The chip industry has a materials problem. Building faster, denser semiconductors takes rare minerals and enormous amounts of energy, and supply has not kept pace with AI driven demand. That gap has become one of the defining stories of the AI boom this year, showing up in everything from earnings calls to stock swings.

That crunch has already rattled markets once this year. Now a small Cambridge startup says artificial intelligence itself might be the fix, and some very well known names just wrote it a very large check to find out if that bet pays off.

Wall Street is watching the chip material crunch

Severe shortages of semiconductors, especially memory chips, hit share prices across the tech sector hard last week. DRAM and NAND prices have surged for months as hyperscale cloud providers lock up production capacity years in advance, leaving less for consumer electronics and automakers.

Micron Technology has become the clearest example of what that scarcity does to a stock. Shares went parabolic earlier this year, rising nearly 40% in five sessions as the memory shortage worsened, and the company has since raised its capital spending plans by billions of dollars to keep up with orders.

Not everyone is convinced the AI trade can keep running at this pace. Michael Burry, the investor famous for calling the 2008 housing crash, has built a basket of bearish bets that includes Nvidia and other chip-adjacent names, arguing valuations have gotten ahead of themselves.

Jensen Huang,  Nvidia CEO, sees it differently. He told reporters in Seoul in June that investors should treat any AI selloff as a buying opportunity, and he has called AI infrastructure a foregone conclusion rather than a bubble waiting to pop. Behind both arguments sits a simpler fact: chipmaking runs on a narrow set of physical inputs, and several of them, like ruthenium and iridium, come from a handful of countries with fragile supply chains.

CuspAI’s $450 million bet on new materials

CuspAI, a two-year-old British startup, has raised $450 million in Series B financing, giving it a $2.6 billion valuation. That is up from just $520 million last September, a nearly five-fold jump in under a year, according to the company.

Kleiner Perkins and NEA led the round. Jeff Bezos’s fund, Bezos Expeditions, took part with what CuspAI called significant participation, alongside newer backers like Glade Brook Capital Partners, Lux Capital, AMD Ventures and Britain’s own Sovereign AI Venture Fund. The round follows a Series A of more than $100 million less than a year ago. Bloomberg previously reported elements of the deal before it was confirmed.

Alongside the raise, CuspAI launched the AI Materials Foundry on Monday, July 20, 2026. The coalition brings together more than 48 technology and industrial partners, including Nvidia, Meta Platforms, Samsung and Hyundai Motor Group, to pool computing power, laboratory access and scientific expertise for materials research spanning the United States, Asia Pacific and Europe.

Chad Edwards, CuspAI’s co-founder and chief executive, said 80% of the company’s research effort this year will go toward material discovery, with a specific goal of cutting the use of scarce metals like ruthenium and iridium. Chipmakers, he said, are “frantically searching for new materials.” He added that much of the new funding will go toward supporting labs with foundry partners in Cambridge, Singapore and the San Francisco Bay Area. Nvidia separately confirmed it is teaming up with CuspAI to hunt for breakthroughs across semiconductors, clean energy and advanced manufacturing.

Bezos doubles down on physical AI

This is not Bezos’s only large bet on AI. He is also co-CEO of Prometheus, the physical AI startup that raised $12 billion at a $41 billion valuation in June. Bezos has been vocal about how such AI tools could reshape manufacturing and even labor markets.

Bezos has reportedly been exploring a fund of up to $100 billion to buy manufacturers outright and rebuild them around AI. Chipmaking is one of the sectors named as a target for that broader strategy.

CuspAI fits a similar pattern, though on a smaller scale. The company started out searching for materials to capture carbon and purify water, using AI to sift through a library of 300 trillion possible chemical structures before narrowing candidates down to a workable handful of ten. It pivoted toward chipmaking over the past year once demand from the semiconductor supply chain, whose products are critical to powering AI, Edwards said as reported on Yahoo.

Google DeepMind has said it will soon start its own work on AI driven material discovery, and investors have already backed several teams led by former OpenAI and DeepMind researchers chasing the same idea. The field is getting crowded fast.

What this means for the AI trade

CuspAI is a small piece of a much bigger story about whether AI can actually solve the physical bottlenecks slowing down its own growth. Chips need new materials, memory is scarce, and energy costs keep climbing. No single funding round changes that overnight.

If CuspAI’s approach works even partially, it could ease some of the scarcity that has been driving swings in chip stocks all year and give chipmakers more room to plan around AI demand instead of chasing it. If it does not, it becomes one more example of AI hype racing ahead of lab results.

Either way, the roster of names now involved, from Nvidia to Bezos to Kleiner Perkins, suggests the biggest players in tech are not waiting around to find out. They are placing their bets now, while the shortage is still headline news.

Benzinga Disclaimer: This article is from an unpaid external contributor. It does not represent Benzinga’s reporting and has not been edited for content or accuracy.

​