AI is built from scarce inputs. The Fund owns the companies that sell those inputs and not the companies racing to consume the supply. Investment focuses on the hardest-to-replicate physical bottlenecks of the AI infrastructure buildout with discipline toward any company whose stock price already assumes a smooth and uninterrupted expansion.
Hyperscalers, sovereigns and defense establishments are committing trillions of dollars to AI infrastructure. These are strategic necessities and not discretionary programs.
Hyperscalers continue to commit hundreds of billions in annual capital expenditure. These are multi-year infrastructure programs already underway that are difficult to reverse and that build on each other. The industrialization of intelligence is under way.
The buildout began in the data center but the value created does not end there. Consumer AI adoption and autonomous vehicles in real-world deployment and humanoid robotics entering homes and workplaces are the next legs of the same structural trend. Each widens the set of companies that benefit as intelligence becomes an industrial input.
Compute, power and materials are the scarce inputs every layer of the AI buildout depends on. Capital cannot eliminate that scarcity overnight. The bottleneck moves from power to packaging to photonics to materials but it does not disappear. The Fund owns the suppliers of those inputs and not the companies racing to consume the supply.
The best holdings are not simply AI beneficiaries. These are companies tied to physical constraints that cannot be resolved with software or capital alone.
Supplier pricing power and extended lead times and constrained capacity that capital cannot quickly replicate. Evidence matters more than narrative.
Materials and packaging and power delivery move before the compute megacaps do. The portfolio holds the entire value chain and not just the headline names.
Revenue growth and margin trends and rate-of-change data and proprietary research and industry developments and management strategy determine whether a position's scarcity thesis still holds. Theme weights and position sizing are adjusted accordingly.
Companies whose valuations already require an uninterrupted and perfectly linear AI expansion. The market prices the best case while the Fund prices the constraint. These are two very different portfolios.
Companies that benefit from AI because everything benefits from AI. The Fund requires a direct structural tie to a physical constraint and not simply demand exposure.
Software responds well to capital because more code can be written. Physical capacity such as transformers and fiber and transceivers and specialty chemicals does not scale overnight. That asymmetry is the opportunity.
A compelling thesis is necessary but not sufficient. The Fund requires evidence that the constraint exists and is reflected in pricing power or lead times and that the entry price does not already assume resolution.
Every holding maps to one of six physical bottleneck themes that span the AI infrastructure value chain from raw materials to compute delivery. Positions are equal-weighted at inception and foreign companies are held through US-listed ADRs where available. A small number of opportunistic positions sit outside the six core pillars.
Fiber cable and optical transceivers and photonic integrated circuits and connectors. The data center spine that moves model traffic at the speed of light.
Wafer fabrication and EDA tools and process equipment and advanced packaging and specialist logic. The geometry of AI compute increasingly lives in the packaging layer and not just the die.
Specialty process chemicals and photoresists and CMP slurries and substrates and advanced polymers. Foundational to every wafer ever made and the least-discussed constraint in the stack.
Power delivery and grid transmission and thermal management and data center construction and utility-scale infrastructure. The AI buildout is first and foremost a power story.
Servers and hyperscale networking and storage and switching and systems integrators. The hardware layer that assembles compute and memory and interconnect into deployable AI infrastructure.
Established operators applying AI to proprietary data and intellectual property and the operating platform to expand margins. These are not software vendors but structural beneficiaries of applying AI internally to an existing business.
Outside the six pillars the Fund holds a small number of opportunistic positions in companies undergoing a sudden structural shift into the AI supply chain or with significant AI-driven business model transformation. These positions are selective and are not a thematic pillar.
The portfolio is constructed with principles designed to maximize participation in the structural AI buildout while maintaining discipline on valuation and position concentration.
Pure equity conviction in the structural AI infrastructure buildout. No short positions and no margin and no derivatives. The buildout has multi-year duration and cyclical hedges would reduce the return profile without a commensurate reduction in fundamental risk.
Six value-chain pillars define the portfolio. Selection is based on physical scarcity and bottleneck position and not on quantitative screens or factor exposures or market capitalization weighting. Every holding must earn its place through bottleneck logic.
Governed by a proprietary monitoring process that continuously tracks each holding for fundamental and technical deterioration. Strength is held or added to and a name can be up sharply and still have room to run. Position sizes are adjusted when deterioration appears and capital re-enters after a confirmed correction. The portfolio is not buy-and-hold and not calendar-driven. It reflects a continuously updated view of where the constraint lives in the stack.
Presvic LLC is the manager of the FourthTurning Global Capital Fund, a concentrated long-only equity portfolio focused on the physical infrastructure of artificial intelligence.
The Fund's name reflects the Strauss and Howe generational theory of cyclical history. The Fourth Turning thesis identifies periods of institutional upheaval and accelerated capital reallocation that recur roughly every eighty years. The AI infrastructure buildout is the defining capital expenditure cycle of the present turning: a multi-decade transformation financed largely by the operating cash flow of the largest buyers with a more leveraged tier of borrowers at the margin.
The Fund invests globally across the full AI infrastructure value chain from specialty process chemicals to fiber transceivers to power transformers and targets the companies that supply the physical inputs the buildout cannot proceed without. Foreign companies are held through US-listed ADRs where available.
The Fund is managed by Jay Plourde, Portfolio Manager and Managing Member of Presvic LLC, based in New York.
The FourthTurning Global Capital Fund is available to qualified institutional investors. For strategy materials or investor documentation or to schedule a conversation please reach out directly.
jplourde@presvic.com