ABB Ltd
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
ABB's moat is the electrification and process-automation stack already installed in grids, data centers, and plants — drives, breakers, 800xA, Ability — not the Motoman-class robot OEM it agreed in October 2025 to sell to SoftBank for about $5.4 billion.
BOTZ still holds ABB as a robotics stock. The Q2 print is an electrification compounder in the process of exiting the robot OEM:
- Electrification Is the Cash: Q2 orders hit a record $12.0 billion, up 28% comparable. Electrification orders rose 58% comparable and Motion 20% comparable, on data-center, grid, and energy-efficiency capex. Automation orders fell 14% comparable only because the year-ago quarter booked a ~$600 million large order. Continuing-operations revenue was $9.48 billion, up 12% comparable, with Operational EBITA margin at 20.2%. That is a grid-and-drives print.
- The Robot Line Is Held for Sale: In October 2025 ABB agreed to divest the Robotics division to SoftBank at about $5.4 billion. Q2 Robotics revenue was $576 million, down from $605 million, with income from operations of $47 million, and the line is presented as discontinued operations. Stranded corporate costs of $25 million in the quarter still sit in continuing ops. The robotics tape that bundles ABB with FANUC and Tesla is bundling a business ABB is selling.
- 800xA and the Installed Stack: What remains is a process and hybrid-automation installed base — 800xA DCS, Ability, Motion drives, Electrification switchgear — plus the pending Rotork actuator acquisition (~$5.5 billion), funded with the expected ~$4.8 billion net cash from the robotics sale. A plant that has validated 800xA or a data center that has specified ABB medium-voltage UPS is not shopping a humanoid. Replacing that layer is a multi-year revalidation, the same species of lock as Rockwell Logix on the discrete side of the Western floor.
Ten Moats Verdict
ABB is AI-resilient as electrification hardware and as a process-plant OS; data-center and grid capex are demand accelerants for the stack that remains. The robotics tape is the AI risk to the *multiple*, not to the business: BOTZ still holds ABB as a robot OEM while Q2 already reports Robotics as discontinued operations awaiting a SoftBank close.
71.4 resilient · 74.5 vulnerable · 80/20 = 72.0 · = 72
800xA, Ability, and ABB drive/commissioning tools are learned plant interfaces, real switching costs for process engineers, but not a single global dialect on the scale of FANUC G-code or Rockwell Studio 5000.
Drive firmware, protection relays, and 800xA control logic are vendor-owned. Customers still encode a large share of the plant's sequences themselves, so the lock is the installed stack rather than an unreplicable ontology.
ABB is an electrification and automation OEM with no public-data aggregation moat.
Power-systems, drives, and process-automation engineers are scarce, and ABB's bench is deep, but it is not a category-defining talent concentration.
Electrification, Motion, and Automation are sold as a complementary stack into the same data-center, grid, and process project — switchgear, drives, PLCs, now actuators via Rotork — so piecemeal defection degrades the engineered package.
Installed-base reliability data on drives, breakers, and 800xA systems feeds service and product revisions. Useful, not a unique un-copyable dataset.
Grid codes, marine class, hazardous-area, and process-safety certifications qualify specific ABB apparatus into a site. Replacing a medium-voltage lineup or a validated DCS is a multi-year recertification, not a vendor swap.
A global channel of ABB-certified system integrators and EPCs reinforces specification. Bilateral industrial incumbency, not a compounding user network.
Service, spare apparatus, and software subscriptions on a huge installed base of drives and switchgear recur for the life of the site. Not per-click embedding, but the aftermarket is material and contracted.
In process plants that run 800xA, the DCS is the authoritative control record for the unit. Discrete factories more often run Rockwell or Siemens; ABB's SoR is strongest in process, marine, and power.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
ABB's moat is the electrification and process-automation stack already installed in grids, data centers, and plants — drives, breakers, 800xA, Ability — not the Motoman-class robot OEM it agreed in October 2025 to sell to SoftBank for about $5.4 billion.
Growth Score
Q2 2026 continuing operations printed record orders of $12.04 billion (+28% comparable) and revenue of $9.48 billion (+12% comparable), with Operational EBITA of $1.93 billion and a 20.2% margin, up 90 basis points. Free cash flow was $881 million. Q3 comparable revenue is guided to low- to mid-teens growth; full-year 2026 is guided to low-double-digit to low-teens comparable growth with a positive book-to-bill and a year-on-year Operational EBITA margin improvement even excluding the Q1 real-estate gain. Robotics is not in that guide: Q2 discontinued Robotics revenue was $576 million, and the SoftBank close is expected in H2 2026.
Valuation Score
At $97.69 on the ADR, ABB sits about 64% of the way from the bear case ($72) to base ($112). The multiple is an electrification-and-automation compounder with a 20% Operational EBITA margin and a $30 billion backlog, not a robot OEM. Paying a robotics multiple for a company that has already signed the robot division away to SoftBank is the basket error the tape is making.
The Plant and the Grid, Not the Arm
BOTZ still holds ABB as a robotics stock. The Q2 print is an electrification compounder in the process of exiting the robot OEM:
- Electrification Is the Cash: Q2 orders hit a record $12.0 billion, up 28% comparable. Electrification orders rose 58% comparable and Motion 20% comparable, on data-center, grid, and energy-efficiency capex. Automation orders fell 14% comparable only because the year-ago quarter booked a ~$600 million large order. Continuing-operations revenue was $9.48 billion, up 12% comparable, with Operational EBITA margin at 20.2%. That is a grid-and-drives print.
