Cogent Communications (CCOI) Oppenheimer 29th Annual Technology, Internet & Communications Conference summary

AI infrastructure demand and industry transformation

• AI infrastructure demand is surging, driving accelerated internet traffic growth and changing traffic directionality as more data is uploaded for AI processing.

• Significant capital is being deployed into AI, with $1 trillion spent and $6 trillion more announced, but shortages in power, data center space, and processing persist.

• AI is expected to transform society and business, making processes more efficient and reducing the need for traditional software and professional services.

• The value from AI will likely accrue at the application layer or to consumers, with service providers facing challenges in monetizing infrastructure investments.

• Professional services, software, and medical industries are at risk of disruption as AI automates and democratizes expertise.

Business impact and growth outlook

• Recent years have been challenging due to revenue declines from acquired businesses, increased capital intensity, and higher leverage, but underlying trends are expected to drive a return to top-line growth.

• The wavelength business is growing rapidly (68% year-over-year), but remains a small portion of total revenue due to delays in data center and infrastructure readiness.

• Market share gains in the wavelength segment are constrained by supply chain and infrastructure bottlenecks, but the company has established a foothold with over 548 customers.

• Bottlenecks in GPUs, memory, power, and data center capacity are expected to persist for two to three years, with gradual resolution as capital and innovation address constraints.

• Internet traffic growth is expected to continue as new applications like AI drive demand, with prices for bandwidth likely to fall as supply chain issues are resolved.

Monetization and industry risks

• Monetizing AI infrastructure remains unclear, with much of the value likely to be embedded in other services rather than directly captured by infrastructure providers.

• The software industry is expected to be fundamentally changed, with AI reducing the need for centrally codified software and lowering the percentage of GDP spent on software development.

• The innovation cycle is expected to invert, with smaller businesses and end users driving new technology adoption rather than large corporations or governments.

• Despite current undervaluation and operational challenges, the business is positioned to benefit from long-term industry trends as bottlenecks ease and demand materializes.