Akamai Technologies, Inc. (AKAM) Earnings

Management highlights

### Strategic Positioning for AI-driven Infrastructure - Akamai leverages its 30+ year legacy global distributed edge platform (deployed across 700+ cities in 130 countries) to become a key infrastructure provider for the AI economy, focusing on distributed AI inference that runs close to end-users rather than centralized training of large foundation models. - The company has secured $2.8 billion in total multiyear commitments for cloud infrastructure services (CIS) in 2026 to date, highlighted by a new 4-year $600 million commitment from a U.S. robotics technology company. ### AI-driven Product Momentum - **Cloud Infrastructure Services (CIS):** Akamai's distributed model delivers improved performance and lower costs for customers compared to centralized hyperscalers, with demonstrated customer wins across AI inference, simulation, analytics, and media use cases that have beat competing hyperscaler offerings. All current GPU capacity is fully sold out due to strong demand. - **Security Portfolio:** AI adoption has created cross-portfolio tailwinds; Q2 security revenue grew 10% YoY, led by strong demand for web application firewall (WAF), API security, and Guardicore segmentation. Recent wins include a $14 million WAF expansion from a top global bank, a $20 million 2-year renewal from a large global telco, and a new customer win with cybersecurity leader CrowdStrike, which switched from a competitor due to inconsistent service. Akamai expects full-year 2026 security revenue to exceed $2.4 billion, making it one of the largest security providers globally. - The July 2026 acquisition of LayerX (rebranded as Akamai Workforce Protector) adds secure enterprise browser and AI usage control capabilities, enabling enterprises to monitor and prevent sensitive data leaks from employee use of public AI tools. This capability integrates seamlessly with Akamai's existing Zero Trust portfolio to deliver a unified workforce security solution for AI use cases. ### Operational and Financial Updates - Non-GAAP operating margin was 25% in Q2, in line with expectations, as the company invests in colocation expansion, depreciation, and headcount to support CIS growth. Q2 capital expenditures (CapEx) were $347 million (32% of revenue), below guidance due to delayed GPU shipments that pushed planned spending into Q3 2026. - Akamai raised $3.5 billion in zero-coupon convertible debt in May 2026 to fund CIS growth. The company has temporarily paused share repurchases to reallocate capital to the high-growth CIS pipeline; as of Q2 end, $565 million remained on the existing repurchase authorization, and the company held $4.6 billion in cash, cash equivalents, and marketable securities.

Guidance

- **Q3 2026 Guidance:** - Total revenue is projected between $1.105 billion and $1.130 billion, representing 5% to 7% YoY growth as reported, and 5% to 8% YoY growth in constant currency. Foreign exchange is expected to create an $8 million YoY negative headwind. - Non-GAAP operating margin is expected between 24% and 26%, with non-GAAP diluted EPS in the range of $1.60 to $1.80. - CapEx is projected between $475 million and $525 million (43% to 46% of Q3 revenue), driven by delayed Q2 GPU shipments and build-out for newly signed large contracts. The company plans to invest up to an additional $500 million to expand GPU capacity, with ~$60 million of this investment occurring in 2026 and the remainder in early 2027. - **Full-Year 2026 Guidance:** - Total revenue is projected between $4.445 billion and $4.530 billion, representing 6% to 8% YoY growth as reported, and 5% to 7% YoY growth in constant currency. - CIS revenue is expected to grow at least 50% YoY in constant currency; security revenue is expected to grow in the high single digits YoY in constant currency; delivery and other cloud applications are expected to decline in the mid single digits YoY in constant currency. - Non-GAAP operating margin is expected to be between 25% and 26%, with full-year CapEx projected to equal ~40% of total 2026 revenue. Non-GAAP diluted EPS is expected between $6.40 and $7.05. - **Long-Term Growth Guidance:** - Management confirmed that driven by the strong multiyear CIS commitment backlog and robust customer pipeline, total revenue growth is expected to accelerate to the low teens in 2027, up from mid-single digit growth in 2026. CIS revenue growth is expected to meaningfully accelerate in Q4 2026, with further acceleration in 2027. - Large multiyear CIS contracts typically deliver non-GAAP cash gross margins of 65% to 75%, and non-GAAP operating margins ranging from the low-20s to low-30s over the life of the contract.

