Cogent Communications (CCOI) TD Cowen 12th Annual Communications Infrastructure Summit summary

Financial performance and strategic actions

• Addressed investor concerns over $750M unsecured debt maturing, with plans to refinance as secured debt and reduce the amount raised.

• Revenue has declined for 12 quarters due to the Sprint acquisition, but EBITDA has grown in 11 of those quarters, with margins expanding by over 2,000 basis points, mainly through cost cutting.

• Sprint's negative EBITDA margin products were purged, reducing their revenue share from 42% to 15% of the total.

• Integration spending is set to end by 2026, with headcount reductions providing a $50M EBITDA tailwind for next year.

• Transfer payments from T-Mobile, which boost reported EBITDA, will end in February 2028.

Capital expenditures and network strategy

• CapEx declined sequentially to $38.5M in 2Q, with further moderation expected as network consolidation and data center sales progress.

• Surge in CapEx was due to converting telephone locations to data centers, reconfiguring the Sprint network, and equipment price increases.

• Equipment prices have risen sharply due to supply chain constraints, with delivery times now up to 24 months; long-term CapEx is expected to normalize at $25M/quarter once issues resolve.

• On-net building count growth is focused on carrier-neutral data centers, while multi-tenant office footprint growth has slowed due to high vacancy and office-to-residential conversions.

• The company remains disciplined in connecting only large, high-return buildings, exiting less attractive properties.

Wavelength business and market outlook

• Wavelength revenue grew 9.2% sequentially and 68% year-over-year, with 3% North American intercity market share achieved in six quarters.

• Targeting 25% market share in wavelengths, with growth expected to accelerate as credibility builds and AI-driven demand increases.

• Backlog remains strong, with demand from regional networks, content distributors, international carriers, hyperscalers, and neoclouds.

• Market growth is expected as AI training facilities come online and supply constraints ease.