding more candidates is not the bottleneck in AI materials science; rather, \u201cthe challenge lies in accurately filtering and synthesizing them.\u201d<\/p>\n

While Sridhar believes Discovered Materials\u2019 unique data and insights will enable the startup to compete with well-funded advanced labs, he acknowledged that much of the work will involve actually going into wet labs to create materials. This is a phase of the process that cannot be expedited.<\/p>\n<\/div>\n

When you purchase through links in our articles, we may earn a small commission. This doesn\u2019t affect our editorial independence.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"

<\/div>\n
\n

Chips executing AI tasks tend to overheat: This is one factor contributing to the high electricity usage in data centers and the need for cooling solutions. Consequently, innovators are leveraging AI to address the issue it has caused.<\/p>\n

Discovered Materials is the newest entrant, aiming to employ swarms of AI agents to uncover new materials that enhance the creation of more efficient integrated circuits. The startup recently announced it has secured a $9 million seed funding round from Lightspeed India Partners, having recently graduated from Y Combinator, with investments from Peak XV Partners and angel investors such as Paul Graham, Gokul Rajaram, and Thariq Shihipar.<\/p>\n

Founders Advaith Sridhar and Akash Ramdas joined forces to establish the company, leveraging Ramdas\u2019 expertise from earning a PhD in materials science at Stanford and Sridhar\u2019s experience with agents at Persona AI and Luma Labs.<\/p>\n

The duo has developed a software pipeline that utilizes Anthropic models within a custom setup to generate leads for materials, then applies foundational physics models they have trained to conduct simulations that confirm the potential of candidate materials.<\/p>\n

\u201c[Ramdas] was making perhaps 20 hypotheses a day during his PhD,\u201d Sridhar shared with TechCrunch. \u201cNow, we can generate thousands of hypotheses daily by having these agents operate round the clock in the cloud, exploring the research directions he provides.\u201d<\/p>\n

Discovered Materials unveiled examples of hundreds of new materials today, along with their \u201cMaterial Discovery Bench,\u201d designed to monitor how cutting-edge models tackle this challenge.<\/p>\n

Similar initiatives have emerged from companies like MatNex, SandboxAQ, and CuspAI, but Discovered Materials is focusing specifically on the thermal challenges faced by semiconductor materials as a pathway to success. The startup claims it has already identified several materials with properties comparable to those used by leading chipmakers but is unable to disclose further specifics.<\/p>\n

One of the hurdles is the engineering trade-off: If a new material could lower heat generation or enhance dissipation, it may still be impractical to manufacture a chip using it, or its electrical characteristics might be diminished.<\/p>\n

\u201cIt\u2019s somewhat like playing whack-a-mole with atomic structures,\u201d remarked Hemant Mohapatra, the Lightspeed partner who spearheaded this funding round, in conversation with TechCrunch. \u201cA material is only practical in the real world if all properties align simultaneously, which adds intrigue to this search problem.\u201d<\/p>\n

Mohapatra anticipates that the market for predicting new substances will become commoditized as models advance. The distinction with Discovered Materials lies in Ramdas\u2019 extensive experience in the domain and the capability to operate a lab capable of swiftly experimenting and validating potential candidates \u2014 a feat he asserts both founders have already accomplished with multiple new materials.<\/p>\n

Once valuable candidates are identified, Sridhar mentions the company will seek to patent the application of these materials in GPUs, or the method of creating chips from the substances, licensing them to chip manufacturers. He aspires to have patent-worthy new materials within the year.<\/p>\n

However, amidst all the enthusiasm, there has yet to be any drugs or materials discovered by AI that have made a significant commercial impact. The closest instance might be Insilico Medicine\u2019s Renterosib, the inaugural drug found with generative AI that has progressed to a Phase II clinical trial. In terms of materials, promising contenders like MatNex\u2019s rare-earth-free permanent magnets and new semiconductor materials developed by Panasonic and Citrine Informatics have been identified, but they have not yet been deployed commercially at scale.<\/p>\n

These approaches may be maturing as AI technology progresses, yet Mohapatra contends that finding more candidates is not the bottleneck in AI materials science; rather, \u201cthe challenge lies in accurately filtering and synthesizing them.\u201d<\/p>\n

While Sridhar believes Discovered Materials\u2019 unique data and insights will enable the startup to compete with well-funded advanced labs, he acknowledged that much of the work will involve actually going into wet labs to create materials. This is a phase of the process that cannot be expedited.<\/p>\n<\/div>\n

When you purchase through links in our articles, we may earn a small commission. This doesn\u2019t affect our editorial independence.<\/em><\/p>\n","protected":false},"author":2,"featured_media":3491598,"comment_status":"open","ping_status":"closed","sticky":false,"template":"Default","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-3491597","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/posts\/3491597"}],"collection":[{"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/comments?post=3491597"}],"version-history":[{"count":0,"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/posts\/3491597\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/media\/3491598"}],"wp:attachment":[{"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/media?parent=3491597"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/categories?post=3491597"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/techingeek.com\/index.php\/wp-json\/wp\/v2\/tags?post=3491597"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}