Our approach & assumptions

We study the AI supply chain as a long-term investment opportunity, starting with the strategies capable of turning a much larger market into lasting value per share. This investigation focuses on power deployment and the businesses that enable it: generation, fuels, utilities, grid and electrical equipment, cooling, construction, storage and related infrastructure. US deployment is the principal demand lens; relevant global suppliers and competitors are included.

The central question is practical: if AI capabilities and infrastructure demand advance rapidly, which companies can execute a compelling strategy, finance their growth, and retain enough of the economics to reward their shareholders? The detailed company cases answer that question within the explicit premise below.

The world this research assumes

Major improvements in AI capabilities during 2027 and 2028 support a prolonged expansion in demand for compute and power. Recursive self-improvement and artificial superintelligence motivate the rapid-progress backdrop. They are scenario premises here, rather than findings established by this company research. The stock targets do not depend on assigning an exact date to either milestone.

The opportunity is not assumed to end in 2031. More capable AI can sustain further investment beyond that point. Successful suppliers can remain growth businesses, reinvesting substantial cash into expansion. The models therefore value continuing businesses while charging them for the capital needed to grow.

AI progress affects supply as well as demand. Better engineering, manufacturing, construction and operating methods can accelerate deployment. Rival capacity, product redesign and customer bargaining can normalize margins or shift value between companies. Today's bottlenecks inform the pace and economics of the buildout; they are not treated as permanent limits or permanent sources of excess profit.

Start with strategy, then test the ability to execute

First, identify the problem a company solves and why its strategy matters in a much larger market. A compelling strategy might increase useful power sooner, expand content per customer project, repeat a qualified manufacturing process, or earn service revenue from an installed fleet.

Next, test the evidence: technical capability, management and operating experience, qualifications, customer acceptance, capacity expansion, competitive advantage and the capital required. A large order or ambitious management objective is one piece of evidence. Delivered output, retained margins and cash collection show whether the strategy works commercially.

Finally, translate execution into operating cash, investment, financing and value per share. The ranking combines strategy, execution evidence and modeled upside. An ambitious plan or a high numerical target alone is insufficient.

The assumptions that connect growth to shareholder value

Useful capacity must reach the customer. Desired compute demand, contracted projects, equipment deliveries and energized load are different stages. Generation nameplate, dependable output and annual energy are different measures. A supplier's equipment sale and its customer's owned fleet can represent the same physical capacity; they cannot be added twice.

Expansion consumes capital. Factories, startup costs, working capital, maintenance, replacement, warranties and service obligations are included where relevant. Weak near-term free cash flow can reflect productive investment. Continued growth creates value only when the incremental cash earned justifies the full capital required.

Customers can help fund the buildout. Urgency can support deposits, paid reservations, dedicated contracts and project participation. The models recognize the corresponding invoice credits, concessions and ownership claims. Terms disclosed for one supplier are evidence of a possible mechanism, not proof that another receives the same deal.

Shareholders retain the cash left after other claims. Debt service, partners, leases, recurring awards, convertibles and new equity affect value per share. Future renewals and financing access are explicit conditions. A mechanically calculable valuation with an unresolved funding shortfall does not qualify as a fully financed target.

What a central target means

The central case describes a specified path of successful company execution under sustained rapid AI demand. It is the principal analytical comparison, not a probability-weighted forecast or a claim that the outcome is most likely. Stronger execution, company disappointment and individual sensitivities show what changes the value.

The detailed operating paths run through 2031. Most targets are equity values on December 31, 2031; Powell uses September 30 and Nextpower March 31. Returns begin at September 11, 2026 reference prices, use each stock's local currency, and include modeled buyer-eligible dividends before taxes, fees and foreign-exchange effects. A future target differs from a present fair value.

Growth beyond the target date is a major valuation assumption. BASX and Hammond manufacturing receive fifteen modeled periods of profitable expansion, followed by slower continuing growth. Their targets are sensitive to growth duration, capital productivity and the return investors require. Those sensitivities are part of the investment conclusion.

Evidence, assumptions and investment judgment

Reported results and executed terms establish dated facts. Management guidance describes company objectives. Later output, margins, financing and economic participation are selected assumptions. The models calculate the implications; the investment judgment weighs strategy, execution evidence and the resulting value together.

The downside tests include company disappointment while AI demand remains rapid. They challenge market share, margins and financing after capital has been committed. The central rankings should therefore be read as conditional on the rapid-AI premise, with separate risks attached to each company's ability to capture the opportunity.

Start with the investment conclusions, then use the full report to inspect the company evidence, valuation conditions and risks. The research and reference prices are dated; this edition does not imply a live market update.