Frank Louthan

Mgmt said Akamai is well positioned for power and capacity given backbone flexibility (can site in cheaper-power states) and acts as an anchor tenant for 5-20MW builds.

Ed confirmed the entire $1.8B comes in as revenue with no CapEx offset. Tom described the Inference Cloud footprint: functions as a service across all 4,300 locations, managed container service in well over 100 cities, full IaaS in several dozen cities, and a couple dozen equipped with new 6000-series GPUs. Vision is to orchestrate workloads across the right compute resource closest to the user, aligned with NVIDIA's AI grid concept.

Tom said they're seeing longer commitments across all services, partly by design and partly driven by customer interest. Delivery growth expected at mid-single-digit decline this year. Ed added that RPO is growing significantly due to longer-term commitments. Sales force is incentivized for longer commits. No major dynamic change in delivery committed vs uncommitted mix, but security and compute driving bigger and longer commitments.

Ed described normal Q3 seasonal softness pattern. Traffic growth is the biggest driver, with pricing declines moderating consistently throughout the year. Longer contracts do not elongate sales cycles with existing customers.

Ed explained other cloud applications includes image/video management, legacy NetStorage (~$50M annualized being end-of-lifed), and cloudlets. Some being transferred to compute partners in exchange for CIS commitments. Growing 3-4% but expected to decline over time.

Ed quantified TikTok U.S. revenue at $40-50M annually, about half included in back-half guidance. But guidance midpoint is up much more than that, driven by strength across the board.

Tom said about one-third complete. Shifting toward hunting, hiring specialists, incentivizing longer-term contracts. Ed outlined three margin drivers: compute scale-up, security mix shift, and delivery stabilization.

Ed noted Edgio was the biggest driver, adding a few hundred logos with over 100 new to Akamai. Organic growth strongest in Guardicore, API security, and compute.

Ed confirmed migration is complete, no significant churn expected. They were selective on which Edgio contracts they acquired. Upsell opportunity exists.

CFO answered in one sentence with a specific modelling assumption, directly addressing how much of the raised revenue guide is non recurring.

Adaire said no capital difference expected as metro-focused strategy is already embedded. Phillip added that new facilities are being built toward higher densities suited for customer requirements.

Management confirmed mid-20% return targets are maintained even at higher densities, consistent with long-standing underwriting standards.

Adaire confirmed 3 GW of powered developable land under control, building only against powered land, and AI workloads already 33% more dense than non-AI. Keith noted 52 energized retail projects underway.

Adaire attributed interconnection strength to Americas-led demand from technology and infrastructure providers and clouds, driven by customer mix rather than power densities.

Keith clarified Equinix does capitalize interest but historically less due to lower debt costs and fewer assets under development; expects this to increase but declined to quantify the forward impact.

Wright described the 870+ acre contiguous Atlanta parcel for a gigawatt campus with power alternatives being explored, plus additional Hillsboro land for 160MW.

Power sees no significant competitive unleased capacity in 2027-2028, with DLR taking more development risk by greenlighting shells and suites before customer lease signing.

Power confirmed taking market share with AI at 18%+ of 0-1MW signings (up from single digits), deal sizes growing, and inference still beyond nascency in enterprise.

Power saw two positives: streamlining domestic data center permitting and promoting US AI tech stack exports, both benefiting Digital Realty's global platform.

Management is increasingly building dedicated buildings for enterprise customers on campuses, targeting 22%+ growth in this category with further acceleration planned.

Chris said price locks here to stay as competitive tool giving customers comfort to switch. On Cox mobile penetration: expects faster ramp than original Charter curve given better platforms, sales channels, marketing, and national brand awareness. Improvement to original curve is good starting point.

Srini outlined structural advantages: 34M trusted account relationships, TLife with 24M monthly actives, advantage credit decisioning from subprime/near-prime capabilities, and dramatically lower customer acquisition costs; André added greenfield advantage with no legacy backbone to protect.

Hans said the FWA deployment pace is unchanged, Verizon is still gearing up to cover 80-90% of planned C-band sites this year, with nearly all basic C-band layer complete by next year. Capital is being allocated primarily to mobility because where C-band is deployed Verizon sees better step-ups and customer upgrades. Hans said CapEx efficiency is a separate win, the network team is executing all 2025 plans more efficiently within budget, which is important heading into the Frontier integration. He declined to discuss July promotional dynamics or forward competitive positioning specifically.