Rebuilt strict screen • not current AI supply-chain beneficiaries
50 U.S. + 50 non‑U.S. public companies with latent AI-useful capabilities
This revision implements the stricter rule: exclude companies already supplying or clearly positioned in AI build-out categories such as optics/CPO, advanced packaging, wafer fabs, semiconductor process tools/materials, data-center power, cooling, transformers, racks, and obvious data-center construction. The remaining candidates are second-order capability owners where the AI angle appears less discovered.
100
Public-company candidates
50
U.S. stocks ranked
50
Non-U.S. stocks ranked
25
Direct AI-chain examples excluded
Strict exclusion rule
I did not treat “could be used in AI” as “already sells into AI.” Candidates were screened to avoid obvious current AI infrastructure suppliers. The dashboard uses “no clear current evidence found” rather than claiming it is possible to prove zero AI exposure.
Excluded: optical interconnect / CPO
Excluded: wafer-fab / semicap
Excluded: advanced packaging / HBM tools
Excluded: data-center power / transformers
Excluded: AI data-center cooling
Excluded: obvious data-center construction/services
Included: latent fluid/water systems
Included: industrial materials/components
Included: sensors, motion, MRO, safety, recycling
Top U.S. candidates
| Rank | Ticker | Company | Alpha | Theme |
|---|---|---|---|---|
| 1 | LNN | Lindsay | 80 | Water/irrigation telemetry |
| 2 | AZZ | AZZ | 79 | Metal coatings/infrastructure |
| 3 | MWA | Mueller Water Products | 77 | Water infrastructure |
| 4 | TWIN | Twin Disc | 77 | Power transmission |
| 5 | BRC | Brady | 77 | Asset ID/safety |
Top non‑U.S. candidates
| Rank | Ticker | Company | Alpha | Theme |
|---|---|---|---|---|
| 1 | PWH.AX | PWR Holdings | 79 | Thermal systems |
| 2 | BFSA.DE | Befesa | 75 | Metals recycling |
| 3 | NTG.DE | Nabaltec | 74 | Ceramic fillers/materials |
| 4 | ROR.L | Rotork | 73 | Flow control/actuators |
| 5 | GHH.L | Gooch & Housego | 72 | Specialty photonics |
Investigation visuals
Sortable ranked table
| Region | Rank | Ticker | Company | Country | Alpha | Theme | AI-chain risk | Confidence | Latent | Discovery | Valuation | Catalyst | Quality | Penalty | Latent pathway | Current AI-chain screen | US mcap $bn | US P/E | Non-US tier | Source |
|---|
Examples deliberately excluded
These names are useful comparables and may still be good businesses, but they fail the user's stricter “not used in AI supply chain yet” test.
| Ticker | Company | Direct category | Reason excluded | Source |
|---|
Methodology note
# Strict Latent AI Alpha Screen — Methodology
Generated: 2026-05-06 America/Los_Angeles context. This is a research-prioritization screen, not investment advice.
## User clarification implemented
The earlier screen was too broad. This revision excludes companies that are already clearly used in the AI supply-chain build-out: AI chips, wafer-fab tools/materials, semiconductor processing, advanced packaging, CPO/optical interconnect, data-center power, cooling, transformers, racks, fire/safety services explicitly sold into data centers, and obvious data-center construction beneficiaries.
## What “not used yet” means in this dashboard
Absence is difficult to prove. I therefore use a stricter but honest label: **no clear current AI/data-center/fab/advanced-packaging/CPO/optics/power participation was found in the spot checks used for this screen.** Each candidate still requires follow-up review of its latest annual report, investor presentation, backlog, customer concentration, and transcript language before capital is deployed.
## Alpha Score
Alpha Score = Latent Fit + Discovery Gap + Valuation Setup + Catalyst Density + Execution/Quality - Hype Penalty.
- Latent Fit: how naturally the company's capability could map to AI infrastructure if demand spills beyond obvious suppliers.
- Discovery Gap: how under-followed or non-consensus the AI pathway appears.
- Valuation Setup: for US names, a rough score based partly on the finance-tool market-cap and P/E snapshot on 2026-05-06 UTC; for non-US names, a qualitative setup because reliable non-US finance snapshots did not return in-session.
- Catalyst Density: plausible events that could make the latent pathway visible.
- Execution/Quality: business quality and ability to scale.
- Hype Penalty: penalty for category adjacency that may already be crowded or too close to current AI supply.
## Output files
- assets/data/companies_strict_latent.csv — ranked 50 US + 50 non-US candidates
- assets/data/excluded_direct_ai_supply_chain.csv — names removed for direct current AI-chain exposure
- assets/data/source_pack.csv — research source links
- assets/images/*.png — visualization pack
## Important limits
This dashboard is not a formal valuation model. It does not claim any company has zero AI-related revenue. It identifies stocks where the public AI thesis appears not yet obvious from the sources reviewed and where the company owns capabilities that could become relevant if supply constraints spread into second-order parts, materials, water, filtration, motion, security, safety, recycling, and MRO layers.
...