The artificial intelligence (AI) playbook is already familiar: build a larger GPU cluster. Add more Blackwell chips. Throw more electricity at the problem. If the chips get hot, build the data center next to a river. If you run out of bandwidth, put in more copper.
This is how Amazon, Alphabet, microsoftand Metaplatforms They’re solving AI in 2026. And it works, until it runs into physics.
Will AI create the world’s first billionaire? Our team just published a report on a little-known company called “Indispensable Monopoly” that provides critical technology that both Nvidia and Intel need. Continue “
A company looked at the same problem and came up with a different answer. Global Foundries(NASDAQ:GFS) is betting that the real bottleneck in AI infrastructure is not computing power. It is the wire that connects the chips and replaces that wire with light.
The copper wall that no one talks about
Inside each AI data center, thousands of chips must share information at enormous speeds. Right now, most of that communication travels over copper, which is running out of space. It generates heat, loses signal at a distance, and consumes energy in ways that become painful at scale. Every time an AI model grows, the copper problem gets worse.
The industry has known this for years. The solution has a name: co-packaged optics (CPO). The idea is to move optical transceivers, components that transmit data through light instead of electricity, directly next to the chip, reducing to almost nothing the distance that data has to travel over copper. The result is a faster, cooler, and more energy-efficient AI infrastructure.
In May 2026, GlobalFoundries announced SCALE (Silicon Photonics Co-Packaged Advanced Light Engine Solution), the industry’s first platform that meets the Optical Computing Interconnection Multi-Source Agreement specifications for AI scaling architectures. The platform uses thick, dense wavelength division multiplexing (DWDM) over each optical fiber to drive bandwidth density and scalability beyond what copper can do, and GlobalFoundries has already demonstrated 8λ and 16λ bidirectional DWDM natively on its platform, a milestone the company describes as critical to everything that follows.
Image source: Getty Images.
The part of the stack that everyone is chasing
Here’s the thing about silicon photonics getting lost in GPU coverage: it’s as much a manufacturing problem as it is a physics problem. Designing a silicon photonic chip is difficult. Building it at scale, with the precision needed for fiber optic alignment, in volume, for hyperscale data centers is more difficult.
GlobalFoundries has spent years developing the process technology to do exactly that. Its silicon photonics platform supports 50 Gbps and 100 Gbps microring modulators, wideband detachable fiber interfaces, and flat insertion loss features future-proofing the platform as wavelength counts increase. In November 2025, the company acquired Advanced Micro Foundry in Singapore, a specialized silicon photonics manufacturer, adding manufacturing assets, intellectual property and engineering depth that would take years to build from scratch.
That acquisition gave GlobalFoundries silicon photonics production capacity in Singapore, a geography that is important for supply chain diversification amid heightened tensions over semiconductors between the United States and China. The company is building a platform that hyperscalers need and that very few manufacturers can offer.
GlobalFoundries fell 10% in one day, but I’m not worried
GlobalFoundries fell almost 10% on May 27, dragged down by the alleged sale of Mubadala shares. Mubadala is Abu Dhabi’s sovereign wealth fund and was previously the majority shareholder of GlobalFoundries. Despite the stock’s decline, the Motley Fool’s long-term approach wins here.
The history of GlobalFoundries is intact. SCALE’s announcement sent shares up 12% in a single session just a few weeks ago. An improvement in Q1 2026 earnings followed. Silicon photonics revenue is expected to nearly double again in 2026, with more than 500 successful designs recorded in 2025 and growing momentum. A liquidation does not change those facts.
Additionally, the U.S. government is supporting GlobalFoundries with a proposed $375 million award to help build a national quantum manufacturing infrastructure.
Every major hyperscaler is wondering how to train larger models faster. GlobalFoundries asks a different question: How do you move data between chips without the infrastructure melting?
Packaged optics are the answer. The company that builds the manufacturing platform to deliver at scale is still, most days, classified as a “semiconductor foundry.” On days when it drops 10% for reasons unrelated to its most important business, that drop becomes an opportunity.
Should You Buy GlobalFoundries Stock Right Now?
Before you buy shares in GlobalFoundries, consider this:
He Varied and Dumb Stock Advisor The analyst team has just identified what they believe are the 10 best stocks for investors to buy now… and GlobalFoundries was not one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when netflix made this list on December 17, 2004… if you invested $1,000 at the time of our recommendation, you would have $463,900!* Or when NVIDIA made this list on April 15, 2005… if you invested $1,000 at the time of our recommendation, you would have $1,294,401!*
Now, it is worth noting stock advisor The total average return is 978.%: An overwhelming market outperformance compared to the S&P 500’s 211%. Don’t miss the latest Top 10 list, available with Stock Advisorand join an investing community created by individual investors for individual investors.
See the 10 actions »
*Stock Advisor returns as of May 30, 2026.
Micah Zimmerman has no position in any of the stocks mentioned. The Motley Fool ranks and recommends Alphabet, Amazon, GlobalFoundries, Meta Platforms, and Microsoft. The Motley Fool has a disclosure policy.
All big tech companies are solving AI the same way. This action is solving it differently. was originally published by The Motley Fool
