← back to ranking

OC · Owens Corning

Construction · mkt cap $9.7B · calls: Q1 FY2026 vs Q4 FY2025
47.0 conviction · conf-adj 47

conf 2/10 partial

enthusiasm:18.0 · trend:8 · quantifies:0 · impact:0 · under_radar:14 · credibility:0 · business_impact:4 · disruption:0 · commitment:0 · confirmation:3

Enthusiasm latest 6 / prev 5 (rising)

Owens Corning’s AI story is bifurcated: internal use centers on predictive manufacturing (20k+ sensors, ~40 plants) and, as of the prior call, supply-chain optimization agents in North American insulation with planned enterprise scaling and future generative/agentic AI under a new CIO—without dollarized savings or margin attribution. The latest call adds an end-market angle: insulation nonres demand from AI/data-center buildout is strong enough to constrain some product lines, but management does not quantify AI-linked revenue. Credibility is moderate on operations (concrete deployment scale) and low on financial impact (no P&L numbers, no analyst-driven AI disclosure in Q&A beyond demand color).

GROUNDED NEXT-FY IMPACT vs CONSENSUS

Grounded on actual base — revenue $10.1B · net income $-0.5B · net margin -5.2% · diluted EPS -6.26

These are next-fiscal-year annual uplift estimates, not next-quarter numbers.

Aggregate next-FY est. rev uplift: % · next-FY EPS uplift: % · vs analysts: unclear · priced in: low (model's call-read: high; verdict above is the hard-data one used for ranking) · confidence: 2/10

ClaimFigureArithmeticNext-FY Rev %Next-FY EPS %
20,000+ process sensors monitored by AI for real-time risk alerts
productivity · soft
over 20,000 sensorsOperational count, not a dollar figure. No $ saving, % productivity, or FTE delta disclosed; cross-claim 20,000 sensors ÷ ~40 plants ≈ 500 sensors/plant is deployment intensity only. Cannot compute after_tax_saving = saving_$ × (1-0.21) with saving_$ undisclosed. EPS denominator is GAAP NI = -$522M (loss), voiding any eps_uplift_pct. No revenue/cost base for this claim appears anywhere in inputs — genuinely unanchored.
AI monitoring deployed in ~40 plants across 3 businesses, expanding
productivity · soft
nearly 40 plantsDeployment-footprint count across Roofing, Insulation, Doors, not a dollar figure. Quality/yield/uptime benefit is plausible but unquantified ('predict risk before it impacts safety, quality or productivity'); no revenue lift, cost saving, or margin $ disclosed. Cannot size; same loss-making EPS denominator problem.
Nonres product lines sold out / near sold out on AI/data-center demand
revenue · soft
'close to being sold out or are sold out in specific product lines' (no revenue $)Revenue tailwind comes from OC selling building products INTO the external AI/data-center buildout (picks-and-shovels SUPPLIER position), not AI improving OC's own products — tagged supplier-side and excluded from the adopter headline. Sold-out status implies full utilization but not the $ magnitude of incremental vs baseline. Management gives no revenue $, no affected-segment size, no incremental volume: rev_uplift_pct = 100×(undisclosed incr_$/$10,103M) is uncomputable. Unanchored — soft.

Assumptions: No phasing applied — none of the three claims provide a next-FY revenue or savings $. Defaults not invoked because no claim is dollar-anchored. For any hypothetical revenue claim, incremental margin would use operating margin ($1,720M/$10,103M = 17.0%), not GAAP net margin (-5.17%, distorted by the one-off-driven -$522M net loss); tax rate 21% for any hypothetical cost saving. EPS sizing would use consensus adjusted NI base (~$1,032.8M FY25E, EPS $12.21), NOT GAAP -$522M NI, which makes any EPS-uplift % meaningless per guardrail and is set null. Segment mapping: claims 1-2 → internal manufacturing ops (cost/quality, adopter); claim 3 → nonres roofing/insulation demand (supplier into buildout). Adopter aggregate excludes the supplier-side data-center demand claim.

Top line: Adopter-side topline uplift: 0% computable. The only revenue-adjacent signal is supplier-side — sold-out nonres product lines feeding AI/data-center/reindustrialization construction — but management gave zero revenue $, no affected-line size, and no incremental volume, so 100×claim_$/$10,103M is uncomputable and supplier_rev_uplift_pct is also null. This is a demand tailwind OC benefits from as a materials supplier into the buildout, not AI improving OC's own products; per the adopter-ranking rule it is excluded from the headline and cannot be sized. est_rev_uplift_pct = null.

