PORTFOLIO RESEARCH NOTE
P6 — AI & Robotics Portfolio
Studies to Date — Gate 1 Review
Bill Cara • June 2, 2026 • Status: research in progress, no positions held
Where I stand on Robotics
An AI and robotics portfolio is one of the hardest a manager can build. The theme is real and it is large, but it is crowded with hype, early-stage promises, and businesses that will not survive the next downturn. So I am taking my time. I am still inside Gate 1 of the Four-Gate Funnel — reviewing the quality of the businesses themselves — and I will not move past it until the corporate work is done.
The names you see in this preamble are a watchlist, not a buy list. Nothing has been bought. Each name has to earn its place by passing through all four gates before I commit a dollar. P6 will open with the usual cash buffer, and every trade will appear here on the day it happens — no exceptions, no back-dating.
This is where AI meets the physical world: factory robots, warehouse robots, drones, surgical robots, self-driving systems, and the software that runs all of it. It is a higher-risk theme than P5, which is the AI Infrastructure and Data Center portfolio. Until I find special opportunities in the individual names, I expect to hold the broad exposure through the sector ETFs and lean on them while the single-stock work matures.
The Four-Gate Funnel
Every name runs the same gauntlet. A stock that fails any gate does not advance, regardless of how good the story sounds, regardless of the portfolio.
Gate 1. Quality. Is the business strong and well run? This is the gate I am working through now.
Gate 2. INSTAT timing. Is my composite score positive and improving?
Gate 3. Confirmation. Do RSI and MACD agree with the timing read?
Gate 4. Chart. Does the Point & Figure chart back it up?
The industry map: nine sub-themes
I have organized the watchlist by job — by what each company actually does in the robotics value chain. Lead names only are shown below; the full list lives on billcara.com.
|
Sub-theme |
What it is |
Lead names |
|
Factory robots |
Arms and machines that build things |
ABBNY, ROK, TER, NOVT, ATS |
|
Warehouse robots |
Movers that run distribution centres |
SYM, PATH, OCDDY |
|
Vision & sensing |
The eyes — cameras and LiDAR |
CGNX, AMBA, HSAI, AEVA |
|
Drones & aerial |
Flying robots, heavy defence tilt |
AVAV, KTOS, ONDS, RCAT |
|
Humanoid & medical |
Robots that walk or do surgery |
TSLA, ISRG, PRCT, RBOT |
|
Components & actuation |
The muscles — motors and gears |
MOG.B, RRX |
|
Automation software |
The brains for office and plant work |
PATH, NOW, MSFT, CRM, IBM |
|
Self-driving & mobility |
Robotaxis and driver-assist systems |
NVDA, TSLA, GOOGL, MBLY, APTV |
|
Robotics & AI ETFs |
Baskets, for broad exposure |
BOTZ, ROBO, ARKQ, ROBT, IRBO |
Ranking the sub-themes: growth versus margins
Before looking at individual stocks, I ranked the nine sub-themes on the combination that matters most for durable returns — fastest growth paired with the best margins. My current working hierarchy, strongest first:
1. Automation software — RPA, workflow, and AI automation.
2. Self-driving & mobility — the software, chips, and driver-assist stack.
3. Humanoid & medical robotics — surgical systems and higher-end platforms.
4. Warehouse robotics — integrators with software-like economics at scale.
5. Vision & sensing — the eyes of the machine.
6. Factory robots — industrial control and factory automation.
7. Components & actuation — the motors and gears underneath.
8. Drones & aerial — defence-tilted, fast-growing off a smaller base.
9. Robotics & AI ETFs — a blend, by definition.
The logic behind the order:
• Automation and factory control is a >$200B base growing high-single to low-double digit, and the software layer carries the best margins of anything in the group.
• Robotic vision is set for low-teens annual growth into the 2030s, driven by AI and the demand for precision.
• Autonomous vehicles and robotaxis are among the highest-growth segments — 30%-plus through the 2030s — but profitability is very uneven between chipmakers, stack providers, and loss-making operators.
• Counter-drone defence is one of the fastest-growing pieces of defence tech in percentage terms, off a small base; the counter-UAS market alone is projected to more than triple toward $20B by 2030.
• The best warehouse-robotics players have shown very high revenue growth, and the leading automated-storage names reach software-like gross margins near 70% once mature.
• Surgical robotics is one of the fastest-growing healthcare-technology categories, riding minimally invasive surgery and ageing populations.
