C

Carrier Global

Global CDO AI Vision — 2026–2029

01
Global CDO AI — Third Interview Presentation

From Climate Leader
to Data Intelligence
Leader

Carrier has built the singular portfolio. The missing piece is the data and AI architecture that converts that uniqueness into durable, measurable competitive advantage in the P&L.

"Carrier isn't in the climate business. It's in the business of making the planet work. Over the next five years, that mission and artificial intelligence will converge in ways no competitor has fully understood yet."
$0B
FY2024 Revenue
0
Buildings Connected
0
Employees Worldwide
0
Month Transformation
Scroll
02

The Pure-Play That Emerged

David Gitlin's portfolio transformation created the world's only true global pure-play in climate and smart energy at this scale. Structural transformation arrived before digital transformation.

$22.5B
Revenue FY2024
Pure-play first full year
$14.2B
Viessmann Acquisition
Jan 2024 — 80% cash, 20% stock
5.66x
Debt/EBITDA Q1 2026
vs. 3.47x historical median
$4.95B
Fire & Security → Honeywell
Portfolio discipline signal
Revenue Trend 2021–2024 ($B)
Business Segments — Revenue Mix
AI Maturity Score
32 / 100
Structural transformation ahead of digital
Portfolio Transformation Timeline
Jan 2024
Viessmann acquisition closed — $14.2B, entry into European residential heat pumps & 150 years of equipment data
Jan 2024
Commercial Refrigeration divested — capital concentration into smart climate
Jun 2024
Fire & Security sold to Honeywell for $4.95B — pure-play transformation complete
2025
QuantumLeap reaches $1B in data center revenue; Abound hits 63,000+ buildings
Jul 2026
Bobby George exits as CIO — opportunity to reframe function as competitive advantage architect
03

The AI Opportunity

The financial argument is direct: every percentage point reduction in unplanned downtime of commercial HVAC systems represents tens of millions in recovered value — and sustainable pricing power for Carrier.

Value Capture by AI Use Case ($M, 3-Year Horizon)
AI Maturity vs. Competitors (Radar)
🏢
Abound Platform
63,000+ buildings · 640M sq ft · 650M kWh saved in 12 months · 40,000 service dispatches avoided
QuantumLeap Data Centers
$1B$1.5B revenue 2025→2026 · Orders growing 4x YoY in Americas · Market to $20B by 2029
🌡️
Viessmann Data Moat
150 years of equipment data · Millions of European homes connected · Most undervalued asset in the $14.2B acquisition
Data Center Cooling Market Trajectory ($B)
04

Transformation Roadmap

Three simultaneous moves: unify the data fabric across Abound, Viessmann, and QuantumLeap; scale AI impact from 63,000 connected buildings to 100,000 field technicians; build the governance system that makes it all auditable.

Phase 1: Foundation (0–12m)
Phase 2: Acceleration (12–24m)
Phase 3: Leadership (24–36m)
Theme: Architecture & Honest Audit
Investment: $220M · Deliverable: Unified data architecture + CAICE CoE operational
Month 1–2
Data landscape audit: Abound API inventory, Viessmann European system mapping, QuantumLeap instrumentation assessment
Month 3–4
CAICE Center of Excellence launched: 80 FTEs hired (industrial data engineers, not theoretical data scientists)
Month 5–6
Unified data lake on AWS (S3 + SageMaker) — first cross-platform data pipeline: Viessmann → Abound schema
Month 7–9
500 AI Champions nominated across BUs; GenAI Task Force elevated to Board-level visibility
Month 10–12
AI Governance Framework live: audit trails, human checkpoints, model review cadences — before first production deployment
Theme: Demonstrate Traction with Real Money
Investment: $340M cumulative · Deliverable: Cross-segment models in production, measurable margin impact
Month 13–15
First cross-segment model: Viessmann EU data improving Abound predictive maintenance in North America commercial
Month 16–18
QuantumLeap → Abound flywheel: data center thermal patterns feeding commercial building optimization models
Month 19–21
TechVantage AI modules live: 15,000 field technicians with AI diagnostic tools; measurable reduction in mean repair time
Month 22–24
Year 2 checkpoint: margin impact in basis points, contract renewal rates, Abound vs. OpenBlue wins measured
Theme: Scale Without Compromise
Investment: $820M cumulative · Deliverable: Carrier = global reference for intelligent physical infrastructure
Month 25–27
Models proven at 1,000 buildings deployed to all 63,000+ Abound buildings with automated drift detection
Month 28–30
Viessmann European Efficiency Service launched: most accurate residential energy efficiency service in Europe
Month 31–33
100,000 field technicians AI-augmented globally; Abound winning on data lock-in, not feature comparison
Month 34–36
Debt/EBITDA trajectory normalized by software margin mix; Carrier AI presented at AWS re:Invent as reference customer
Cumulative Financial Impact ($M)
Investment Allocation by Phase ($M)
05