- The Robot Line Is Held for Sale: In October 2025 ABB agreed to divest the Robotics division to SoftBank at about $5.4 billion. Q2 Robotics revenue was $576 million, down from $605 million, with income from operations of $47 million, and the line is presented as discontinued operations. Stranded corporate costs of $25 million in the quarter still sit in continuing ops. The robotics tape that bundles ABB with FANUC and Tesla is bundling a business ABB is selling.
- 800xA and the Installed Stack: What remains is a process and hybrid-automation installed base — 800xA DCS, Ability, Motion drives, Electrification switchgear — plus the pending Rotork actuator acquisition (~$5.5 billion), funded with the expected ~$4.8 billion net cash from the robotics sale. A plant that has validated 800xA or a data center that has specified ABB medium-voltage UPS is not shopping a humanoid. Replacing that layer is a multi-year revalidation, the same species of lock as Rockwell Logix on the discrete side of the Western floor.
Ten Moats Verdict
ABB is AI-resilient as electrification hardware and as a process-plant OS; data-center and grid capex are demand accelerants for the stack that remains. The robotics tape is the AI risk to the *multiple*, not to the business: BOTZ still holds ABB as a robot OEM while Q2 already reports Robotics as discontinued operations awaiting a SoftBank close.
71.4 resilient · 74.5 vulnerable · 80/20 = 72.0 · = 72
800xA, Ability, and ABB drive/commissioning tools are learned plant interfaces, real switching costs for process engineers, but not a single global dialect on the scale of FANUC G-code or Rockwell Studio 5000.
Drive firmware, protection relays, and 800xA control logic are vendor-owned. Customers still encode a large share of the plant's sequences themselves, so the lock is the installed stack rather than an unreplicable ontology.
ABB is an electrification and automation OEM with no public-data aggregation moat.
Power-systems, drives, and process-automation engineers are scarce, and ABB's bench is deep, but it is not a category-defining talent concentration.
Electrification, Motion, and Automation are sold as a complementary stack into the same data-center, grid, and process project — switchgear, drives, PLCs, now actuators via Rotork — so piecemeal defection degrades the engineered package.
Installed-base reliability data on drives, breakers, and 800xA systems feeds service and product revisions. Useful, not a unique un-copyable dataset.
Grid codes, marine class, hazardous-area, and process-safety certifications qualify specific ABB apparatus into a site. Replacing a medium-voltage lineup or a validated DCS is a multi-year recertification, not a vendor swap.
A global channel of ABB-certified system integrators and EPCs reinforces specification. Bilateral industrial incumbency, not a compounding user network.
Service, spare apparatus, and software subscriptions on a huge installed base of drives and switchgear recur for the life of the site. Not per-click embedding, but the aftermarket is material and contracted.
In process plants that run 800xA, the DCS is the authoritative control record for the unit. Discrete factories more often run Rockwell or Siemens; ABB's SoR is strongest in process, marine, and power.
Growth Analysis
Growth Drivers
Key Risk
Electrification order growth cools from the Q2 +58% comparable rate as data-center and grid capex digest, while Rotork integration and stranded robotics costs keep the Operational EBITA margin from expanding the way the FY guide implies.
Score Derivation
72.9 base + 4.0 trajectory + 4 margin − 5 risk = 76
Base 73 (10% midpoint of 8-12%) + 4 trajectory (all three drivers accelerating) + 4 expanding margin − 5 moderate cycle/Rotork-integration risk = 76
Price Scenarios (12–24 Months)
Where We Are vs Targets
Loading live price…
Data-center and grid order growth cools, the SoftBank robotics close slips, and Rotork is absorbed at a price that does not earn the $5.5 billion bid.
- Comparable revenue growth falls back to mid-single digits as the Q2 Electrification order spike proves a pull-forward
- The SoftBank Robotics close is delayed past H2 2026, leaving stranded costs in continuing operations and the ~$4.8 billion cash undeployed
- Rotork integration misses the promised immediate Operational EBITA accretion and the $5.5 billion cash bid looks expensive into a slower automation book
FY2026 lands in the low-teens comparable-growth guide, the robotics sale closes, and Electrification plus Motion compound on data-center and grid capex without a robotics multiple.
- Full-year comparable revenue growth prints in the low-double-digit to low-teens range with a positive book-to-bill and Operational EBITA margin up year on year
- The SoftBank Robotics divestment closes in H2 2026 and net cash proceeds of ~$4.8 billion fund Rotork plus the $2 billion buyback capacity
- Electrification and Motion hold double-digit order growth even as the +58% Q2 Electrification comparable rate normalises
Data-center electrification stays in a multi-year build, Rotork lifts the automation margin, and ABB re-rates as the grid-and-drives compounder it already is rather than as a residual robotics holding.
- Medium-voltage data-center products (HiPerGuard and related) become a disclosed growth line that sustains Electrification orders above 20% comparable
- Rotork is EPS-accretive on the guided timetable and Automation comparable orders turn positive without large-order distortion
- Free cash flow compounds off the H1 $2.13 billion run-rate and the buyback is used in size against a still-net-cash balance sheet (net debt/EBITDA 0.3 at Q2)