Skip to content

Meet Mosaic

Find ideas that fit your investment framework

Mosaic applies your research framework company by company—investigating the questions that matter, preserving the evidence behind each conclusion, and helping you compare the findings.

Mosaic researches companies against your framework, with evidence behind every conclusion.

Which US industrial companies are direct beneficiaries of structural capital spending?

Investment framework

US Industrial-Capex
Compounders

Company universe · Four qualitative gates · Ranking rubric

US-listed Industrials $10–100B market cap +4

Best matches for this framework

  • ELO1,355.5
    STRL.USSterling Infrastructure, Inc.
  • ELO1,347.9
    MTZ.USMasTec, Inc.

Sterling Infrastructure

STRL.US

Screening conclusion: QUALIFIES

Sterling passes all four framework gates. Direct capex exposure, expanding backlog and mission-critical execution support its fit.

1. Capex-megacycle exposure — PASS

E-Infrastructure serves data centers, semiconductor fabrication, manufacturing… (STRL 10-K FY2025).

This directly supports exposure to AI data-center buildout and semiconductor-fab construction, with additional exposure to US manufacturing and power infrastructure.

Your investment framework, applied company by company

  1. Define what a company must prove

    Turn your investment thesis into a reusable research method. Define what qualifies a company, what rules it out and the evidence Mosaic must find.

    Set what qualifies a company, what rules it out and the evidence Mosaic must find.

    RunsDefinition

    US specialty industrial / A&D / specialty-materials companies at $10–100B market cap that are direct beneficiaries of AI datacenter, semiconductor fab, US reshoring, defense industrial base, and IIJA/grid capex mega-cycles. Modeled on EMCOR (EME), Comfort Systems (FIX), Sterling Infrastructure (STRL), and Carpenter Technology (CRS).

    Research section

    US Industrial-Capex Compounders — framework v1

    US specialty industrial / A&D / specialty-materials companies at $10–100B market cap that are direct beneficiaries of AI datacenter, semiconductor fab, US reshoring, defense industrial base, and IIJA/grid capex mega-cycles.

    Qualitative context

    Tasks 1–4 each demand a PASS/FAIL verdict with cited evidence from the company’s own filings (10-K/10-Q/earnings-call/investor presentation); a company must PASS all four to qualify. Task 5 is the peer-valuation construction, independent of pass/fail.

    Ground every judgment in the Portrait document corpus; do not infer end-market exposure from industry classification alone.

    Tasks

    [capex-megacycle-exposure] Assess the subject’s direct US capex mega-cycle exposure. Report the specific end-market disclosure, revenue exposure and a PASS/FAIL verdict.

  2. Research your universe

    Apply the same research approach company by company. Investigate the questions that financial filters alone cannot answer.

    Start with a small sample, examine the findings and refine the instructions before expanding the work.

    Test your research instructions on a small company sample. Review and refine them before expanding to your universe.

    Industrial capex sample

    US-listed · $10–100B market cap

    • Advanced Energy Industries, Inc.AEIS.USDoes not qualify
    • Alcoa CorporationAA.USDoes not qualify
    • Sterling Infrastructure, Inc.STRL.USQualifies
    • MasTec, Inc.MTZ.USQualifies
    • Allegion plcALLE.USDoes not qualify
    • Allison Transmission Holdings, Inc.ALSN.USDoes not qualify
    • AngloGold Ashanti plcAU.USDoes not qualify
    • APi Group CorporationAPG.USQualifies
    • Arxis, Inc.ARXS.USQualifies
    • nVent Electric plcNVT.USQualifies
    • Comfort Systems USA, Inc.FIX.USQualifies
    • Dycom Industries, Inc.DY.USQualifies
    • Quanta Services, Inc.PWR.USQualifies
  3. Inspect each conclusion

    Open the company dossiers to examine the findings, supporting analysis and gaps. See why a company meets your requirements—or falls short.

    The framework sets the questions. The research gives you work to examine. The investment judgment remains yours.

    Check each company’s findings, evidence and gaps. See why it qualifies; keep the investment judgment yours.

    STRL.USSterling Infrastructure, Inc.
    summary.mdresearch.md

    Sterling Infrastructure, Inc. (STRL)

    Screening conclusion: QUALIFIES

    Sterling passes all four qualitative framework gates. Its strongest fit is direct exposure to AI/data-center and semiconductor/manufacturing construction through E-Infrastructure, supported by rapidly expanding backlog and a differentiated mission-critical execution niche.

    1. Capex-megacycle exposure — PASS

    Sterling’s FY2025 10-K says E-Infrastructure serves “data centers, semiconductor fabrication, manufacturing, distribution centers, warehousing, power generation and more” (STRL 10-K FY 2025).

    The principal caution is that data-center revenue is not separately reported and the business retains fixed-price contract, customer-concentration, and project-cancellation risks.

  4. Compare what holds up

    Bring the company research together to compare the findings. For screening work, rank researched companies against your own investment priorities.

    Research-only tasks keep the findings without a ranking. When you use a ranking, the order expresses relative fit to your rubric.

