BLUF: The architectural fight in smart-building AI for the next 36 months is not about who has the best chiller, the best controller, or the best dashboard. It is about who owns the orchestration layer above the OEM. A working Series A cap-table in this space — the most-watched one in CRE-tech right now — names four roles in order: the incumbent HVAC OEM, the dominant AI compute provider, the largest industrial REIT, and a top-tier global asset manager. Read the cap table as architecture, not finance, and the strategic shape of the next decade of CRE becomes legible.
The four-role cap-table read
An emerging-leader Series A in the AI-HVAC orchestration category — PassiveLogic — closed in late 2025 with a participant list that, regardless of dollar amounts, names the four roles that must coexist for the category to scale. Reading the cap table by role rather than by name is the clearest expression we have of where the value actually accrues.
| Role on Cap Table | What They Bring | What They Validate |
|---|---|---|
| Incumbent HVAC OEM | Substrate access — installed equipment base, field-services org, OEM-grade reliability commitments | The orchestration layer cannot ignore the iron. Whoever runs the layer must integrate with the largest installed-base OEMs in their region. |
| Dominant AI Compute Provider | Inference fabric, Omniverse + OpenUSD, NIM microservices, edge + cloud topology | The orchestration layer runs on a known compute and digital-twin stack that is converging fast. Single-cloud bets are getting risky. |
| Largest Industrial REIT | Demand-side validation — multi-million sqft portfolio, retrofit pipeline, named operating discipline | The product has to clear "would an institutional owner buy it?" — not "would a pilot site try it?" The REIT seat on the cap table is the institutional buyer signal. |
| Top-Tier Global Asset Manager | Capital, governance, deal-flow access to other institutional owners | The category is now allocated as infrastructure, not as software. Asset-manager participation is the structural reclassification. |
One company — the round subject — appears once in this analysis and not again. The point of the article is the shape, not the name. The next decade of category leadership will go to whoever assembles the same four-role mix earliest and most credibly. That includes incumbents who reposition, new entrants who recruit the right capital, and platform players (including AISB) whose architectural posture is consistent with the shape.
What "orchestration above OEM" actually means
"Orchestration above OEM" is a specific architectural choice with specific consequences. The orchestration layer:
- Reads from any qualified BMS or controls layer below it — Niagara, Honeywell EBI, Siemens Building X, open-Brick deployments, mixed substrates.
- Holds a coherent digital-twin state independent of the OEM — typically on OpenUSD / Omniverse or an equivalent open description.
- Runs AI workloads (fault detection, sequence-of-operations optimization, demand-response scheduling, predictive maintenance) on a compute fabric the owner can switch.
- Exposes a procurement-language interface the owner can specify in an RFP without naming the OEM. ASHRAE Guideline 36, IPMVP M&V Options A through D, and Brick Schema are the canonical specifications.
- Closes the loop back to OEM substrate via well-defined control points the OEM still warrants. The orchestration layer does not attempt to replace the OEM's safety, reliability, or warranty surface.
This is the composer pattern. Trane Cloud + ARIA + BrainBox, Schneider's EcoStruxure with the Aveva-Omniverse bridge, and Siemens Building X with the Xcelerator marketplace are three OEM-led examples of the same pattern. AISB Fusion Model is a vendor-neutral example. The OEM-led examples have a substrate advantage; the vendor-neutral examples have a portfolio-orchestration advantage. Both are legitimate plays.
The alternative: standalone productizer stacks
The opposite pattern is the productizer: ship the AI bundled with the iron, defend margin through vertical integration, lock the substrate. JCI Nantum AI (Q3 2026 commercial rollout, Q2 FY26 +45% adjusted EPS), Honeywell Forge, Carrier Abound, and Daikin Intelligent Equipment all sit in this camp. The productizer pattern has a real revenue line and a real schedule. It is not wrong. It is a different architectural bet.
For owners with a multi-OEM portfolio — the typical Fortune 500 corporate occupier, the typical institutional REIT, the typical APAC government landlord — productizer stacks force a vendor rationalization the owner did not ask for. That is the structural friction the composer camp is monetizing.
