Five bets are horizontal by construction. Building any of them per-vertical is the failure mode.
Three are product builds. Two are decisions this session must not defer. All five fund once and land in all nine sub-industries.
Every platform bet has a dated forcing function behind it.
| Bet | Type | Forcing function | Lands in |
|---|---|---|---|
| H1 Single traceability object | Product build | Battery passport 18 Feb 2027, fixed and non-slipping | All nine |
| H2 Landed cost and trade-authority agility | Product build | Four legal authorities in six months | All nine |
| H3 As-maintained configuration with statutory trace logs | Product build | Machinery Regulation 20 Jan 2027; CRA Art 14 11 Sep 2026 | Machine builders first |
| H4 Own, or explicitly rent, an OT and semantic data layer | Defensive decision | Schneider–Cognite, $3.1bn, agreed 30 Jun 2026 | Existential |
| H5 Publish an AI commercial model | Commercial decision | Buyers now ask in every RFP | Every deal |
Five regimes are converging on the same data object. Build one object, not five features.
One item/lot/serial-level, supplier-attributed, evidence-linked object with a unique identifier, machine-readable carrier, differentiated access rights and 5-to-10-year (or lifetime+10) retention. Serve DPP/ESPR, Battery Passport, EUDR, CRMA, PPWR, FSMA Rule 204 and steel melt-and-pour from it.
Risk: DPP registry status is unconfirmed and Art 15 customs obligations only start when the registry is operational. Scope to the dated regimes — battery, EUDR, melt-and-pour, FSMA — and treat DPP as the generalisation, not the anchor.
The pain is not the tariff rate. It is re-deriving cost every time the legal basis changes.
Four legal authorities in six months, with the rate barely moving. Master data, landed cost and HTS classification have to be re-derived each time.
No competitor is using the authority-churn framing. Everyone is selling tariff impact dashboards against a rate that is roughly stable. Manitowoc filed a CBP prior disclosure on five years of its own Section 232 methodology — that is the buyer.
Fed Beige Book, Jul 2026: pass-through is heterogeneous — "some contacts reported passing along all the cost increases, some passed on only part, and some did not pass along."
The Machinery Regulation turns every field modification into a potential conformity event.
Treat "what software version is on this asset, who changed it, and when" as a first-class EAM/FSM object with statutory retention: machine software self-inventory, a 5-year safety-software upload trace log, digital instructions for lifetime+10 years, and modification records that carry conformity consequences.
Do not become a security vendor. The defensible argument is work management, not threat detection. No competitor is publicly positioning against the Machinery Regulation software-inventory duty.
A partnership is fine until the partner buys the layer above you.
Make a deliberate build / buy / deepen-partnership call on time-series, telemetry and a governed semantic layer. The status quo is an unmanaged dependency, not a strategy.
Aggravating factor: no public statement defines competitive boundaries, overlap rules or co-sell economics in the AVEVA–IFS partnership. That is a real commercial gap and it should be closed internally before the Cognite deal closes, not after.
If buying, the target profile is a time-series/anomaly-detection and semantic-layer asset — not another application. If renting, the deliverable is a written boundary and co-sell agreement with AVEVA and Siemens plus a documented multi-source ingestion posture, so no single partner's M&A can strand the roadmap.
Every serious competitor has published a commercial model for agents. IFS has not.
"We'll tell you later" loses to Infor's "it's all included". With an IPO signalled for 2027–28, a coherent public model is also an equity-story asset — the market repriced the seat model in public on 3 February 2026, when Thomson Reuters fell 18%, RELX 14% and Wolters Kluwer 13%.
Adjacent decision. Panorama argues buyers are now "effectively evaluating two vendors" — the application and the model behind it. Oracle embeds Gemini, SAP puts Claude behind Joule, Microsoft runs OpenAI. The Nexus Black/Anthropic relationship is a live differentiator and a live single-model-dependency risk. State a model strategy alongside the pricing model.
| Vendor | Published model |
|---|---|
| Oracle · Infor | Included |
| Epicor | Outcomes-based plus a free marketplace |
| SAP | AI Units, moving to outcomes — prices compute, not outcomes, pushing ROI risk onto the buyer |
| Microsoft | Copilot Credits, hard cut-off at 125% of prepaid capacity |
| ServiceNow | Re-bundled into tiers |
| Salesforce | Three models simultaneously; rate card down to $0.50 per field-service appointment |
| IFS Loops · Nexus Black | No published model. Nexus Black is explicitly bespoke co-investment |
Ordered by conviction, not by TAM. Two of the nine are deliberately "do less".