The scoring work to date
To keep this disciplined rather than impressionistic, I scored every sub-theme on five criteria using a simple 1-to-5 scale, with 5 being best. This is the structure I will turn into a reusable matrix for the full P6 universe.
The rubric
|
Criterion |
1 — Weak |
3 — Average |
5 — Excellent |
|
Industry CAGR (3–10y) |
< 5% |
10–15% |
> 25% |
|
Gross margin band |
< 30% |
50–70% |
> 80% |
|
Capital efficiency |
heavy capex (>15%) |
5–15% of sales |
asset-light (<5%) |
|
Moat quality |
commodity |
brand / integration |
IP, switching costs, regulation |
|
Breakout trigger |
vague |
probable in 2–3 yrs |
imminent / high-impact |
Sub-theme scorecard
Averaging the five criteria gives a quick fundamental read on each sub-theme. Green is strong (4.0+), amber is middling (3.0–3.9), red is weak (below 3.0). Valuation is deliberately kept out of this average and applied separately — a great business at a poor price still earns a poor expected-return rank.
|
Sub-theme |
CAGR |
Margin |
Cap. eff. |
Moat |
Trigger |
Average |
|
Automation software |
4 |
5 |
5 |
4 |
4 |
4.4 |
|
Self-driving & mobility |
5 |
4 |
3 |
4 |
5 |
4.2 |
|
Humanoid & medical |
4 |
5 |
3 |
5 |
4 |
4.2 |
|
Drones & aerial |
5 |
3 |
3 |
4 |
4 |
3.8 |
|
Warehouse robots |
4 |
4 |
2 |
3 |
4 |
3.4 |
|
Vision & sensing |
3 |
3 |
3 |
3 |
3 |
3.0 |
|
Robotics & AI ETFs |
3 |
3 |
4 |
2 |
2 |
2.8 |
|
Factory robots |
3 |
3 |
2 |
3 |
2 |
2.6 |
|
Components & actuation |
2 |
2 |
2 |
3 |
2 |
2.2 |
Note: “Cap. eff.” scores capital efficiency — the more asset-light the model, the higher the score. “Trigger” scores how near and how powerful the likely breakout catalyst is.
An early read on the names
Inside the stronger sub-themes, a handful of lead names stand out on the fundamental work so far. I stress again that this is Gate 1 thinking only — quality and structure — before any timing, confirmation, or chart work. The valuation figures in my files are placeholders to be refreshed with live data, so treat the order below as a research queue, not a ranking to act on.
• Automation software is, on the numbers, the best single combination of growth, margins, and low capital intensity in the whole theme. PATH is the highest-beta pure play on “agents do the workflow,” NOW is the quality compounder, and MSFT, CRM and IBM are the ballast that cross-sell automation inside large suites.
• Self-driving & mobility offers the highest end-market growth, but the profit pools are split unevenly. NVDA and MBLY give the cleanest exposure to the high-margin silicon and perception stack; TSLA stays a high-variance call; GOOGL’s Waymo value is buried inside a conglomerate.
• Humanoid & medical pairs a proven cash machine in ISRG with genuine asymmetry in the smaller surgical and experimental names — PRCT and RBOT — plus the long-dated optionality in TSLA’s humanoid work.
• Warehouse robotics is lumpy but explosive when mega-contracts land; SYM has shown revenue growth near 60% in stretches, and OCDDY brings software-like margins.
• Drones & aerial is the fast-growth defence sleeve; AVAV is the most established name, with the smaller plays carrying binary, event-driven risk.
Read against my own framework, the working shortlist of names worth the deepest Gate 1 work is PATH, MBLY, ISRG with PRCT/RBOT as the asymmetric pairing, SYM, and AVAV — with the sector ETFs held as a risk-managed placeholder while conviction is built. None of this is a commitment. It is the map of where the corporate due diligence goes next.
What comes next
The immediate task is to finish Gate 1 properly: complete the corporate review on the shortlist, then formalize the scoring matrix into a stock-level sheet that feeds my normal workflow. The columns I want on every name are industry growth, gross-margin band, capital intensity, moat quality, and the specific breakout trigger — with a valuation overlay layered on top to convert “high growth plus high margins” into a genuine expected-return rank.
Only once a name clears Gate 1 does it earn the right to be tested at Gate 2 against its INSTAT score, and so on through confirmation and the chart. Until then, P6 stays in cash but for its buffer. I will not rush any portfolio. The discipline is the edge.
— Bill Cara
billcara.com