α
AlphaSteve
← Brain index

Alpha-territory expansion — two new screening universes

Added 2026-05-28 following user feedback that AlphaSteve has been searching well-covered US large/mid-cap names where price-discovery is efficient and mispricings are smaller. Classical deep-value (Klarman, Marks, Greenblatt) is explicit that alpha lives where coverage is thin and marginal sellers are uninformed. This file codifies two new screening universes the kit will now monitor alongside the existing six-screen funnel — not as a replacement.

The directive is unchanged: find durable equity mispricings. What changes is where the kit looks.

Universe 1 — Forward-tech picks-and-shovels

Mission anchoring. Mega-cap names exposed to forward-tech themes (NVDA for AI, PLTR for defense AI, RTX/LMT for defense, MSFT/GOOG for quantum) are already heavily covered and priced. The picks-and-shovels layer — tier-2 and tier-3 suppliers, component makers, materials providers, ground-segment integrators — is where coverage thins out, narratives are slower to form, and the deep-value frame has the most surface to work on. These are exactly the companies that capture economic profit during a structural shift in a vertical without being the consensus expression of the theme.

Sub-universe 1A — Robotics suppliers

What's in scope: companies whose revenue mix has material exposure to industrial robotics, collaborative robotics, mobile robotics, warehouse automation, surgical robotics, or autonomous-vehicle robotics. Not the integrators (ABB, Fanuc, Kuka are well-covered Japanese/European primes). The suppliers to integrators.

Starting roster (illustrative, not exhaustive; the scan should look beyond this list):

  • Cognex (CGNX) — machine vision; well-covered but high-quality enough to monitor for dislocations
  • Symbotic (SYM) — warehouse robotics; Walmart-exposed
  • Teradyne (TER) — semi test + robotics (Universal Robots, MiR mobile robotics)
  • Rockwell Automation (ROK) — industrial automation, well-covered but cyclical-dislocation candidate
  • Brooks Automation / Azenta (AZTA) — automation
  • Trimble (TRMB) — precision positioning, geospatial
  • Allient (ALNT) — motion control, small-cap, thinly covered
  • Ouster (OUST) — lidar, small-cap
  • Aeva Technologies (AEVA) — FMCW lidar, micro-cap
  • Innoviz Technologies (INVZ) — lidar, micro-cap
  • Microbot Medical (MBOT) — surgical robotics, micro-cap
  • Asensus Surgical (ASXC) — surgical robotics, micro-cap

Sub-universe 1B — Quantum supply chain

What's in scope: companies whose revenue or technology stack is materially exposed to quantum computing or quantum sensing — superconducting fabs, cryogenic cooling, photonic integrated circuits (PICs), specialized lasers, materials enablement, dilution refrigerators, quantum software stacks, quantum sensors.

Starting roster:

  • IonQ (IONQ) — trapped-ion quantum, public
  • Rigetti Computing (RGTI) — superconducting quantum, public
  • D-Wave Quantum (QBTS) — annealing quantum, public
  • Coherent (COHR) — photonics, including quantum-relevant lasers
  • Bruker (BRKR) — scientific instruments, quantum-relevant
  • Aehr Test Systems (AEHR) — test systems, semi + quantum exposure
  • PsiQuantum (private — track for IPO) — photonic quantum
  • Quantinuum (Honeywell-controlled, partial public via HON) — track via Honeywell exposure
  • IBM (IBM) — large-cap; not the pure-play but the most embedded quantum stack
  • Atomera (ATOM) — quantum-relevant semi tech, micro-cap

Sub-universe 1C — Defense electronics outside the primes

What's in scope: companies whose defense revenue is from subsystem electronics, mission systems, embedded compute, ISR sensors, autonomous-systems software, or proprietary aerospace components — not the Big 5 primes (LMT, RTX, NOC, GD, BA).

Starting roster:

  • Mercury Systems (MRCY) — mission electronics, RF / EW
  • Curtiss-Wright (CW) — defense embedded, ruggedized compute
  • Ducommun (DCO) — defense electronics + aerostructures
  • Heico (HEI) — defense + commercial aerospace components
  • TransDigm (TDG) — proprietary aerospace components, well-covered
  • Elbit Systems (ESLT) — Israeli defense electronics
  • Leonardo DRS (DRS) — defense electronics
  • Kratos Defense (KTOS) — autonomous systems, defense AI
  • AeroVironment (AVAV) — small UAS, defense AI
  • Booz Allen Hamilton (BAH) — services-led but defense-tech-exposed
  • CACI International (CACI) — defense IT services + tech
  • Parsons (PSN) — defense + infrastructure
  • OSI Systems (OSIS) — security electronics

Sub-universe 1D — Space supply chain

What's in scope: companies enabling commercial space — launch vehicles, satellite buses, ground-segment integration, propulsion, in-space servicing, sat-comm infrastructure, Earth observation enablers, space ISR, debris management.

