AI Infrastructure Demand Broadens Beyond Chips as Data Center, Observability Names Guide Higher

The AI trade is widening. Equinix used the Technology Leadership Forum 2026 platform to frame its outlook around available power and capacity rather than demand — a familiar refrain across the colocation complex, and one that supports pricing power into 2026 as hyperscaler pre-leasing absorbs incremental supply. Dynatrace management argued that AI deployment is itself the growth driver for observability, as distributed inference workloads multiply telemetry volumes and raise the cost of unmanaged latency. Fastly, long a share-donor in the CDN price war, pointed to genuine edge cloud traction as inference moves closer to the end user. Taken together, the commentary supports a second-derivative thesis: infrastructure and tooling vendors capture AI spend with lower model-obsolescence risk than the application layer. Software sentiment is following. Monday.com gained 20.4% in July and post-earnings analysis has shifted to the next catalyst set rather than defending the quarter. In semiconductors, ASML's roughly 5% monthly decline has drawn a high-profile wealth-manager call for 50% returns, reflecting a willingness to underwrite lithography's structural moat through order-book volatility. The counterweight remains at the top of the stack: analysts continue to flag two structural AI problems at Alphabet, with some arguing private vehicles offer cleaner exposure. The message for allocators is that AI leadership is rotating, not fading.