Bottom line: Adopter-side bottom-line uplift: 0% computable. The 20,000-sensor / ~40-plant operational AI deployment (≈500 sensors/plant) is a real footprint targeting safety/quality/productivity but carries no disclosed opex savings, downtime-avoidance $, or productivity %, so after_tax_saving = saving_$ × 0.79 cannot be computed. Separately, OC is GAAP loss-making (NI = -$522M, -5.17% margin), so EPS-uplift % off that base has no valid denominator; even against consensus adjusted NI of ~$1,032.8M (EPS $12.21) there is no savings $ to flow through at 21% tax. est_eps_uplift_pct = null per the loss-making guardrail.

[impact n/m (all claims soft/unanchored); EPS uplift n/m (loss-making base)] Consensus FY25E: revenue ~$10,132M (+0.29% vs actual $10,103M), EPS $12.21, NI ~$1,033M (12 analysts). Trajectory: FY23→FY24 revenue +10.4% ($9,872M→$10,902M), FY24→FY25E -7.1% ($10,902M→$10,133M); EPS peaked $15.51 in FY24, falling to $12.21 in FY25E (-21.3%). With zero hard adopter-side rev/eps uplift computable (all three claims soft), there is no quantifiable gap above consensus. The supplier-side sold-out/data-center demand narrative is a known cyclical demand fact already embedded in the nonres strength behind the -7% revenue normalization vs the FY24 peak — not new unmodeled upside. Forward statements (supply-chain AI agents, generative/agentic AI via new CIO, plant-monitoring expansion) contain no $ anchors. Net: no positive consensus gap can be demonstrated in numbers — treat as priced in.

MODEL CONSENSUS (impact)

partial

Both answers fully agree on all numbers (all null), soft flags, priced_in=high, unclear, confidence 2. Only minor type-label and margin-basis wording differed.

Conflicts reconciled
FieldOpus 4.8GPT-5.5
Rev uplift %
EPS uplift %
Priced inhigh
vs analystsunclear
Confidence2
Top lineNo sizable adopter-side AI revenue is quantified. The only AI-linked revenue signal is supplier-side: nonres product lines feeding data-center/reindustrialization demand are 'sold out or close to sold out,' but management gave zero revenue $, no affected-line size, and no incremental volume — so 100*claim_$/$10,103M is uncomputable. This is a demand tailwind OC benefits from as a materials supplier into the buildout, not AI improving OC's own products; per the adopter-ranking rule it is excluded from the headline and cannot be sized either. est_rev_uplift_pct = null.
Bottom lineThe two adopter AI initiatives (20,000+ sensors; ~40 plants) target safety/quality/productivity and likely yield real but undisclosed cost/yield benefits. None carries a $ saving or % productivity figure, so after_tax_saving = saving_$ x 0.79 cannot be computed. Separately, OC is GAAP loss-making (NI = -$522M), so an EPS-uplift % has no valid denominator — even a known saving would produce a meaningless/negative figure. est_eps_uplift_pct = null per the loss-making guardrail.
ReasoningNothing AI-specific is quantified to dollars, so there is no incremental figure to compare against the consensus trajectory. Consensus already carries the nonres/data-center demand: FY25 revenue avg $10.13B (vs $10.10B actual) and net income avg ~$1.03B — the sold-out commentary is a known cyclical/demand fact embedded in those numbers, not a new uplift. The internal sensor/plant AI program is an efficiency story with no disclosed magnitude, so it adds no measurable gap above consensus. Net: no positive consensus gap can be demonstrated in numbers -> treat as priced in.

Rows highlighted where the two models disagreed.

QUANTIFICATIONS
Process sensors monitored with AI: over 20,000 (Today (Q1 FY2026 prepared remarks), bottomline)
“Today, we are monitoring and analyzing over 20,000 process sensors in our plants using AI to provide real-time alerts that help our teams predict risk before they impact safety, quality or productivity.”
Plants with AI monitoring deployment: nearly 40 plants (Current across Roofing, Insulation, and Doors (Q1 FY2026), bottomline)
“These capabilities are deployed in nearly 40 plants across our 3 businesses with plans to continue expanding.”
Nonres product-line capacity vs AI/data-center demand: close to being sold out or are sold out (specific product lines; no revenue $) (Current nonres environment (Q1 FY2026 Q&A), topline)
“Anything related to AI, data centers, the reindustrialization of North America has been strong for us on the nonres side. So in some of those areas, we're close to being sold out or are sold out in specific product lines that feed into those high-growth segments within nonres.”
PAST (realized)
CURRENT (now)
FORWARD (guidance)
TRACK RECORD — PROMISE vs DELIVERY

/100 (no quantified promises)   no-quantified-promises  6 calls reviewed

Across six calls Owens Corning never made a quantified AI promise (number plus timeframe); AI talk is limited to qualitative scaling of supply-chain agents and later operational metrics (20k sensors, ~40 plants) with no prior measurable targets to score.

