Matt Tuttle

Buy GEV as the genuine near-term AI infrastructure bottleneck — gas turbines and grid interconnect are capacity-constrained now, with GE Vernova's order book sold out into the 2030s, making it the direct beneficiary of hyperscaler campus buildout over 2026–2028.

Buy TSM as a debasement-resistant asset — foundry capacity cannot be rapidly replicated, demand is structurally inelastic, and pricing power provides real-terms protection against dollar erosion driven by record T-bill issuance and sustained 5%+ long rates.

Buy GLW as a structural beneficiary of the glass substrate roadmap for advanced AI packaging; speaker argues substrate suppliers are systematically mispriced relative to the GPU names that dominate investor attention.

Buy GLD as central bank gold demand (863t/year) reflects structural de-dollarization and counterparty risk hedging post-2022 sanctions, not a CPI trade; measured against the global monetary base rather than CPI, gold remains early in its repricing.

Buy energy producers via XLE; allocators remain structurally underweight following spring positioning, and a prolonged Strait closure combined with oil printing new highs creates a re-rating catalyst for the sector.

Short Micron because Google's TurboQuant 6x LLM memory compression could structurally reduce AI-driven memory demand — a risk not priced into MU's +141% YTD run, where bulls assume AI workloads sustain elevated memory consumption.

Europe is building a "Kill Switch" to insulate itself from US tech leverage. France and others have stated desires to move off platforms like Teams and Zoom. This is the "Soft Power" rebuild. Regulatory pressure and "sovereignty" mandates will force EU enterprises and governments to switch to local providers for cloud, telecom, and integration. This creates a "Euro Stack" of winners in semi-conductors (ASML, Infineon), telecom infrastructure (Deutsche Telekom, Ericsson, Orange), and data/cloud services (SAP, Capgemini, Nebius). Long the "European Digital Sovereignty" basket. These companies have a regulatory moat protecting them from US competition within the EU. European tech execution has historically lagged; US tech giants are deeply entrenched and difficult to displace.

Buy Coherent as a named recipient of Nvidia's photonics capital deployment; the AI interconnect bottleneck thesis and Jensen Huang's explicit capacity shortage warning point to a multi-year demand ramp for optical component suppliers.

Buy Lumentum as a direct beneficiary of Nvidia's $6.5B photonics investment wave; AI's bottleneck has shifted from compute to interconnect, and silicon photonics capacity is structurally undersupplied at the start of a multi-year supply chain build cycle.

Buy Marvell as a named photonics/interconnect beneficiary of Nvidia's capital; the thesis that AI infrastructure spending is rotating from compute silicon to interconnect favors Marvell's custom silicon and optical DSP exposure.

Buy Antero Resources as a specific energy name the speaker explicitly tags; the SPR depletion thesis and physical infrastructure underinvestment narrative support owning natural gas producers ahead of a potential supply shock repricing.

Buy Cheniere Energy as a directly named infrastructure play; the speaker explicitly cites LNG terminals as an ignored physical constraint, and the SPR depletion thesis amplifies the asymmetric repricing case for LNG export infrastructure.

Buy energy services via OIH; with the SPR depleted and physical infrastructure underinvested, energy services companies are positioned to benefit from the structural rebuild cycle the speaker argues is being ignored by a market focused on semis.

Buy duration (TLT) as Warsh's hawkish-on-QT, dovish-on-rates stance is misread by a front-end market still pricing him as a generic dove; once sequencing is understood, the rate-cut signal dominates and long-end bonds rally.

Buy VRT as SK Hynix's iHBM launch — embedding cooling inside the memory stack — signals thermal management is the next structural AI infrastructure bottleneck, directly intensifying demand for rack-level liquid cooling solutions.