Segment performance

Total Q2 2026 revenue was $1.1 billion, up 5% year-over-year (YoY) both as reported and in constant currency. International revenue was $549 million, up 6% YoY (up 7% in constant currency), representing 50% of total Q2 revenue. - Cloud Infrastructure Services (CIS): Revenue was $99 million, up 39% YoY both as reported and in constant currency. This accounts for approximately 9% of total Q2 revenue. - Security: Revenue was $595 million (corrected from transcript typo of $4 million), up 10% YoY as reported and 9% YoY in constant currency. This accounts for approximately 54% of total Q2 revenue. - Delivery and other cloud applications: Revenue was $396 million, down 6% YoY as reported and 5% YoY in constant currency. This accounts for approximately 36% of total Q2 revenue.

Risks & headwinds

- Foreign exchange fluctuations create headwinds for 2026 second half revenue, with an expected $9 million negative impact, down from the prior expectation of a $20 million positive impact due to recent U.S. dollar strength. - GPU and data center component pricing can fluctuate, and supply chain delays can push CapEx and revenue recognition into future quarters, as seen with delayed Q2 2026 GPU shipments. - Large CIS deals require upfront capital investment for data center capacity, power, and hardware before revenue recognition, which can create temporary margin pressure in early quarters of a contract ramp. - Power and data center capacity constraints could limit growth if the company is unable to secure sufficient capacity at attractive costs to meet strong customer demand. - The execution risk associated with ramping up large contracted capacity on schedule to meet customer commitments, which could impact revenue ramp timelines. - General macroeconomic uncertainty and geopolitical developments could impact customer demand and business performance, as noted in the company's SEC filings.

Analyst Q&A

• Q: Given all current GPU capacity is sold out after recent large contract wins, will Akamai be unable to sign additional new CIS deals in the near term? What is the state of the current CIS pipeline? /

  A: Management reports the CIS pipeline remains very strong across all industries and deal sizes. Akamai is already ordering additional GPU hardware to meet ongoing demand, with a typical 6 to 9 month lead time for large deals, so the sold-out status does not prevent the company from continuing to sign new customers. Customers are often pre-reserving capacity months in advance, similar to reserve instance models in the public cloud market, allowing Akamai to match capacity build to committed demand. (307 characters)

• Q: What is the typical timeline between signing a large multiyear CIS contract and recognizing revenue, and can the margin profile of these deals hold even if component prices rise? /

  A: For large custom CIS deals, the timeline from signing to revenue recognition is typically 6 to 9 months, as Akamai procures hardware and secures data center capacity after a contract is signed. Pricing is collaborative with large customers; the market environment has rising hardware prices, and Akamai builds pricing adjustment mechanisms into long-term contracts to pass along cost increases, maintaining margin profiles. Most cost components (data center leases, hardware, maintenance) are locked in long-term contracts, with only small annual escalators already factored into contract pricing. (432 characters)

• Q: What opportunity do you see for edge AI inferencing leveraging your unique global edge platform, and how does AI drive growth for your security business? /

  A: AI infrastructure requires a continuum from core to edge: large foundation model training is done centrally, but inference for latency and bandwidth-sensitive use cases (like robotics) is best done close to the end device. Akamai is the only provider that can deliver both core and edge AI capacity on a single unified platform, optimizing both cost and performance. For security, AI agent proliferation creates new demand: LayerX protects against sensitive data leaks from employee AI use, and Akamai's bot management identifies and manages authorized vs unauthorized AI crawlers/scrapers, helping customers maintain brand visibility and site security. (461 characters)

• Q: What is Akamai's financing strategy to fund the large upfront capital needs of the growing CIS business, and what is the CIS revenue ramp outlook for H2 2026? /

  A: Akamai ended Q2 with $4.6 billion in cash on hand, maintains an investment grade credit rating, and has substantial unused debt capacity. Convertible debt has been the most attractive financing option to date, and the company has temporarily paused share buybacks to reallocate capital to CIS growth. Management expects minimal CIS acceleration in Q3 due to delayed GPU shipments, but projects a sharp acceleration (a 'hockey stick') in Q4 2026 as contracted capacity comes online, with continued acceleration into 2027. (398 characters)

• Q: What specific advantages does Akamai offer CIS customers that lead them to choose Akamai over hyperscaler competitors? /

  A: Akamai's key advantages include a long-standing reputation for reliability, a uniquely massive global distributed footprint that places compute capacity close to end-users and data, lower egress costs than centralized hyperscalers, and a full-stack offering that combines compute, storage, content delivery, and enterprise cybersecurity on one platform. This combination delivers better performance, lower latency, and lower total cost than competing offerings for most distributed AI workloads. (281 characters)