In August 2026, Prysmian announced a $1.25 billion investment across North Carolina, Tennessee and South Carolina. The plan includes $1.02 billion for fiber operations in Claremont, North Carolina, alongside expanded optical-cable capacity and up to 600 jobs. Crucially, the program includes upstream glass-preform and fiber-drawing stages as well as final cable manufacturing.

One clarification is essential: the announcement says Prysmian will more than double its own U.S. optical-fiber production capacity. It does not state that total U.S. manufacturing capacity will double. Even so, the investment’s scale and vertical scope indicate how procurement conditions are changing.

Prysmian’s Investment Is Backed by Contracted Demand

The expansion is notable because it is linked to long-term commercial commitments, not a broad market forecast alone. In July 2026, Prysmian announced an agreement worth up to €5.5 billion over as many as ten years to supply Molex with optical cables for data-center deployment. It also said its agreements and initiatives with hyperscalers and data-center infrastructure providers could generate more than €10 billion in incremental cumulative revenue through 2035.

The mix spans high-density data-center products, outside-plant cable, broadband distribution cable and drop cable. Demand is therefore not concentrated in one cable category.

AI Data Centers Are Changing the Fiber Demand Curve

Traditional forecasts often begin with homes passed or subscriber growth. AI infrastructure has a different profile: dense connectivity inside data halls is followed by building-to-building routes, carrier-hotel connections, metro rings, diverse long-haul paths and links to power-rich sites outside established hubs.

Research from the Fiber Broadband Association and RVA estimated that a threefold increase in U.S. hyperscale data-center capacity by 2029 could require twice as many fiber route miles, 2.3 times as many total fiber miles, and an average of 135 new route miles for each new hyperscale data center. These are forecasts rather than guaranteed purchases, but they explain why manufacturers are investing across multiple product layers.

• Inside the data center: higher fiber density, more parallel links and faster architecture refreshes.
• Across the campus: routes among data halls, substations, meet-me rooms and network buildings.
• Between facilities: metro and regional interconnection routes.
• Beyond established hubs: middle-mile and long-haul paths to new locations.

One program may need loose-tube cable for an outside-plant route, high-fiber-count ribbon cable for constrained ducts, indoor-rated cable for data halls and pre-terminated assemblies for fast deployment. That breadth changes supplier qualification.

BEAD Adds Demand, but BABA Changes Who Can Supply It

The Broadband Equity, Access, and Deployment (BEAD) Program provides $42.45 billion for high-speed internet expansion. It is not the entire U.S. fiber market: private capital, telecom operators, utilities, cooperatives, municipalities and data-center developers continue to fund substantial deployment outside the program.

The procurement issue is BEAD’s interaction with the Build America, Buy America Act (BABA). Under the BEAD-specific waiver framework, optical fiber and fiber-optic cable remain strategically protected construction materials. For compliant cable, processes such as ribboning, buffering, stranding and jacketing must occur in the United States; optical-fiber requirements reach further upstream into preform fabrication and fiber drawing.

In March 2026, AFL, Corning, Lightera and Prysmian stated that U.S. manufacturers produced approximately 135 million fiber-kilometers annually and estimated peak BEAD demand at less than 5% of that capacity. Buyers should treat this as a supply-side position rather than an independent capacity audit.

The practical conclusion is nuanced: BEAD creates a compliance-defined demand pool, while AI data centers and privately financed networks may be the larger long-term reason for wider capacity expansion. Domestic cable jacketing alone is not equivalent to a fully domestic optical-fiber supply chain.

Local Manufacturing Does Not Automatically Mean Shorter Lead Times

More domestic capacity can improve supply resilience, but it does not mean every cable design will immediately become cheaper or faster. The supply chain extends from high-purity materials and glass preform through fiber draw, coloring or ribboning, buffering or stranding, jacketing or armoring, testing and logistics.

Expansion in preform and draw stages reduces dependence on imported fiber and improves upstream control. Local plants can respond more quickly to project changes and reduce transoceanic shipping exposure. Yet lead time still depends on whether the required design is on the new lines, qualification and ramp-up schedules, long-term contract allocation, specialty inputs, changeovers, minimum runs and regional logistics.

Capacity is not the same as available inventory. A factory can have substantial annual output while still offering a long lead time for a low-volume, nonstandard cable.