    Compare company findings. For screening, rank by fit to your rubric; research-only tasks retain the findings without ranking.

    Ranked companies (26)

    Ranked by ELO score from company-by-company comparisons.

    Search by name or ticker ELO band Export to Excel
    RankCompanyELO
    1STRL.US Sterling Infrastructure, Inc.1,355.5
    2MTZ.US MasTec, Inc.1,347.9
    3NVT.US nVent Electric plc1,346.2
    4FIX.US Comfort Systems USA, Inc.1,341.9
    5DY.US Dycom Industries, Inc.1,324.4
    6PWR.US Quanta Services, Inc.1,276.5

Frequently asked questions.

What is Mosaic?

Mosaic is Portrait Research’s AI workflow for investigating a set of companies against your investment framework. It applies the same research questions company by company and retains the findings for you to review.

Use Mosaic to assess a thesis across a sector, investigate a watchlist or build a shortlist. You can collect the same information for each company without ranking them, or screen and rank candidates against the criteria you define.

What investment research can I run with Mosaic?

Mosaic supports research questions that need to be investigated consistently across companies. For example, you can ask it to:

  • Assess software growth quality through retention, margins and evidence of revenue durability.
  • Investigate pharma patent-cliff exposure and the pipeline intended to replace lost revenue.
  • Examine semiconductor companies’ disclosed China revenue exposure, export-control impacts and management’s response.
  • Identify industrial companies with direct exposure to data-center power and cooling, and examine the supporting disclosures.

Your brief determines the questions, period and evidence required. These are examples of research you can request; the findings depend on the sources available for each company.

Can Mosaic use my firm’s investment framework and company universe?

Yes. Describe your research in Mosaic’s Easel panel: the companies, questions, qualifying conditions and evidence you want examined. Start with a company list or CSV, or use quantitative filters to define the universe within Portrait’s available coverage.

Mosaic saves the instructions as a reusable framework. You can inspect and revise them before running research, including whether you want information gathered for each company or an assessment against screening conditions.

How is Mosaic different from a stock screener or an AI research chat?

Mosaic turns a research brief into a repeatable company-by-company workflow, with retained dossiers and optional ranking. A stock screener helps narrow a universe using filters; a research chat helps you explore questions and follow up on individual findings.

Use Mosaic when the same question needs investigation across a set of companies—for example, whether reported growth reflects durable customer demand, or whether exposure to a theme is supported by company disclosures.

Compare Portrait’s research workflows
What sources does Mosaic use, and can I check the findings?

Mosaic researches company filings, earnings-call transcripts, investor presentations and financial data available through Portrait. Coverage varies by company and period, and access to licensed sources depends on your firm’s data entitlements.

You can inspect retained company summaries and supporting research files, then check material claims against accessible source evidence. AI-generated conclusions still need analyst review. Missing disclosure or an incomplete research task should not be treated as an investment verdict.

See how to review Mosaic research
Can I review the research before running Mosaic across a larger universe?

Yes. Start with a research trial of up to 20 companies. Inspect the company findings, check how your instructions were applied and review the measured usage before deciding to expand. For screening work, you can also rank the trial.

Trials consume Mosaic credits. A completed trial under the current research instructions is required before a full run; revising those instructions requires a new trial. The 20-company cap applies to the trial, not the full research universe.

Review the sample-to-full workflow
How does Mosaic rank companies, and can I export the results?

For screening work, Mosaic compares researched companies against your ranking criteria. The ordering and scores express relative fit to that rubric, rather than probabilities, valuations or expected returns. Research-only tasks do not rank companies.

You can review the underlying company research and export the ranking table to Excel, including the company, ticker, rank, score and summary. The workbook contains the ranking results; it does not include every company research file.

Can I apply my framework at a historical date?

Where Mosaic backtesting is enabled, you can create a framework at an available completed month-end snapshot. The date is fixed for that framework, and historical research follows the supported data and document cutoff.

Historical research has limits: company relationships use the current graph, and instructions can contain hindsight. It is not a guarantee of a fully point-in-time or bias-free investment backtest. Confirm availability and the appropriate scope for your team.

Read the historical research boundaries
How do Mosaic access and pricing work?

Mosaic is available in the web app for enabled accounts with Easel access, relevant data entitlements and a separate Mosaic credit balance. Research, including trials, and ranking use those credits.

A completed trial helps estimate the usage of a larger run; the estimate is not a fixed price. A Mosaic demo is an opportunity to discuss your research scope, source requirements, access and usage with Portrait.

Read about usage and estimates
What should I bring to a Mosaic demo?

A research question and an example company universe are a useful starting point. You do not need a finished framework: bring the questions your team asks repeatedly and the evidence you would want to review.

Use the demo to explore how Mosaic could fit your research process, from defining the brief to reviewing company findings and deciding whether a ranking is useful.

Book a Mosaic demo

See your investment framework at work.

Mosaic applies your questions and criteria company by company across an investment universe, with the research behind the results.

Book a Mosaic demo

We’ll use these details only to respond to your request. See our Privacy Policy.