PRICED-IN (REFINED)
LOW (room left)

Est. revisions flat  ·  Fwd P/E 10.1  ·  EV/Sales 1.6x

AI claim maps to Roofing, Insulation, Doors

Analyst ratings are stable (buy/hold counts barely moved over six months; consensus Hold), price targets show only a modest near-term lift (lastMonthAvg 137.25 vs lastQuarterAvg 135.67) while remaining below lastYearAvg 140.06, and forward consensus EPS steps down from 15.51 (2024) to 12.21 (2025)—not a pattern of upward revisions baking in AI upside. Valuation is not stretched at ~10x forward P/E and ~1.6x EV/Sales, well below levels where efficiency narratives are typically capitalized. AI-driven manufacturing, pricing, or demand benefits would most plausibly flow through Roofing and Insulation (largest revenue lines), yet neither estimate momentum nor multiples suggest that thesis is already in the price.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
2Q4 FY20242Q1 FY20252Q2 FY20252Q3 FY20252Q4 FY20257Q1 FY2026

AI enthusiasm across 6 calls — trend ↗ rising

Five calls with no AI story; Q1 FY2026 added specific plant-sensor predictive monitoring across ~40 sites.

RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY

5/10 qualitative impact   moderate  medium-term · mixed evidence

Where AI matters: plant predictive maintenance & supply-chain ops

OC has credible adopter-side scale—20k+ AI-monitored sensors across ~40 plants plus insulation supply-chain agents—but no dollarized savings or margin attribution, and data-center/nonres strength is supplier demand into the AI buildout, not AI improving OC’s own products.

Caveats: No disclosed opex, downtime, or EPS uplift from operational AI despite scale metrics; Generative/agentic AI and enterprise agent rollout remain forward narrative without P&L proof; Sold-out nonres lines on data-center demand are cyclical supplier tailwind, not durable adopter-side AI monetization; Industry-wide efficiency gains could modestly cap pricing power without commoditizing the product

AI DISRUPTION / CANNIBALIZATION RISK  tailwind · 2/10

Core revenue is physical insulation, roofing, and doors where AI mainly optimizes manufacturing and logistics; it does not automate away installed building-envelope materials or structurally compress billable units the way services/labor-arbitrage models face.

OPTIONS / MARKET STRUCTURE

option liquidity: fair

ATM IV
TYPICAL BID-ASK
OPEN INTEREST

proxy inputs — dollar-ADV $161M · beta 1.35 · px $120.22

source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.

CONFIRMATION — INSIDERS · 13F · LANGUAGE
Mixed — insiders selling, institutions adding, management language 6/10 measured.
INSIDERS selling 2 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 100 new / 117 closed positions; 356 increased / 179 reduced; institutional ownership +2.50pp; -15 net 13F holders
MGMT LANGUAGE 6/10 measured Brief ops AI block: firm present deployment and sensor/plant counts; future expansion hedged, no AI P&L targets.
commit “Today, we are monitoring and analyzing over 20,000 process sensors in our plants using AI”
commit “These capabilities are deployed in nearly 40 plants across our 3 businesses”
hedge “with plans to continue expanding”
VERBATIM AI QUOTES
“This includes expanding our use of intelligent monitoring and AI-enabled tools to improve asset reliability, reduce unplanned downtime and support a structurally lower cost position. Today, we are monitoring and analyzing over 20,000 process sensors in our plants using AI to provide real-time alerts that help our teams predict risk before they impact safety, quality or productivity. These capabilities are deployed in nearly 40 plants across our 3 businesses with plans to continue expanding.”
— Brian Chambers, Q1 FY2026
“Anything related to AI, data centers, the reindustrialization of North America has been strong for us on the nonres side. So in some of those areas, we're close to being sold out or are sold out in specific product lines that feed into those high-growth segments within nonres.”
— Todd Fister, Q1 FY2026
“We are also enhancing our winning cost position through the use of advanced analytics and AI to drive efficiency, support customer growth and strengthen market leadership. For example, we are applying AI through the use of supply chain optimization agents that help us respond quickly to network adjustments and maintain strong service levels at reduced cost. This capability is being used in our North American fiberglass insulation business today, but will be scaled to our other businesses throughout the year.”
— Brian Chambers, Q4 FY2025
“To accelerate our digital efforts, Annie Baymiller was recently promoted to the role of Executive Vice President and Chief Information Officer. She will lead the advancement of our digital technology capabilities and next-generation tools, including generative and agentic AI that will be instrumental to unlocking new capabilities and generating additional value.”
— Brian Chambers, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Stephen Kim (Evercore ISI)): Curious if you could give us a little bit of insight into what you're seeing in the supply-demand dynamics across [Insulation segments] North America resi, nonres, and Europe for the rest of the year.
A: Todd Fister: On nonres, "Anything related to AI, data centers, the reindustrialization of North America has been strong for us on the nonres side" with some product lines "close to being sold out or are sold out" that feed high-growth segments; res described as stable with disciplined idle; Europe stable with pockets of strength.