Governance Framework

Companies that deploy AI fastest in industrial environments are the ones that built their governance framework before their first incident — not after. Speed and safety are not opposites.

AI Council Structure

Board Level

Technology & Innovation Committee (Neil Barua chair) — quarterly AI risk & opportunity review

Executive AI Council

CDO AI + CFO + Chief Legal + CHRO + 3 BU Presidents — monthly; owns AI investment prioritization

CAICE (Center of AI & Competitive Excellence)

250 FTEs by Year 3 — data engineers, ML ops, AI safety specialists, domain experts embedded in BUs

BU AI Champions (3,000 certified)

One per major functional team — first responders for AI adoption, local governance stewards

Three Non-Negotiable Axes

1. Complete Transparency
Every material AI decision has an auditable trace reviewable by an external auditor. No black boxes in critical infrastructure decisions. Explainability documented before deployment, not retroactively.
2. Verified Human Control
No autonomous system operates in critical physical environments (hospitals, data centers, cold chain) without an explicit, documented human checkpoint. Not a checkbox — a verifiable gate with ownership.
3. Mandatory Continuous Improvement
Models have review dates, not indefinite retirement. Mandatory quarterly performance audits vs. baseline. Models degrading below threshold are automatically flagged for retraining or decommission.

Data Mesh Architecture

Domain-Owned Data Products
Abound DomainQuantumLeap DomainViessmann DomainSupply Chain Domain
↕ Federated Governance Layer (schema registry, lineage, PII controls)
AWS Data Lake (S3)SageMaker ML PlatformBedrock LLMsKinesis Real-Time
↕ Cross-domain consumption API
Predictive MaintenanceEnergy OptimizationField AI ToolsCustomer Efficiency SaaS
06

AI Activation Model

47,000 employees don't need a PowerPoint about digital transformation. They need honest answers to three questions they never ask aloud: Will this take my job? Can I learn it? Will my day actually be better?

Adoption S-Curve — AI-Augmented Employees
Segment Activation Progress
Field Technicians (100K global)72%
Building Operations (Abound users)85%
Engineering & R&D60%
Sales & Customer Success45%
Corporate & Support Functions30%

AI Champion Personas

🔧
Marco — Field Technician
Frankfurt · Viessmann Heat Pumps
Arrives knowing exactly what he'll find. AI diagnosed the fault before he crossed the door. His diagnosis accuracy improved 40%. He's the loudest AI advocate in his region — not because management said so, but because it made him better.
🏗️
Sarah — Building Ops Director
Chicago · 3.2M sq ft Commercial Portfolio
Reduced energy cost per sq ft by 18% in 12 months using Abound recommendations. When her lease renewals came up, she led with AI-verified energy performance data. Closed two renewals at premium rates.
Rajesh — Data Center Ops
Virginia · Hyperscaler Facility
QuantumLeap AI flagged a thermal anomaly in Rack 47 at 2am — 6 hours before threshold breach. Zero downtime. Zero data loss. He now treats the system like a co-worker with superhuman attention span.
07

AWS: From Vendor to Multiplier

Carrier's AWS relationship is not aspirational — it is deployed operational infrastructure. Abound runs cloud-native on AWS and is listed in AWS Marketplace. Lynx cold chain was co-developed with AWS. The strategic question is how to extract 3x more value from what's already built.

Current State — Fragmented

3 Disconnected Data Silos
Abound
AWS S3 + IoT
Viessmann
On-prem EU
QuantumLeap
Hybrid
No common schema · No cross-segment ML · Heavy reliance on AWS Professional Services & Accenture · High marginal cost per model deployment

Target State — Strategic Alliance

Unified Carrier Intelligence Fabric
AWS Data Lake (S3 + Lake Formation + Glue)
All 3 domains · Common schema · GDPR + SOC2 compliant
SageMaker
ML Ops Platform
Bedrock
Foundation Models
Greengrass
Edge AI (HVAC units)
Marketplace
100K+ enterprise buyers
Co-development agreement · Shared roadmap visibility · AWS as go-to-market multiplier
AWS Migration — 3-Phase Capability Unlock
08

Pattern Intelligence

What the market sees. What the board sees. And what a systems thinker sees that neither group has connected yet.

