Artifact · v0 (real content)

Hill Scorecard

Four candidate hills scored against six criteria. Which market is worth climbing first, and why Physical AI is not a strategic pivot but a naming exercise for work already in progress.

1.4 · Market and Sizing · artifact id: hill-scorecard-v0.html · 2026-05-28 · v0 draft

The four hills

These are not hypothetical markets. Each one appeared in the May 15 discovery call, and the B1/B5/B3 research streams verify the market conditions behind each. The scoring below reflects current conditions as of May 2026.

Framing note. This scorecard does not declare a winner. It shows which hills are accessible now versus which require prerequisites you have not yet met. Read it alongside the alignment memo before making any hill selection.

Scoring criteria

Criterion What it measures Weight
TAM 2026Total addressable market, sourced from cited research20%
Time to first revenueMonths from activation decision to first signed deal20%
Capability fitHow well SmartOne's existing stack matches what buyers need20%
Competitive moatHow defensible the position is against the comparable set15%
Capex / activation costWhat it takes to open the door (low = good)15%
Founder fitHow well the hill matches Habib and Shahysta's stated priorities10%

Scorecard: four hills

Hill 1: Physical AI and World Model Data Services

Phase 1
CriterionFindingScore (1-5)
TAM 2026 Broad Physical AI market: $81-108B (Kaisore Research). Data annotation services within it: $5.1B (Verified Market Research). AV training data sub-segment alone: $1.3B (Intel Market Research). The ratio claim holds: broad Physical AI is 16-21x the annotation services market. 5
Time to first revenue Already generating revenue here (Amazon robotics pilot underway, autonomous truck and drone work done). This is a repackaging exercise, not a new market entry. First new branded deal: 3-6 months. 5
Capability fit Temporal annotation, trajectory prediction, 3D spatial reasoning, LiDAR segmentation all confirmed in SmartOne's production portfolio. SOC 2 Type II and ISO 27001 pass the buyer gate that eliminates most emerging-market competitors (Data Science Society 2026 buyer survey). Gap: no packaged Physical AI pipeline offer; the work exists but is unnamed. 4
Competitive moat Three simultaneous differentiators no competitor in the comparable set holds: bootstrapped independence, full-time institutional workforce (not gig), and Physical AI depth already in production. The Scale AI conflict window is open. Durable moat is the workforce model; the conflict window is time-bounded. 4
Capex / activation cost Low. No new legal entity required. No new certifications required. The activation cost is positioning and packaging: a product brief, named service lines, and a case study set. Estimated 4-8 weeks of founder time plus HARBOR engagement scope. 5
Founder fit On the call you said "the signals are converging towards physical AI, robotics." The concern about workforce protection is real; Physical AI at scale preserves the Madagascar model. Strong fit. 5
COMPOSITE SCORE: 4.7 / 5.0  ·  Recommended primary hill

Hill 2: Commercial Account Expansion

Phase 1 / near-term
CriterionFindingScore (1-5)
TAM 2026 Not a market; it is a revenue mechanics play. The addressable base is SmartOne's existing warm pipeline and the accounts adjacent to current clients. Value: converting 40-60% of monthly spend to annual commitments changes SmartOne's financial posture without requiring new market entry. 3
Time to first revenue Fastest path: the robotics pilot is live. Converting it to a 12-month commitment requires one conversation, not a new sales cycle. 30-90 days to first annual commitment if the pricing model is in place. 5
Capability fit Perfect fit. The work is already being done. The gap is the commitment structure, not the capability. 5
Competitive moat Low moat by itself. Annual commitment structures are standard. The moat comes from delivering Physical AI work well enough that the annual commitment is not a risk the client resists. 2
Capex / activation cost Very low. Requires a pricing model spec and a commitment template. No new hires, no new certifications, no new legal structure. 5
Founder fit Directly addresses Shahysta's mandate: secure revenue as fast as possible. Reduces the survival-mode pressure that constrains every other decision. Essential enabler rather than a long-term hill. 5
COMPOSITE SCORE: 4.2 / 5.0  ·  Recommended as parallel immediate action