Why the cap-table read is a procurement signal
Three reasons CRE buyers should treat the four-role cap-table mix as procurement signal, not press release:
- Institutional capital has reclassified the category. Top-tier asset-manager participation on the cap table means AI-HVAC orchestration is now allocated as infrastructure — long-dated, balance-sheet-friendly, governance-heavy — not as venture software. That changes who the buyer is (asset management committee, not innovation team), the diligence depth, and the contract length.
- OEM and REIT seats validate operational fit. An incumbent HVAC OEM and a largest-class industrial REIT do not co-invest in a layer that will fail a field-services review. The roster is, in effect, a pre-IPO operational diligence pass on the architecture.
- Compute-provider seat names the substrate. When the dominant AI compute provider takes a position, the digital-twin and inference fabric becomes the de facto standard. OpenUSD, NIM microservices, and the Omniverse-based functional twin become procurement language by 2027 whether or not buyers like the dynamic.
What this means for the next 24 months of CRE-tech
Three predictions, dated and falsifiable:
- By end of Q4 2026, every major HVAC OEM RFP issued by a Fortune 500 occupier or top-quartile institutional REIT will include a written "orchestration architecture" section that maps onto the composer/productizer split. Vendors that cannot give a clean DSX-status answer (see the DSX Partner Roster) will lose pre-shortlist.
- By mid-2027, at least two more institutional-class cap tables in the AI-HVAC and AI-building-ops categories will name the same four roles. The four-role mix is becoming a category-level structural feature, not a one-off.
- By end of 2027, OEM-led composer stacks (Trane Cloud, Schneider EcoStruxure + Aveva, Siemens Building X) and vendor-neutral composer stacks (AISB and peers) will jointly hold a majority of orchestration-layer spend in DSX-IN portfolios, while productizer stacks will hold a majority in single-OEM portfolios. The market will visibly bifurcate along architectural lines.
How AISB sits in the composer pattern
AISB Fusion Model is vendor-neutral composer-by-construction:
- Above any qualified substrate — Tracer SC+, Niagara, Honeywell EBI, Building X, open-Brick. Substrate is owner's choice; orchestration is independent.
- On open standards — Brick Schema for asset description, Speckle.dev and IFC 4.3 for design and coordination interop, IPMVP for M&V claims, ASHRAE Guideline 36 + Addendum G for sequences of operation.
- Across CDE platforms — Procore, Autodesk Construction Cloud, Aconex, e-Builder, SharePoint fallback. Multi-CDE pattern detection is a keystone CRE-CON moat that single-platform AI cannot match.
- Below the owner's procurement language — every AI claim is anchored to ASHRAE, IPMVP, IBC, NEC, or NFPA section reference. Claims without an anchor auto-downgrade in confidence.
This is the explicit product posture. It is also the reason the cap-table read matters to AISB and to anyone evaluating AISB against a productizer-pattern alternative. The four-role mix is not who AISB is competing with — it is the shape of the category AISB is operating in.
What to ask the agent
- Where does my portfolio sit on the composer-versus-productizer split?
- Run my shortlisted vendors through the DSX-IN / DSX-OUT roster.
- What does cross-CDE pattern detection give me that single-platform AI cannot?
Companion reading
- DSX Partner Roster as Architecture Evidence
- The CRE Standards Gap — why BACnet/IoT data silos keep stalling AI in buildings
- The Agentic Operating Model for CRE — 4-pillar pillar page
- AISB Proof — operational evidence map
Methodology note: cap-table role analysis derived from publicly available Series A disclosures and standard institutional-capital classification. Specific dollar amounts and percentage stakes are not material to the architectural read and are deliberately omitted. Verifiable dated proof points (Trane 14K buildings, JCI Q2 FY26 earnings, Phaidra 80% Max-Q reduction, Delta 20% Taipei HQ savings) are sourced from vendor disclosures and the Q2 2026 CRE Competitor Radar.