The robotics industry has over 3,000 companies globally, and roughly 15% of robotics startups fail within their first five years. When your operation depends on a vendor's continued existence, their financial health matters as much as their robot's specifications. Yet most buyers spend 90% of their evaluation time on hardware and 10% on the vendor behind it.
This framework flips that ratio where it matters.
Financial Stability Assessment
A robot is a 5-10 year commitment. Your vendor needs to survive at least that long.
For public companies (FANUC, ABB, Intuitive Surgical, Stryker): Review quarterly earnings, revenue growth, and robotics segment margin. These vendors aren't going anywhere — the risk is product line discontinuation, not company failure. Check annual reports for R&D spending as a percentage of revenue (healthy: 8-15%).
For venture-backed startups: Request their most recent funding round details. Key metrics: total capital raised, months of runway at current burn rate, revenue trajectory, and path to profitability. A company that raised $50M in Series C but burns $8M/month has six months of runway — that's a risk. Ask directly: "At current burn rate, how many months of runway do you have?" Evasive answers are themselves an answer.
Warning signs:
- Leadership departures (CTO, VP Engineering) within the last 12 months
- Layoffs exceeding 15% of headcount
- Pivot from hardware to "platform" or "data" — often signals hardware economics aren't working
- Acquisition rumors — acquisitions can mean product discontinuation
- Unusual discounting or aggressive RaaS terms — may signal desperation for revenue
Mitigation strategies: Negotiate source code escrow for critical software. Ensure data export rights in your contract. For critical deployments, require a maintenance bond or establish a service agreement with a third-party integrator who can support the hardware independently.
Support Network Evaluation
The support experience after purchase matters more than the sales experience before it.
Geographic coverage: Where are the vendor's service technicians located relative to your facilities? For critical operations, target 4-hour on-site response time. Map service center locations and ask for guaranteed response times by geography — not just averages.
Support tiers: Understand what's included in standard support vs. premium support. Key distinctions:
- Phone/remote support hours (24/7 vs. business hours)
- On-site response time (4 hours vs. next business day)
- Parts availability (depot stock vs. ship-from-factory)
- Software update frequency and testing requirements
- Dedicated account manager vs. general support queue
Escalation paths: How does a critical failure get handled? Ask the vendor to walk you through their severity classification (Sev 1/2/3), escalation procedures, and executive escalation process. If they can't articulate this clearly, their support organization isn't mature enough for enterprise deployments.
Support metrics to request:
- Average resolution time for critical issues (target: under 8 hours)
- First-call resolution rate (target: over 60%)
- Customer satisfaction score (target: over 85%)
- Average time to dispatch on-site technician
- Spare parts fill rate (percentage of parts available for immediate shipment)
Reference Check Framework
Vendor-provided references are curated. You need independent verification.
Request a complete customer list — not cherry-picked references. If the vendor refuses, ask for all customers in your industry vertical and geographic region. Some will refuse this too (citing customer privacy), but push for at least 5-10 references beyond the 2-3 they volunteer.
Reference check questions (ask the customer, not the vendor):
Deployment experience:
- How long from purchase order to fully operational? Was it within the quoted timeline?
- What surprised you during deployment that the vendor didn't warn you about?
- If you could do it over, what would you do differently?
Ongoing support:
- How often does the robot go down, and how long does recovery take?
- Rate the vendor's support responsiveness on a 1-10 scale.
- Have you ever escalated a support issue to executive level? What happened?
Financial:
- What was the total cost of ownership versus what the vendor initially quoted?
- Were there unexpected costs? What were they?
- Would you buy from this vendor again?
The killer question:
- "If this vendor went out of business tomorrow, how would you cope?"
Their answer reveals how dependent they are — and how exposed you'd be.
Pilot Program Negotiation
A pilot tests the robot in your actual environment, not a demo room. Negotiate it properly.
Pilot structure:
- Duration: 60-90 days minimum (30 days is too short to capture edge cases)
- Scope: test the primary use case plus at least one secondary use case
- Success criteria: define specific, measurable targets before the pilot starts (throughput, uptime, error rate, integration milestones)
- Staffing: the vendor should provide an on-site applications engineer for at least the first 2 weeks
Pilot pricing models:
- Paid pilot (best): you pay 10-25% of purchase price; this amount applies toward the purchase. Paid pilots get better vendor attention.
- Free pilot (caution): the vendor is desperate or confident. Either way, define who pays for shipping, installation, connectivity, and support during the pilot. "Free" pilots often have $10K-$20K in hidden costs.
