Akamai Technologies (AKAM) Goldman Sachs Communacopia + Technology Conference 2026 summary

Strategic transformation and AI compute expansion

• Executives highlighted a major inflection point, leveraging decades of network investment to build a next-generation AI compute network, integrating Linode assets and expanding large-scale CPU and GPU deployments globally.

• The company has signed over $2.8 billion in multi-year cloud infrastructure commitments, with a robust pipeline and accelerating growth expected into 2027 and 2028.

• Operational focus is on securing colocation and power, with capital readily available and supply chain management tightly linked to customer demand and contract timing.

• Revenue from new deals is expected to ramp significantly from Q4 2026, with $15–20 million per major deal and continued acceleration into 2027.

• Contracts are structured with escalators to hedge against rising costs, and take-or-pay terms ensure revenue predictability despite potential delays.

Financial model, margins, and capital allocation

• CapEx to revenue conversion for cloud infrastructure services (CIS) is generally $0.50–$1 per $1 spent, with operating margins for large deals in the mid-20% to mid-30% range.

• Revenue yield per megawatt typically ranges from $15–$25 million, with higher yields possible for certain CPU deals or smaller customers.

• The company aims to fund growth primarily through internal cash flow, with strong EBITDA and free cash flow margins expected as CIS scales.

• Maintaining investment-grade credit is a priority for favorable colocation terms, but temporary downgrades are manageable.

Technology, customer journey, and market positioning

• Modern hardware and chipsets are deployed for new compute locations, with no need to overhaul legacy infrastructure; expansion is achieved by adding new, larger sites.

• Customer relationships often begin with large compute deals and expand into cross-selling security and delivery services, leveraging existing enterprise trust.

• The distributed network model enables both centralized and edge deployments, with edge inference use cases growing, especially where latency is critical.

• The company’s unique approach of aggregating colocation across many providers creates a scalable, flexible infrastructure advantage.