Hill 3: Europe Growth

Phase 2 / conditional
CriterionFindingScore (1-5)
TAM 2026 French enterprise AI market: $633.6M in 2024, growing to $3.89B by 2030 at 36.3% CAGR (Grand View Research). French data annotation addressable market estimated $300-500M by 2026 (working estimate; European data services market is approximately $2B at European GDP weighting). DORA enforcement and EU AI Act full application (August 2026) create compliance-driven procurement demand. 4
Time to first revenue B5 research timeline was 9-15 months from a cold activation. SmartOne already has a team in France covering Europe and SMARTONE EU SAS in Paris. That existing footprint reduces the ramp. Realistic range from focused activation to first signed framework agreement: 6-12 months. First paid POC: months 4-7. Conditional on survival-mode pressure easing in months 1-4. 3
Capability fit SmartOne is already operating in France. A small business and healthcare team covers Europe today. SMARTONE EU SAS is the legal vehicle behind them. Habib has been president of the entity since November 2021. Safran is a current client. Palantir DGSI credential (Palantir renewed multi-year DGSI contract December 2025) is a named reference. ESSEC alumni network is a structural warm path. SOC 2 and ISO 27001 meet EU AI Act documentation requirements. Capability fit is strong; structured sales motion capacity is the constraint. 5
Competitive moat Strongest geographic moat in the set. No comparable (Scale AI, Surge, iMerit, Toloka) has a French entity, French-speaking leadership, and active French defense client history simultaneously. EU AI Act compliance demands create a lasting procurement advantage for certified vendors with European entities. 5
Capex / activation cost Moderate. Entity and team exist; no new incorporation required. Growth-stage actions: French-language materials (2-3 weeks), Safran framework conversation, ESSEC alumni outreach. First sales motions are cost-of-time, not cash-heavy. Estimated activation cost: $10-20K in translation work and sales materials plus 2-3 months of founder attention to structure the motion deliberately. 4
Founder fit France was explicitly identified on the call as viable based on the Mistral-driven ICP convergence with the US 2022-2023 market. Both founders carry the ESSEC network and the language. Strong fit, but contingent on Phase 1 stability first. 4
COMPOSITE SCORE: 4.2 / 5.0  ·  Recommended for Phase 2 activation (months 4-9). Existing footprint reduces activation risk versus a cold-country entry.

Hill 4: Defense and NATO Data Services

Phase 3 / conditional on firewall
CriterionFindingScore (1-5)
TAM 2026 Pentagon CDAO data labeling budget: $582M FY25 request. Scale AI CDAO OTA ceiling: $500M (Bloomberg, May 2026). UK MoD Asgard AI framework: 4-year program, 26 firms. France DGA Airbus framework: 50M euro (December 2025). Canada DIA AI investment: CAD 13.8M (March 2026). Aggregate NATO-ally data AI spend: significant but diffuse and partner-specific. 5
Time to first revenue B3 research realistic timeline: 3-5 years minimum from decision to first cleared contract delivery. Steps include legal structure design (3-6 months), US subsidiary formation (1-3 months), FOCI mitigation negotiation with DCSA for complex multi-jurisdictional ownership (12-24 months), Facility Security Clearance (6-12 months), key personnel clearances at Secret level (6-12 months per person). This is not Phase 1 or Phase 2 work. 1
Capability fit Capability for non-classified defense-adjacent work is strong (sensor fusion, temporal annotation, Palantir-adjacent history). Capability for classified work requires a US cleared subsidiary with a separate cleared annotation workforce. Madagascar cannot be used for ITAR-controlled data (deemed export rule). A separate, parallel cleared US workforce must be built from scratch. 2
Competitive moat Long-term moat is real: once FOCI mitigation is complete, a cleared Canadian-headquartered Physical AI data firm occupies a nearly unique position. Near-term moat is zero because Scale AI already owns the cleared data annotation category. BAE Systems Inc. (UK parent, US cleared subsidiary under SSA) is the structural precedent. 3
Capex / activation cost Very high. Estimated costs: US subsidiary setup $150-350K first-year infrastructure; FOCI mitigation legal fees $200-500K for complex multi-jurisdictional structure; FCL and clearance timelines 18-36 months for TS-cleared personnel; cleared annotation workforce build 2-4 years. Total activation: $750K-$2M before first cleared contract delivery. Your own framing on the call: "back pocket, not pertinent for now." 1
Founder fit Explicit on the May 15 call: this is not the focus of work ahead. The defense path is filed for later consideration. The score here is honest: it is low because the founders have not chosen this hill for now. 1
COMPOSITE SCORE: 2.2 / 5.0  ·  Dormant. Revisit when alignment is settled and revenue is stable.

Summary

Hill 1
Physical AI
Score: 4.7
Hill 2
Acct Expansion
Score: 4.2
Hill 3
Europe Growth
Score: 4.2
Hill 4
Defense
Score: 2.2
Now
Hills 1+2 in parallel
The Physical AI hill is already the hill you are on. The robotics pilot, the autonomous truck work, the drone annotation, the sensor-fusion client history: these are Physical AI data services, not data annotation with Physical AI aspirations. The question is not whether to enter this market. You are in it. The question is whether to name it, package it, and price it to match the market it belongs in.
Hill 4 (defense) carries a hidden cost that the TAM does not reflect. SmartOne's cost advantage (Madagascar workforce) is structurally unavailable for ITAR-controlled classified work. The company would need to build a separate, more expensive US cleared workforce. Load that cost fully before treating the defense option as a revenue opportunity rather than a long-term call option.

Key sources