Will More U.S. Capacity Reduce Fiber Cable Prices?

Additional capacity can reduce scarcity premiums and improve competition, but it does not guarantee a lower total project cost. The Fiber Broadband Association’s 2026 deployment-cost report found that 92% of surveyed respondents experienced cost increases in 2025. Median underground deployment cost reached $18 per foot and median aerial cost reached $8 per foot; labor represented 72% of underground and 64% of aerial cost.

Cable price is only one component. A lower-priced cable can cost more if it slows pulling or blowing, increases splicing labor, fits poorly with ducts or closures, creates scrap, delays documents, fails acceptance tests or misses a construction window.

Total landed and installed cost = cable price + freight + duty + inventory + installation labor + testing + schedule risk + compliance risk.

Why Second-Source Planning Becomes More Important—not Less

A larger primary supplier can be attractive, but strategic projects still face concentration risk. Long-term data-center agreements may reserve capacity; BEAD projects may receive scheduling priority; and a regional event, raw-material issue or qualification delay can affect multiple plants. Suppliers sharing the same fiber, compound or armor input may not provide genuine supply-chain diversity.

The best time to approve a second source is during design and pilot deployment—not after a primary supplier misses delivery.

What Global Cable Buyers Should Change in Their RFQs

1. Identify funding before supplier selection. Confirm whether the project uses BEAD, another federal award, state or municipal funds, or private capital.
2. Separate fiber origin from cable origin. Request preform, fiber-drawing and cable-manufacturing locations with the applicable evidence.
3. Lock the technical design before comparing price. Confirm fiber standard and count, construction, jacket, tensile and crush performance, bend radius, attenuation, temperature, installation method, reel length and approvals.
4. Request allocation and lead-time statements. Separate normal production timing from committed project capacity; use rolling forecasts and reservation mechanisms for multi-year builds.
5. Price alternative architectures. Compare loose-tube, ribbon, rollable-ribbon, microcable and pre-terminated options where appropriate.
6. Qualify the second source early. Schedule sample testing, closure compatibility, pulling or blowing trials and documentation review before construction.

Where International Suppliers Still Fit in the U.S. Market

Domestic expansion does not eliminate international suppliers; it makes market selection more important. BABA applies to covered federally assisted infrastructure, not every U.S. project. Private data centers, enterprise campuses and privately financed builds may source internationally, subject to customer requirements, tariffs, trade restrictions and technical approval.

International manufacturers can also remain relevant for specialized or low-volume designs, hybrid cable, specialty jackets, custom breakout constructions, short runs and project-specific packaging when no domestic standard product fits. They can serve OEM and private-label programs, connectivity hardware and complementary product lines, as well as emergency or overflow capacity when the project permits imported material and an alternate design has already been approved.

The Market Is Splitting into Three Procurement Lanes

1. Compliance-led procurement: BEAD and other covered projects, where domestic manufacturing evidence is a gate to entry.
2. Performance- and schedule-led procurement: AI data centers and strategic routes, where allocation, density, deployment speed and reliability dominate.
3. Value- and customization-led procurement: private enterprise, regional networks, specialty applications and OEM programs where qualified international supply can remain competitive.

Prysmian’s investment touches all three lanes, especially the first two: it adds upstream resilience for BABA-sensitive work and large-scale capacity for AI infrastructure while spanning broadband and high-density data-center products.

Conclusion

Prysmian’s U.S. expansion is evidence that optical fiber has become strategic infrastructure. More capacity should improve long-term availability, but it will not remove price volatility, allocation risk or design-specific shortages. The stronger procurement strategy is to segment projects accurately, calculate total installed cost, confirm manufacturing origin, reserve capacity and qualify a technically equivalent second source before it is needed.

Frequently Asked Questions

References

• Prysmian: $1.25 billion U.S. manufacturing investment
• Prysmian: data-center agreements and capacity expansion
• NTIA: demonstrating compliance with the Buy America requirement
• U.S. Department of Commerce: Build America, Buy America resources
• Fiber Broadband Association: fiber requirements for AI and data-center growth
• Fiber Broadband Association: 2026 fiber deployment-cost findings
• Corning: U.S. manufacturers’ BEAD supply commitment