| # | Sub-industry | Fit | Pool 2026 → 2031 | CAGR | The bet |
|---|---|---|---|---|---|
| 01 | Industrial & heavy equipment (MTO/ETO) | 5/5 | $4.65 → $7.1bn | 8.9% | Build the servitization operating system for capital equipment |
| 02 | Aerospace & defence | 5/5 | $3.72 → $5.7bn | 8.8% | Double down on MRO plus defence industrial-base readiness |
| 03 | Energy equipment & renewables | 5/5 | $0.62 → $0.9bn | mfg slice only | Follow the value inversion from the factory to the installed base |
| 04 | Automotive & mobility | 3/5 | $7.75 → $10.9bn | 7.07% | Win on warranty and aftermarket cost intelligence, not volume manufacturing |
| 05 | Metals & building products | 4/5 | $1.24 → $1.8bn | 8.2% | Melt-and-pour provenance plus asset-performance cost-out |
| 06 | Food & beverage / CPG | 4/5 | $3.72 → $5.0bn | 6.0% | Use compliance-grade traceability as the wedge into the mid-market |
| 07 | High-tech, electronics & semis | 3/5 | $4.34 → $7.2bn | 10.7% | Own capital-project-to-operations for fabs and EMS, not fab MES |
| 08 | Chemicals & process | 3/5 | $2.48 → $3.9bn | 9.3% | Lead with carbon and asset economics; partner for process physics |
| 09 | Life sciences & pharma | 2/5 | $2.48 → $4.6bn | 13.0% | Enter narrowly through equipment qualification — or stay out on purpose |
In capital equipment the installed base is the product.
BCG: a quarter to half of industrial companies fail to realise their services potential specifically because of "a lack of information about installed equipment, including its condition, location, usage, and service history." That is a description of a missing installed-base register, not a missing analytics tool.
Do not size this from ETO software reports. Four houses size ETO at $1.9bn to $23.3bn — a 12× spread. Conflict It is unusable and must be built bottom-up. June alone ran +21% real YoY on large-plant business, which favours large-capital-project and multi-jurisdiction execution over volume-manufacturing efficiency.
Demand is solved. Capacity is not. Sell into the constraint.
The cleanest demand-plus-pain combination in the pack. The primes' constraint is capacity, not demand.
Resolve before committing investment. Aviation MRO software CAGR is 2.57% (Fortune) versus 8.0% (Straits) on the same ~$8bn base Conflict — a $7bn divergence by 2034. The 2.57% figure was arithmetically verified as a genuine analytical position, not a typo. This single choice swings the whole A&D aftermarket business case.
The largest pool in the table is the wrong fight. Win warranty, not volume.
The best-evidenced opportunity in the pack, because it is SEC-audited rather than survey-based.
The buying trigger here is cost control and compliance, not growth. Build warranty forecasting and change-of-estimate analytics as a named, marketed capability; supplier-risk-linked planning against a distressed tier-2 base; and battery-passport traceability via H1, where the 18 Feb 2027 date is fixed.
Do not chase volume MES or plant-floor execution here. This is where SAP, Rockwell (Lucid) and Siemens are strongest and where IFS has no differentiated claim.
The factories are shrinking. The installed base is not.
Record deployment — 165 GW of new wind in 2025, +40%; solar above 600 GW — alongside cleantech manufacturing investment more than halving 2023→2025 and falling again in 2026, with over 60 GW/yr of module capacity sitting idle. Building manufacturing ERP for module plants is building for a shrinking population.
Build O&M and service execution for installed renewable and grid assets — not manufacturing ERP for the factories. Leverage Copperleaf asset investment planning, already proven at Avacon, E.ON Sweden and WEL Networks, plus the Siemens autonomous-grid partnership.
Do not present the $0.62bn as the opportunity. It is deliberately small because it is only the manufacturing slice. The strategic pool for this bet sits in EAM/FSM and the utilities vertical, not in the manufacturing segmentation table.
The new steel safeguard turns trade compliance into a master-data problem with a date.
Reg (EU) 2026/1384 from 1 July 2026 cuts quota ~47% to 18.3 Mt/year, raises out-of-quota duty to 50% ad valorem, and requires importers to submit verifiable evidence of first melting and casting — Commission evidentiary specification due 31 August 2026. Quota carry-over is now conditional on average use exceeding 80% over three quarters, which makes quota availability itself a planning variable.
Historically, falling volume with rising value favours cost-out and asset-performance software over growth-driven ERP replacement. Build melt-and-pour origin as a supplier-attributed master-data object (H1), quota-aware and CBAM-aware planning, and APM cost-out against ageing brownfield plant. ArcelorMittal is already an IFS FY2025 win — use it.
