What Nokia's Manufacturing Expansion Means for AI Networks

A quiet corner of Pennsylvania's Lehigh Valley is about to become a significant address in US tech.

Nokia has announced a major expansion of its advanced test and packaging (ATP) operations in Allentown, Pennsylvania.

It will increase domestic production capacity of the optical technologies that underpin both AI and telco networks.

The investment is expected to increase manufacturing capacity at the site by up to ten times its current level, with new production expected to be available by the end of the third quarter.

Nokia also expects the expansion to nearly double its Pennsylvania workforce to more than 500 employees across engineering, manufacturing and R&D.

The connectivity challenge behind AI

The announcement reflects a growing recognition that AI infrastructure requires far more than advanced processors alone.

As AI models become larger and workloads more distributed, network performance has become a core consideration.

Optical technologies play an important role in enabling the fast, low-latency connections required between compute clusters and data centres.

Nokia's Allentown facility is one of only a small number in the US capable of carrying out ATP processes for photonic chips that are integrated into optical modules used across AI and telco infrastructure.

Justin Hotard, President and CEO of Nokia, said: “The AI supercycle is fundamentally reshaping network and infrastructure requirements in the US and globally.

“Our expansion in Allentown is a direct investment in that future – scaling domestic manufacturing of the optical networking technologies that power AI infrastructure.

“It also reflects the strong partnership between Nokia, the United States, and the Commonwealth of Pennsylvania to support advanced manufacturing, create jobs and strengthen US technology leadership and global competitiveness.”

The company says its optical networking technologies can reduce energy consumption in communications networks by as much as 75% as AI infrastructure pushes higher power demands.

A push for domestic production

Today, less than 2% of global semiconductor ATP activity takes place on American soil. Nokia wants to change that.

By increasing domestic production of photonic components, Nokia is aiming to support more resilient supply chains for critical communications infrastructure.

The project includes approximately US$30m in investment from Nokia, supported by around US$4m in assistance from the Commonwealth of Pennsylvania and approximately US$10m through the federal CHIPS investment tax credit.

Governor Josh Shapiro said: “Nokia is doubling down on the Lehigh Valley and ensuring that the future of chip production continues to run through this region because we’ve made smart investments to make Pennsylvania more competitive and proven that our Commonwealth is a great place to do business.

“As demand for semiconductors continues to grow across industries, we’ll continue to position Pennsylvania as a leader in innovation, with a supportive, thriving business climate that helps companies compete on a global scale.

“From advanced manufacturing to the research and development of new technology like advanced chip packaging, Pennsylvania has all the resources to be a world leader in chip production.”

Supporting AI-ready telecom networks

As AI applications become embedded within enterprise and consumer services, operators face growing pressure to support higher traffic volumes and more demanding performance requirements.

Nokia's investment forms part of a broader multi-year plan to invest US$4bn in US research, development and manufacturing focused on AI-ready network connectivity.

Bill Frauenhofer, Executive Director of Semiconductor Investment and Innovation at the Department of Commerce, said: “Nokia’s investment in Pennsylvania is directly advancing America’s AI leadership.

“Supported by CHIPS and Science Act funding, Nokia is deepening its commitment to innovation and the production of photonic chips in the United States.

“This project enables critical optical technology and strengthens America’s semiconductor supply chain.”

As AI infrastructure expands, the networks connecting data centres and end users are as vital as the computing power inside them.

The real constraint on AI performance may be the optical plumbing that ties everything together, and right now, most of that plumbing is made elsewhere.

Nokia is making a bet that it doesn't have to stay that way.