The Viessmann Data Moat
Analysts model the $14.2B acquisition as: heat pumps + brand + European distribution network. Nobody has modeled the 150 years of equipment operational data — hourly energy consumption signatures from millions of European homes, fault prediction patterns with decades of history, climate-region correlations no competitor holds at that geographic scale. That dataset is platform value, not product value. The market hasn't discounted it yet.
Contrarian AssetPlatform Play
The QuantumLeap → Abound Flywheel
QuantumLeap and Abound are not two competing bets fighting for investment budget. They are two halves of a unique sector flywheel. Data centers generate the richest, most densely instrumented operational data in built infrastructure — temperature per rack, consumption per server, efficiency per cooling unit. Those patterns can directly retrain Abound models for complex commercial buildings. The company that connects these two data flows first will build optimization models no single-segment specialist can replicate.
FlywheelUnique Moat
The Real Battle: Data Lock-In vs. Feature War
The Abound vs. OpenBlue (Johnson Controls) competition isn't decided in Gartner quadrant comparisons. Johnson Controls is already specifying digital connectivity as a contractual condition in their bids — using software as a lock-in mechanism, not a feature differentiator. Carrier has the largest installed base of equipment in the market, which is the most valuable data source. The winning strategy: make customer data so valuable inside Abound that switching cost becomes prohibitive. This requires a fundamentally different customer data strategy than anything HVAC has attempted.
Competitive IntelRetention Strategy
The Debt Lever Nobody Is Modeling
At 5.66x Debt/EBITDA vs. 3.47x historical median, Carrier's balance sheet recovery depends on EBITDA growth, not just debt repayment. Software-generated revenue carries 60-70% gross margins vs. 30-35% for hardware. A 5-point mix shift toward recurring software revenue at Abound and QuantumLeap adds roughly $200M+ to EBITDA without a single additional unit of hardware sold. Analysts are not modeling this. The CDO AI role is the direct lever for executing that mix shift.
Balance SheetMargin Expansion
The 18-Month Window — First-Mover Advantage Is Not Permanent
Honeywell has acquired the Carrier Fire & Security unit and Siemens Smart Infrastructure is accelerating its Building X platform. The 18-month window during which Carrier can establish genuine AI differentiation in commercial buildings before competitors have comparable data scale is real but finite. Every month of organizational delay in deploying cross-segment models is a month of competitive runway that cannot be recovered. This is not a planning problem — it is an execution problem with a clock on it.
UrgencyCompetitive WindowExecute Now
09

Financial Impact

Every investment must demonstrate return in 18–24 months, not 5-year theoretical horizons. This digital agenda is designed with capital discipline as a Day 1 constraint, not an afterthought.

5-Year ROI Projection ($M Cumulative Impact)
$1.84B
5-Year Cumulative Value ($M)
2.2x
Return on $820M Investment
Value Drivers — 3-Year Contribution ($M)
Investment vs. Cumulative Return ($M)

Numbers to Memorize

$22.5B
Revenue FY2024 — first full year post-Viessmann integration; the real size of the pure-play
$14.2B
Viessmann acquisition size (Jan 2024; 80% cash, 20% stock) — largest bet in Carrier's history
63,000+
Buildings actively managed by Abound at April 2026 — the SaaS platform at operating scale
640M sq ft
Total surface connected to Abound — equivalent to managing an entire city's built space
$1B→$1.5B
Data center revenue: 2025 actual vs. 2026 target — most analyst-watched growth vector in portfolio
5.66x
Debt/EBITDA Q1 2026 vs. 3.47x historical median — investment discipline constraint from Day 1
650M kWh
Energy saved by Abound in 12 months ending Sept 2025 — the climate P&L in real numbers
40,000
Service dispatches avoided by Abound in 12 months — the software margin argument in one number
10

Interview Preparation

Five key messages and the ten questions the committee will ask. Click to expand each.