Starting roster:

  • Rocket Lab USA (RKLB) — launch vehicle + space-systems
  • Redwire (RDW) — space-systems components, in-space manufacturing
  • AST SpaceMobile (ASTS) — direct-to-device satellite comms
  • Iridium Communications (IRDM) — sat-comm constellation
  • Planet Labs (PL) — Earth observation
  • BlackSky (BKSY) — Earth observation, ISR
  • Spire Global (SPIR) — RF data, weather, AIS
  • Sidus Space (SIDU) — micro-cap space services
  • Karman Holdings (KRMN) — recent IPO, propulsion/space hardware
  • Maxar Technologies (private now) — track via private market
  • Voyager Space (private) — track for IPO
  • Intuitive Machines (LUNR) — lunar surface services
  • Astera Labs (ALAB) — connectivity, partial space exposure

Universe 2 — Small/micro-cap with thin sell-side coverage

Mission anchoring. Klarman's Margin of Safety and Greenblatt's You Can Be a Stock Market Genius both make the same point: the deepest mispricings live where price-discovery is weakest, which means small-cap and micro-cap names with minimal sell-side coverage and dominated by retail or technical (forced) sellers. The kit's existing screens cover the Russell 1000 plus extensions; the Russell 2000 tail and the sub-Russell-2000 universe is structurally where alpha is more available, provided the screens are designed for it.

Screening criteria

  • Market cap: $50M floor (avoid distressed micro-cap with insufficient operational data) to $2B ceiling (the conventional small-cap cutoff)
  • Sell-side coverage: ≤5 analyst estimates available, with explicit preference for ≤3 (the practical "thin coverage" threshold)
  • Liquidity: minimum $1M average daily volume (so the kit can actually trade without market impact; the constraint is structural — names below this floor are unactionable regardless of mispricing magnitude)
  • Listing: US-listed (NYSE / NASDAQ / NYSE American); OTC and pink-sheets excluded (data quality and liquidity)
  • Geographic exposure: US revenue ≥50% (familiar accounting standards, US regulatory environment)
  • Operational profile: profitable in at least one of the last three fiscal years OR positive free cash flow last twelve months (avoid pure-cash-burn stories — those are venture-stage, not deep-value)
  • Quality floor: ROIC ≥10% in at least one of the last three years; debt/equity ≤1.5 (excludes structurally-broken balance sheets)

Screening process

Each scan run, the AM and PM deep-value scan tasks pull a fresh list from:

  • Russell 2000 constituents filtered by the criteria above
  • Cross-reference with Form 4 insider-buying clusters
  • Cross-reference with 52-week-lows screens
  • Cross-reference with EV/sales, EV/EBITDA, P/B compression screens

Names surfacing on 2+ screens are first-read candidates. Names surfacing on 3+ screens are Tier 1 candidates for promotion.

Specific anti-patterns to filter

  • Reverse-merger SPACs with no operating history
  • Single-product biotech without commercial revenue
  • Cannabis equities (regulatory uncertainty makes the structural-value calculation untrustworthy)
  • Chinese reverse mergers (accounting-standard risk)
  • Names with auditor-resignation or restatement history in the last 36 months

Integration with existing six-screen funnel

The existing funnel (deep-value-scan-skill) runs six screens daily: 52-week lows, EV multiple compression, insider clusters, value-line screens, capital-cycle inflections, and Tier-1 candidates. The two new universes add two parallel screens to that funnel:

  • Screen 7 — Forward-tech picks-and-shovels: Each scan run, the four sub-universes (robotics, quantum, defense electronics, space) get a fresh pass against the existing kit's evaluation criteria (EPV, margin of safety, kill criteria). Names with material setups get promoted to first-read.
  • Screen 8 — Small/micro-cap thin-coverage: Each scan run, the small/micro-cap universe defined above gets a fresh pass against the same kit criteria.

The funnel logic is unchanged. The thesis pipeline accepts inputs from both new screens.

What does NOT change

  • The doctrine. Greenwald-modified EPV, Klarman MoS discipline, Marks second-level thinking, capital-cycle lens. All carry. This is the same investment philosophy applied to a different search universe.
  • The verdict gates. Five-gate decision framework unchanged. Buy / pass-with-trigger / pass / avoid still operate. A small-cap thinly-covered name has to clear the same gates as a large-cap; it doesn't get a quality discount because the coverage is thin.
  • The portfolio discipline. Position sizing, kill criteria, calibration tracking — all carry. Small-caps may carry tighter position-size caps for liquidity reasons (see Portfolio section on small-cap sizing), but the discipline framework is the same.

Recalibration condition

The kit's six-month evaluation framework (six-month-test) measures whether AlphaSteve is operating to standard across six dimensions. The November 2026 review will explicitly assess whether the two new screening universes are surfacing actionable opportunities at a rate above what the existing six-screen funnel surfaced in months 1-6.

  • If yes: the new universes are doing their job; consider extending coverage to additional under-researched territories (Japan small-caps, EM ex-China, special situations).
  • If no: retire the new screens. The cost of running screens that don't surface alpha is opportunity cost on the kit's attention.

This is the same falsifiability discipline applied elsewhere — the change is empirically testable and gets re-evaluated against a pre-specified bar.

Linked