Disclaimer: This analysis is for informational and educational purposes only and does not constitute investment advice. All investments carry risk, including the risk of loss. Past performance does not guarantee future results. Please consult with a qualified financial advisor before making investment decisions.
Position Disclosure: The author holds positions in the following symbols: AAOI, AAPL, ALAB, AMD, AMPX, AMZN, APLD, APP, ASTS, AUR, BBAI, BFLY, BKSY, BRK.B, COHR, CRDO, ESTC, ETH, GOOG, GOOGL, HURA, HUT, IBIT, INOD, IONQ, JOBY, KTOS, LITE, MBLY, META, MSFT, MU, NBIS, OKLO, OUST, PDYN, PL, PLTR, POET, QBTS, QUBT, RCAT, RGTI, RKLB, RLAY, RR, SDGR, SGOV, USMV, VGT, VICR, VRT.
Every week, we score ~120 public companies on one question: as AI changes how businesses think, build, and compete, who benefits most? These five ranked highest this week — a sample of the full analysis available to members.
How We Picked Them: Selected from our full coverage based on growth potential, AI advantage, and how well-positioned each company is as of July 28, 2026. We diversified across sectors so you see a range of opportunities, not just one hot corner of the market.
Also available as a PDF download.
Applied Digital Corporation (APLD)
ai
energy
cloud
crypto
Synopsis
The operating setup is stronger than the old speculative-hosting label suggests: signed
hyperscaler demand is real and powered campuses are scarce. The investment question is whether management can translate that scarcity into per-share value before capital costs and counterparties take too much of the upside.
Thesis
Applied Digital can grow from a speculative builder into a scarce-power
AI campus landlord if it keeps converting signed
hyperscaler leases into live
megawatts on schedule and funds that buildout with more project capital than common equity; the upside is non-linear because
backlog is already large, but shareholder capture depends on capital costs and customer concentration.
Last Economy Alignment (0.7/1.0)
Cheaper cognition should expand demand for power-dense AI campuses, and Applied Digital controls a scarce physical bottleneck with high switching costs once a tenant commits. It is not meaningfully exposed to software price collapse, but utilities, lenders, and self-building
hyperscalers can still capture part of the economics.
Critique
This is not a software-to-zero problem; it is a capital-spread problem: revenue can scale while common equity underperforms if debt, preferred capital, and
hyperscaler bargaining absorb the lease economics.
Cerebras Systems Inc. (CBRS)
ai
semiconductors
hardware
cloud
Synopsis
There is a real chance to become a valuable AI infrastructure provider for latency-sensitive workloads. The constraint is not interest in the product; it is whether capacity, utilization, and financing scale fast enough to justify an already ambitious valuation.
Thesis
Cerebras can plausibly grow from a differentiated AI hardware vendor into a high-utilization low-latency compute platform, but from today’s valuation the equity case depends on converting contracted demand into durable cloud economics, not on endless
rerating.
Last Economy Alignment (0.7/1.0)
Cerebras owns real AI compute and benefits directly from rising
inference demand, but capture is capped by weaker ecosystem defaultness, heavy capacity needs, and supplier and customer concentration.
Critique
If buyers end up routing on blended token cost rather than latency and reliability, Cerebras could fund the hardest capacity while larger ecosystems keep the customer relationship and compress its pricing power.
Snowflake Inc. (SNOW)
software
cloud
enterprise
ai
Synopsis
The opportunity is real if Snowflake becomes the trusted operating layer for enterprise AI rather than just the place data sits. That can support a 2x+ equity outcome by 2031, but the company must prove it can keep the economics as
hyperscalers and open interfaces crowd the stack.
Thesis
Snowflake can still create a 2x+ equity outcome by 2031 if it upgrades from
data warehouse to governed AI execution and trust layer, because agentic workloads should pull more enterprise data, policy checks, and audited actions onto the platform; the key question is whether it captures that value before
hyperscalers and open interfaces compress the economics.
Last Economy Alignment (0.7/1.0)
Cheaper cognition increases demand for governed data and AI execution, and Snowflake’s usage model, workflow integration, and trust surfaces let it capture more of that flow. It is not a full chokepoint because
hyperscalers own the compute layer and
agent interfaces could dilute surplus capture.
Critique
The bear case is that AI increases activity on top of enterprise data, but Snowflake becomes a background utility as
agents, open standards, and native cloud stacks own the interface and pricing power.
Vicor Corporation (VICR)
hardware
energy
semiconductors
ai
defense
Synopsis
The company controls a painful part of AI infrastructure: efficient power delivery into denser compute. The opportunity is real, but the stock now needs proof that
backlog, licensing, and added capacity can turn scarcity into durable, diversified earnings.
Thesis
Vicor sits inside a real AI hardware bottleneck: moving more power into hotter, denser compute systems. The upside is substantial if it converts
backlog through a nearly full first
fab, scales a second
fab on time, and turns patent enforcement into repeatable licensing rather than episodic royalty spikes.
Last Economy Alignment (0.7/1.0)
Cheaper cognition increases compute intensity, and compute intensity increases the value of dense, efficient power delivery. Vicor benefits because its moat is physical hardware, process know-how, and licensable IP, not software seats that AI can commoditize.
Critique
If
hyperscalers use Vicor mainly to qualify licensed or bundled alternatives, Vicor may become the technical reference point but not the lasting profit pool, leading to pricing pressure and multiple compression.
Microsoft Corporation (MSFT)
cloud
software
enterprise
ai
cybersecurity
Synopsis
The core question is not whether Microsoft has AI demand, but whether it can capture that demand at attractive economics across cloud, software, and trust layers. If it proves that transition, double-digit compounding can persist from an unusually large base.
Thesis
Microsoft is one of the few companies that can monetize AI at infrastructure, workflow, and trust layers at once; if it shifts value capture from seat add-ons toward usage, governed execution, and security while keeping
Azure utilization high, a roughly 2x equity outcome by July 2031 is realistic even from a mega-cap base.
Last Economy Alignment (0.8/1.0)
AI makes Microsoft's cloud, workflow, identity, audit, and
RBAC control points more valuable, while its
installed base lowers adoption friction. The main leak is
seat deflation and model commoditization, but Microsoft has credible paths to recapture value through usage and governed workflow execution.
Critique
If AI budgets migrate faster from premium seats into cheaper, vendor-neutral usage and
agents, Microsoft could fund the compute layer while losing too much app-layer pricing power and compressing returns on heavy
capex.