The Five Key Messages

Message 1
Carrier doesn't need a new CIO. It needs a Competitive Advantage Architect.
There is a fundamental difference between managing data and AI as IT support and using them as the primary engine of value creation. I'm here to do the second — and I have a specific perspective on how to do it in Carrier's unique 2026 context.
Message 2
The AI business case at Carrier is not speculative — it's already in the numbers.
40,000 dispatches avoided, 650M kWh saved, data center orders growing 4x YoY. My job is not to invent the business case — it's to build the architecture that scales those proof-of-concepts by a factor of 10 in 36 months.
Message 3
I understand the financial context completely — and I respect it as a Day 1 constraint.
With Debt/EBITDA at 5.66x vs. the 3.47x historical median, every technology investment must demonstrate return in 18–24 months, not 5-year theoretical horizons. My digital agenda is designed with capital discipline built in from the first day, not as an afterthought.
Message 4 — Contrarian
The Abound vs. OpenBlue battle isn't won in Gartner comparisons. It's won through data lock-in.
Johnson Controls is already using software as a lock-in mechanism in their bids. Carrier has the largest installed equipment base — the most valuable data source. The winning strategy: make customer data so valuable inside Abound that switching becomes prohibitively costly. This requires a customer data strategy unlike anything the HVAC industry has attempted.
Message 5 — Priority
Viessmann data integration is the biggest competitive advantage opportunity — and no one in the market is looking at it correctly.
Analysts model heat pumps + brand + European distribution. Nobody has modeled the data from millions of connected European homes. Connecting that to Abound's architecture — the work I'm committing to in the first 90 days — is how the $14.2B bet becomes genius, not just scale.