Honest caveat. Metals/steel ERP, building products software and cement software are not sized discretely by anyone. The $1.24bn is our allocation and the CAGR is proxied from mining software. Estimate Treat the bet as evidence-led on regulation and sector economics, not on a TAM figure.
Compete on a dated regulation and an installed reference base — not on a growth story.
FSMA Rule 204, compliance date 20 July 2028, described in the industry as "the most significant regulation the food industry has ever faced." The 24-hour records-response requirement makes paper and spreadsheets untenable. FMI notes adoption broadening "especially in mid-sized companies and regional food processors" — which is a segment, not a rounding error.
Build lot and batch genealogy engineered for a 24-hour recall response, riding the H1 traceability object. Frame cyber as production continuity, not security. Pair with connected-worker execution — Poka already carries Nestlé, Tetra Pak and Mars — plus IFS Cloud references at Coca-Cola and William Grant & Sons.
At ~6% this is the slowest-growing sub-industry in the table.
Sell the fab build, not the fab floor.
The sharpest capex upcycle in the pack, and the fastest-growing manufacturing-ERP vertical at 10.70%. SEMI's CEO: "AI is resetting the scale of semiconductor manufacturing investment."
Timing risk. The US incentive cliff is December 2026 and there is no public record of final CHIPS awards for the 12 firms with preliminary agreements. Weight the bet toward EU Chips Act, Japan, Korea and India rather than assuming US federal money.
Be the carbon ledger of record. Partner for the process physics.
The trigger is not growth. It is a compliance and cost crisis. CBAM's definitive regime went live 1 January 2026 with real published prices — €75.36/tCO2 Q1, €75.28 Q2 — and a 21-month liability accrual before payment, since certificate sales start only February 2027 and the first declaration is due 30 September 2027. EUROFER's April 2026 nine-point list says the mechanism still leaks.
State this openly. The weakest-sourced sub-industry in the pack: no Gartner, IDC or LNS sizing exists and both CAGR sources are low-tier. Energy cost also cuts both ways — DSAG's chairman explicitly names energy prices as a reason ERP investment is being postponed.
Fastest growth, weakest fit. Enter narrowly, or decide not to enter and say so.
The fastest-growing software vertical in manufacturing at 11.2–15.5%, and the only one where every independent house agrees on high growth. Pharma MES $2.37bn (2025) → $4.62bn (2030) at 14.3%; life-sciences QMS $3.7bn → $10.1bn at 13.8%. On-premise still held the largest MES share as recently as 2023, so the cloud runway is real.
An explicit "no" is a legitimate outcome of this session; drift is not. The failure mode is a half-funded pharma push that neither wins regulated accounts nor frees capital for bets 01 and 02.
SAP's ECC base must move, cannot get AI without moving, and increasingly does not trust the destination.
Business Suite 7 mainstream maintenance ends 31 Dec 2027, extended to 2030 at a premium. Joule requires shifting at least 50% of maintenance spend to cloud.
The differentiated claim is composite, not financials-first. An ECC migration is usually scoped as a financials and core-ERP replacement. IFS's claim is that asset, service and project execution belong in the same replacement decision — and SAP's own asset/service estate is mid-consolidation while that decision is being made. Five products (MRS, FSM, RSH, C4P RM, MSO) are merging into SAP FSA, MRS sunsets 31 Dec 2030, two parallel vocabularies are in play, and ERP integration only reaches GA in Q4 2026. DSAG publicly called MRS re-licensing "unacceptable."
Every ECC account that migrates is removed from the pool permanently.
The 2027–2030 stretch is the whole opportunity — which is exactly the FY27–29 horizon this session is setting.
Run the wedge hardest in EMEA. DACH has the highest documented ECC concentration, Europe is the largest region for ESG software at 34.9%, and the regulatory forcing functions — Machinery Regulation, CBAM, DPP, CSRD, NIS2 — are European. North America's manufacturing capex has reversed and the incentive cliff arrives Dec 2026.
There is no credible public ERP win-rate or displacement dataset — not from Gartner, IDC, Panorama or Third Stage. Any win-rate number used in the session must come from IFS's own CRM.
Two competitors have published the research that argues IFS's composite case better than IFS marketing does.
Never use these. The widely circulated "189% average overrun / 215% in manufacturing / 73% of discrete projects fail" figures are not in the Panorama report.
Feedback on this section
Where is this wrong? Comments are visible to everyone reading the portal, and your name is attached.