Steven Cress's 2026 AI Picks: A Deep Tech Strategist's S-Curve Analysis

Steven Cress operates from a quantitative, growth-at-a-reasonable-price (GARP) framework. His strategy is systematic, relying on data models to identify companies where fundamental strength aligns with valuation. This approach, honed through a career in finance and fintech, aims to cut through market noise and volatility. As he noted in a recent review, the goal is to find stocks that are "really strong in terms of their fundamentals, not overvalued" – a precise calibration for the current cycle.

His 2026 thesis is a direct application of this philosophy to the AI paradigm shift. He is targeting companies positioned on the steep, early-to-mid phase of the AI infrastructure S-curve. The rationale is clear: the global AI market is projected to grow at a CAGR of 30.6% to exceed $3 trillion by 2033. This isn't a fleeting trend; it's a multi-year adoption curve that will drive sustained capital expenditure.

The focus is squarely on the foundational compute stack. The explosive demand from large language models creates a fundamental need for more chips, memory, and storage. This isn't about chasing the latest consumer app; it's about investing in the rails that will carry the entire next technological era. Cress's quantitative system is built to identify the companies that are building these essential infrastructure layers, where the growth trajectory and the valuation are in sync.

The 2026 AI Stock Picks: Rationale and S-Curve Positioning

The picks themselves are a deliberate portfolio across the compute stack. They target companies whose revenue growth is directly tied to the surge in AI server shipments and the record-breaking data center capital expenditure. This isn't about speculative bets on future applications; it's about capturing the fundamental infrastructure build-out that is already underway. The list includes both established leaders and emerging players with strong fundamentals, aiming for diversified exposure to the accelerating adoption phase.

A key rationale for this selection is the shift in the AI hardware paradigm. The early phase was dominated by a pure performance race, but the next phase is about efficiency. As models grow larger, the cost of power and cooling is becoming a critical bottleneck. This makes infrastructure for power delivery, thermal management, and high-speed networking not just supportive, but essential. Companies that solve these efficiency challenges are positioned for exponential growth as the AI compute curve steepens.

The specific picks, drawn from Cress's recent commentary, illustrate this strategy. They include Micron and CommScope as quant strong buys in the AI sector, alongside Seagate Technology. Micron represents the memory layer, critical for the massive data throughput required by AI workloads. CommScope is a leader in the networking and connectivity infrastructure that links servers within the data center. Seagate provides the storage backbone for the petabytes of data generated and consumed. Together, they form a foundational stack where growth is directly linked to the AI capex cycle.

This approach is a classic deep tech play. It focuses on the infrastructure layer where the adoption curve is accelerating, not on the consumer-facing applications that often get the headlines. By targeting companies with strong fundamentals and a clear link to the AI infrastructure build-out, Cress's picks aim to ride the S-curve at its most productive phase.

Updates to Previous Picks and the Evolving S-Curve

The evolution from Cress's 2025 picks to his 2026 list is a clear signal of the infrastructure S-curve maturing. The earlier recommendations, like those highlighted in his August 2025 piece, were positioned for the initial, explosive phase of the build-out. At that time, the narrative was about pure performance and capacity, with hyperscalers committing more than $300 billion in capital expenditures for AI expansion. The focus was on capturing the first wave of demand.

The 2026 list reflects a necessary evolution as adoption accelerates and the stack matures. The paradigm is shifting from a race for raw compute to a race for capital efficiency. As models grow larger and data center investments reach record levels, the total cost of ownership-including power, cooling, and network bandwidth-is becoming a critical bottleneck. This dynamic is what makes the updated picks more strategic. They target companies solving these efficiency challenges, ensuring their growth is sustainable as the AI compute curve steepens.

This adjustment also incorporates a view on market cycles, particularly the potential impact of monetary policy. The 2025 context was one of rising hopes for Fed rate cuts, which fueled a broad market rally. The 2026 outlook must account for the possibility that such cuts could be delayed or that higher rates persist, affecting the valuation of high-growth infrastructure plays. By focusing on companies with strong fundamentals and a clear link to the AI capex cycle, Cress's updated strategy aims to ride the S-curve while building a buffer against the volatility that often accompanies shifts in the monetary environment. The list is a refined bet on the infrastructure layer, where the adoption trajectory is now more visible and the need for efficient scaling is paramount.

Catalysts, Risks, and What to Watch for the S-Curve Thesis

The S-curve thesis for AI infrastructure is now in its steep, accelerating phase. The key question for investors is not if the build-out will happen, but at what pace and with what efficiency. The near-term signals will confirm whether the fundamental adoption curve is holding or if it faces a disruptive inflection.

First, monitor quarterly guidance from major cloud providers and semiconductor vendors. This is the most direct pulse check on the AI capex cycle. The record more than $300 billion in capital expenditures committed this year sets a high bar. Any shift in that plan-whether a pause, acceleration, or reallocation-would be a major catalyst. For instance, if a hyperscaler like Microsoft or Amazon signals a slowdown in AI server orders, it would challenge the growth trajectory for companies like Micron and CommScope. Conversely, upward revisions to capex budgets would validate the thesis and likely drive the stock prices of foundational infrastructure plays higher.

Second, watch for announcements on new compute architectures. The current stack is focused on scaling existing GPU-centric models, but the paradigm is shifting toward efficiency. Breakthroughs in chip design, memory hierarchy, or even alternative compute paradigms could represent a potential inflection point. While disruptive, such innovations would likely benefit the entire ecosystem by lowering the cost of ownership, which is becoming a critical bottleneck. The market will reward companies that are positioned to adapt or lead in this new efficiency race.

Finally, track regulatory developments and geopolitical factors that could impact the supply chain for critical components. The multi-year build-out depends on a stable flow of semiconductors, rare earth materials, and advanced manufacturing. Any new trade restrictions, export controls, or policy shifts could introduce significant friction and cost. This is a key risk to the smooth execution of the S-curve, as supply chain bottlenecks could delay deployments and pressure margins across the infrastructure stack.

The bottom line is that the thesis is robust, but it is not immune to real-world execution risks. The path forward is paved with clear signals: capex guidance, architectural innovation, and supply chain stability. Watching these will separate the durable infrastructure plays from the speculative ones.