Anthropic signs $11.6bn Akamai cloud deal and gets warrant for 5% of Akamai

Anthropic signed a seven-year, $11.6bn agreement to use Akamai cloud infrastructure, with a potential additional $9bn expansion that could bring total contract value to about $20bn. Akamai also issued Anthropic a warrant for 7.7 million shares on an as-converted basis. The deal represents a major long-term cloud-infrastructure commitment tied to AI demand and is materially positive for Akamai's revenue visibility.

Analysis

The investable issue is not the headline contract value but whether this converts Akamai from a low-growth delivery/security multiple into a credible AI-inference infrastructure platform. A large, technically demanding customer can improve utilization of Akamai’s distributed compute footprint, spreading fixed network and server costs over materially higher revenue; incremental margins should expand sharply once capacity is installed. The warrant is economically important: until its strike price, vesting triggers, and performance conditions are disclosed, headline backlog likely overstates the value accruing to common shareholders.

Near term, AKAM should outperform on backlog visibility and a likely upward revision to cloud growth expectations, but the stock’s durability over the next 1-3 months depends on management quantifying revenue-recognition timing, required GPU/server capex, and whether the commitment is take-or-pay. The principal downside is that deployment requires front-loaded capex and power/network spend before utilization ramps, creating FCF pressure despite reported bookings. A more subtle risk is customer concentration: if this becomes a meaningful share of cloud revenue, any Anthropic model-efficiency gains, supplier diversification, or hyperscaler price concession could impair the expected run-rate.

The contrarian view is that the market may be underestimating the strategic value of inference distribution while overestimating near-term P&L conversion. Anthropic’s workload is a validation event for edge-oriented AI serving, which could pressure premium valuations for edge/application platforms such as NET and FSLY if Akamai can bundle compute with its installed security and delivery relationships. Conversely, AKAM should not receive a full AI-cloud re-rating unless it demonstrates repeatable non-Anthropic demand and cloud gross-margin expansion within the next two earnings cycles.

Key Decisions for Investors

• Establish a starter long AKAM only after the warrant and contract economics are filed or discussed on the next earnings call; add if management confirms take-or-pay commitments, phased capacity deployment, and no material FY free-cash-flow dilution. Base case is a 6-12 month multiple re-rating; exit if cloud growth does not accelerate by the second reported quarter after deployment begins.
• Use a 1-3 month relative-value expression: long AKAM / short NET in equal dollar amounts, sized modestly. AKAM has the more direct infrastructure-validation catalyst, while NET carries greater valuation sensitivity if edge AI remains a feature rather than a standalone growth engine; cover the short if NET reports materially faster AI-related RPO or large-enterprise growth than AKAM’s cloud segment.
• Do not underwrite the full contract value into earnings estimates until the company discloses annual revenue ramp, capex, and warrant dilution. Set an alert for guidance: a meaningful increase in capex without corresponding cloud revenue guidance is a signal to reduce or avoid the long, as the thesis would shift from operating leverage to balance-sheet execution risk.
• For options-oriented exposure, prefer a defined-risk AKAM call spread dated beyond the next two earnings reports rather than near-dated calls. The catalyst path requires disclosure and capacity buildout, not merely an initial news-cycle move; the trade is invalidated by evidence that the workload is primarily low-margin pass-through capacity or that warrant dilution offsets the incremental contract economics.