Bloomberg Tech: Anthropic’s AI Spending Surges (Podcast)

Anthropic signed an $11.6 billion AI-computing contract with Akamai, the largest agreement in Akamai's history, underscoring continued large-scale infrastructure spending on generative AI. Microsoft is consolidating its consumer and workplace Copilot offerings, stepping back from maintaining a separate personal chatbot effort. Trump characterized his meeting with Xi Jinping as productive, but the report indicated few tangible outcomes, leaving U.S.-China trade and geopolitical implications uncertain.

Analysis

AKAM’s rerating depends less on the headline contract value than on contract duration, minimum-take-or-pay terms, and incremental GPU/network capex. If the commitment is largely capacity reservation with pass-through hardware costs, revenue can scale faster than EBITDA and free cash flow; if it includes firm utilization guarantees, it could transform investor perception from legacy CDN to scarce AI-infrastructure provider and justify multiple expansion versus Cloudflare (NET). The near-term risk is customer concentration: any model-efficiency breakthrough, financing constraint, or shift toward hyperscaler-owned capacity could leave AKAM with underutilized specialized infrastructure.

The more investable second-order effect is pressure on AI compute intermediaries with weaker networks or higher funding needs, notably private CoreWeave and smaller GPU-cloud peers. AKAM’s global edge footprint can reduce inference latency and data-egress expense, making distributed inference a plausible higher-margin follow-on opportunity over 6-18 months; this is not yet proven by a single customer commitment. Watch management’s next earnings call for capex-to-revenue conversion, remaining performance obligations, gross-margin guide, and whether Anthropic-related revenue is recognized ratably or usage-based.

MSFT’s product consolidation is strategically rational but exposes a weaker consumer willingness-to-pay signal than the market has embedded in broad Copilot optionality. Over the next 1-3 months, focus should shift to paid-seat conversion and ARPU in Microsoft 365 rather than product-launch cadence; enterprise bundling can protect distribution but may cannibalize standalone monetization. The uncertain US-China backdrop is more a volatility input for semiconductors and hardware supply chains than a direct AKAM or MSFT earnings catalyst absent concrete export-control or tariff changes.

Key Decisions for Investors

• Initiate a small long AKAM / short NET pair over the next 1-2 weeks only if AKAM provides minimum-commitment and duration disclosure; target 15-20% relative upside over 6-12 months, with exit if incremental capex rises faster than contracted revenue or gross-margin guidance falls by more than 200 bps.
• Do not chase AKAM on the initial move. Add on a post-event pullback only after confirming that contract economics are not predominantly pass-through capacity; lack of disclosure is an alert condition, not a long signal.
• Maintain a tactical underweight in MSFT versus GOOGL for 1-3 months if Copilot consumer consolidation is accompanied by flat paid-user or ARPU commentary. Cover the relative short if Microsoft reports accelerating M365 Copilot attach rates or raises AI revenue guidance.
• Monitor AI-infrastructure credit and utilization indicators for 6-18 months: a widening in GPU-cloud financing spreads, reduced hyperscaler capex guidance, or evidence of lower inference demand would invalidate the AKAM scarcity thesis and favor reducing all AI-infrastructure exposure.