- Rental (practical): monthly rental at market rates, with purchase credit. Cleanest structure for both parties.
Data from the pilot: Insist on access to all performance data — uptime logs, error logs, throughput metrics, connectivity events. This data is your negotiating leverage for the purchase contract and your baseline for post-deployment SLA enforcement.
Exit criteria: Define clear conditions under which the pilot fails and you walk away with no further obligation. If the robot fails to meet 80% of primary success criteria, the pilot is a failure regardless of what the sales team says.
Contract Terms That Protect You
Robot purchase contracts favor vendors by default. Here's what to negotiate.
Pricing protection: Lock in per-unit pricing for fleet expansion within 24 months. Volume discount tiers should be explicit. If you're buying RaaS, cap annual price increases (CPI-linked, max 5%).
Uptime SLAs: Define uptime as a percentage of scheduled operating hours (not calendar hours). Target 95% for AMRs, 98% for surgical robots. Include financial penalties for SLA misses — typically service credits equal to 2-3x the prorated downtime cost.
Software terms: Clarify what happens if the vendor changes their software model. Can they move from perpetual license to subscription mid-contract? Are updates mandatory or optional? Who bears the cost of testing updates against your integration?
Data rights: You own your operational data. The contract should state this explicitly. The vendor may retain rights to anonymized, aggregated data for product improvement — that's reasonable. But your specific operational data (routes, throughput, inventory, personnel) is yours.
Termination and exit: What happens if you need to switch vendors? Can you operate the robots without the vendor's cloud services? What's the contract buyout formula? Is hardware transferable to a new owner?
Warranty: Standard robot warranties are 12-24 months. Negotiate for 24 months. Clarify warranty scope — does it cover parts and labor, or just parts? What voids the warranty (third-party modifications, uncertified parts, environmental conditions)?
Exit Strategy Planning
Every vendor relationship should have an exit plan. Not because you expect to leave, but because planning for it reveals dependencies.
Vendor lock-in risks:
- Proprietary communication protocols that prevent multi-vendor fleets
- Cloud-dependent operations that stop if the vendor's servers go down
- Custom integrations that only the vendor can maintain
- Proprietary end-effectors or accessories with no third-party alternatives
Mitigation:
- Prefer vendors with open APIs and standard protocols (ROS 2, OPC UA, REST APIs)
- Ensure local operation capability without cloud connectivity
- Document all integrations and ensure your team (or a third-party integrator) can maintain them
- Stock critical spare parts independently
Use our Robot Finder to compare vendors across these dimensions, or talk to our AI Advisor for vendor recommendations tailored to your requirements.
Frequently Asked Questions
How do I check if a robot startup is financially healthy?
For private companies, direct inquiry is your best tool. Ask about total funding raised, current burn rate, revenue growth rate, and customer count trajectory. Cross-reference with Crunchbase or PitchBook data. Check Glassdoor for employee reviews mentioning layoffs or financial concerns. Ask for bank references or a D&B report if the purchase exceeds $500K.
What should a robot vendor's support SLA include?
At minimum: defined severity levels with response time guarantees (Sev 1: 2-4 hours, Sev 2: 8 hours, Sev 3: next business day), on-site response time for hardware failures, spare parts availability guarantee, software update schedule, and financial penalties for SLA misses. Premium SLAs add 24/7 phone support, dedicated account management, and preventive maintenance visits.
How many vendor references should I check?
Check at least 5 references, and ensure they include: one customer in your industry, one at similar scale, one who has been a customer for 2+ years, and at least one reference you sourced independently (not provided by the vendor). If the vendor's total customer count is under 20, check at least 30% of their install base.
Should I choose a large corporation or a startup for my robot vendor?
Large corporations (FANUC, ABB, Siemens) offer stability, global support networks, and proven reliability — but they move slowly, charge premiums, and may prioritize larger customers. Startups offer innovation, flexibility, and competitive pricing — but carry financial risk, have smaller support networks, and may pivot or fail. For mission-critical operations, lean toward established vendors. For competitive advantage and latest technology, startups can be worth the risk with proper mitigation.
What's the most common vendor selection mistake?
Selecting based on demo performance rather than real-world deployment evidence. Demos are controlled environments — your facility is not. The most reliable predictor of deployment success is reference checks with customers running similar applications at similar scale. Second most common: ignoring total cost of ownership in favor of unit price.