The Hard Questions

Q1: Why are you the right person for this — what CDO AI role of this scale have you run?+
The Question Behind the Question:
They're testing whether you oversell or have a credible answer. Don't claim you've run a $22.5B industrial pure-play — nobody has. Answer: "I haven't run a role at this exact scale — and I'd be suspicious of anyone who claimed they had, because this role, in this company, at this moment, doesn't have a direct precedent. What I bring is [specific: industrial IoT architecture experience / cross-segment data integration at scale / AI deployment in safety-critical physical environments]. The 90-day audit I described isn't a placeholder — it's how I de-risk the transition and move fast without moving reckless."
Q2: With 5.66x Debt/EBITDA, how do you justify $820M in AI investment over 3 years?+
The Right Frame:
"The 5.66x constraint is exactly why this investment is necessary, not despite it. Software margins at Abound run 60-70% gross vs. 30-35% hardware. A 5-point mix shift toward recurring software revenue adds ~$200M to EBITDA without a single additional hardware unit. The $820M is not a cost — it's the mechanism for reducing the debt burden faster than the bear case assumes. Year 1 investment is $220M. The payback on service dispatch avoidance alone — 40,000 dispatches at ~$800 average — is $32M per year from one use case. The committee should challenge me if the numbers don't hold. But the frame is: this is the deleveraging vehicle, not competing with it."
Q3: Bobby George left. What went wrong, and why would your approach be different?+
Answer with Precision:
"I don't have visibility into what drove Bobby's decision, and I'm not going to speculate. What I can say is that the role he held was architected as a traditional CIO role — systems, infrastructure, IT governance — with AI added as a secondary mandate. That architecture is the wrong frame for 2026. The question Carrier needs answered is not 'how do we run IT efficiently' — it's 'how does data become a competitive asset that changes our pricing power, our retention, and our P&L mix.' Those are different jobs. The role I'm proposing to perform is the second one."
Q4: How do you handle AI failures in critical environments — hospital HVAC, cold chain, data centers?+
This is a Safety and Trust Test:
"The governance framework I described — three non-negotiable axes — isn't corporate defensive bureaucracy. It's the engineering discipline that makes fast deployment possible. A model failure in a hospital isn't a software bug — it's a potential patient safety event. My answer to 'how do you handle it' starts with 'we don't get into a position where there's an autonomous AI decision in a critical system without a documented human checkpoint.' The companies that have deployed industrial AI fastest are the ones that built governance before their first incident. Carrier will be in that category. The audit trail, the human verification gate, the mandatory review cadences — those aren't slowing us down. They're what lets the board and the regulators say yes to scale."
Q5: What's your timeline to first measurable P&L impact?+
Commit to a Number:
"18 months to first verifiable margin impact. Not a demo, not a proof of concept, not a pilot — a number in the management accounts. The Abound service dispatch use case is already proven at 40,000 avoidances per year. Scaling that model across the remaining addressable building base while simultaneously adding Viessmann fault data generates measurable OpEx reduction in the first year. The 18-month commitment is: X basis points of margin improvement in the Abound segment, verifiable against Q1 2027 baseline. I'll put that on the table if the committee wants a commitment tied to performance metrics."
Q6: How do you manage the Viessmann cultural integration while also pushing AI transformation?+
Two-Speed Problem:
"The Viessmann integration and the AI transformation are not competing priorities — they're mutually reinforcing if sequenced correctly. The data audit in Month 1 is not IT due diligence — it's building the relationship with Viessmann's engineering leadership by demonstrating that I want to learn their data architecture, not override it. The AI Champions program gives Viessmann teams local ownership of the transformation rather than feeling like they're being absorbed. The German market context specifically requires visible local leadership. My approach is to have a Viessmann-side AI Champion lead visible from month 3 onward — someone from inside their engineering organization who has credibility with the team."
Q7: OpenBlue from Johnson Controls is ahead. Why can Carrier win?+
The Contrarian Answer:
"Johnson Controls has more features in production today. That's true. But they don't have Carrier's installed equipment base — which is the actual source of the most valuable training data. OpenBlue learns from building systems; Abound can learn from the equipment that creates the conditions those systems respond to. The winning strategy is not to build more features than OpenBlue. It's to make the customer's own equipment data so deeply integrated into Abound's performance models that the switching cost becomes prohibitive. A building operations director who has 3 years of AI-optimized performance data inside Abound is not switching to OpenBlue for a feature — they'd be destroying a competitive asset they own. That's the moat we're building."
Q8: How do you work with the Board and CFO on this agenda?+
Demonstrate Financial Fluency:
"The CFO and I need to be aligned before anything goes to the Board — not because I need approval for every decision, but because the Debt/EBITDA context means every investment proposal needs to be presented as a capital allocation decision with explicit payback timeline and downside protection. I would establish a quarterly CFO alignment session where we review the digital portfolio against financial KPIs: revenue attributed to AI-enhanced contracts, OpEx savings from service dispatch avoidance, software margin mix shift progress. The Board gets a semi-annual view — strategic progress plus the Organizational AI Capability Index that CHRO co-owns. I want the Board to feel that AI governance at Carrier is a board-level capability, not a CIO's pet project."
Q9: What would make you quit this role in 12 months?+
An Honest Answer Builds Trust:
"If after 12 months the organizational model doesn't give this function direct influence on how Abound's product roadmap is prioritized, or if the data integration decisions for Viessmann are being made by the business units without a unified architecture standard, I would surface that as a structural failure and have a direct conversation with David Gitlin. I'm not describing a scenario where I'd quietly exit — I'm describing a scenario where I'd escalate clearly. What I need to succeed is not headcount or budget — it's clarity that the CDO AI function has a seat at the table where product strategy decisions are made, not just a mandate to 'do AI things.' If that's not the design, I'd rather have that conversation before I start."
Q10: What's the first thing you do on Day 1?+
Be Specific, Not Generic:
"Day 1: Three conversations. First, with the Abound product team — I want to understand what data is being collected that is not being used, and what questions their customers are asking that the current platform can't answer. Second, with the Viessmann CTO in Munich — I want their read on the state of their data infrastructure and what they'd want from a unified architecture. Third, with the CFO — to establish the financial KPIs that will measure my success and the investment gates we'll use together. I'm not going to walk in with a transformation plan on Day 1. I'm going to walk in with questions that demonstrate I understand the business deeply enough to be dangerous. The plan gets written in Month 2, after I've listened."
The Closing Statement

"Five years from now, I want Carrier to be the undisputed global reference for what it means to make physical infrastructure genuinely intelligent. Not the AI reference in the HVAC sector — the global reference for how an industrial company converts measurable climate impact into sustainable business advantage and real returns for its shareholders. Abound managing 250,000 buildings worldwide. QuantumLeap as the hyperscalers' choice for the next generation of AI infrastructure. Viessmann data having created Europe's most accurate residential efficiency service. And a Debt/EBITDA that has normalized because high-margin software has changed the P&L mix. I'm here for this role because I believe with genuine conviction that Carrier has the unique pieces to accomplish all of that. My work — the work I'm here to propose — would be making sure those pieces don't stay as pieces."