The real cost of machine vision: what the camera quote leaves out
When a machine vision project is scoped, the number that gets written into the business case is almost always the camera price. In 2026 a Basler ace 2 Pro lists around €1,629 on Basler's own shop. A useful lens adds another €260. Three weeks later, when the quote from the system integrator lands, the total has tripled. That gap between the headline hardware number and the real cost is where most machine vision business cases break.
This post walks through the full cost stack of a classic fixed-installation machine vision station and a smartphone-based alternative. It uses the same components and sources we published in our iPhone vs. industrial cameras benchmark so the numbers are verifiable end-to-end.
The five cost buckets no camera quote mentions
A production vision station needs more than a camera. The five additional buckets are: lens and mount, lighting, industrial PC or GPU compute, enclosure plus cabling, and integration labour. Each one carries a range. Collectively they account for 70 to 85% of first-year total cost on a typical industrial camera station.
1. Lens and mount
A C-mount machine vision lens from Basler, Edmund Optics or Kowa lands between €250 and €900 depending on working distance and resolution. A precise mount with alignment adjusters adds another €150 to €400. For high-end applications requiring telecentric lenses (flatness checks, dimensional gauging), a single lens can exceed €2,500.
2. Industrial lighting
This is the bucket most scoping exercises get wrong. A reliable ring light or bar light with a driver, a diffuser and a light-tight enclosure runs €400 to €1,500 per station. Dome lights for reflective surfaces push €1,500 to €3,000. A poorly lit station is the number-one cause of false rejects in the field, which is why we wrote a whole lighting guide for AI visual inspection.
3. Compute
An industrial PC capable of running a deep-learning model in real time is not a cheap accessory. A fanless IPC with an NVIDIA RTX A2000 or a Jetson AGX Orin module sits between €2,000 and €4,500. For multi-camera setups or high-throughput lines, the number moves further up. Compute is the single line item that often surprises finance leaders.
4. Enclosure, cabling and cabinet
Industrial enclosures to IP54 or IP67, PoE switches for GigE Vision cameras, shielded cables in the right length, mounting brackets and a small control cabinet all add up. A reasonable allowance is €500 to €1,200 per station.
5. Integration labour
The least visible but largest cost bucket. A system integrator typically quotes between 40 and 120 hours for a first-of-kind installation, at €120 to €180 per hour. That is €4,800 to €21,600 per station for a classic vision project. For a line with five stations, system integration alone can exceed €50,000 before any hardware is shipped.
The classic bundle, totalled
A realistic conservative bundle for a single Basler ace 2 Pro station with mid-range lens, ring lighting, Jetson AGX Orin compute, enclosure, cabling and integration labour lands around €10,500 to €13,500 in Year 1. A premium bundle with telecentric lens, dome lighting, fanless IPC with RTX A2000 and extended integration lands between €18,000 and €26,000. Neither number includes the classic machine-vision software licence (from €2,500 upwards per seat for Cognex VisionPro or MVTec MERLIC).
The smartphone alternative, totalled
An iPhone-based inspection station built on an Enao Vision subscription looks different. A refurbished iPhone suitable for on-device inference lands at €400 to €600. A mount, cable, bracket and ring light together cost €200 to €300. The Enao subscription replaces the IPC, GPU, vision software licence and most of the integration labour. Hardware total: under €1,000 per station. The subscription is the OpEx line item and is sized per station.
The gap is not marginal. On a single station, the smartphone bundle lands at 5 to 15% of the classic bundle's Year-1 cost. The gap widens on multi-station lines because lighting and compute savings compound. We walk through the shift from upfront hardware to a subscription model in our post on the move from CapEx to OpEx in manufacturing.
Where the classic bundle still wins
The classic industrial-camera bundle is still the right choice for three workloads. First, high-speed linear scanning (film, web, strand) where line-scan cameras and specialized lighting are needed. Second, dimensional gauging with telecentric lenses to sub-millimetre accuracy. Third, extreme environments (high-vibration, high-temperature, flammable atmospheres) where no smartphone will survive. For those, the full classic stack is warranted and its cost is justified. For the rest, which is most of the real-world inspection problems, the smartphone bundle lands at a fraction of the cost.
The fully-loaded cost per inspected part
A useful way to reason about this: take the Year-1 bundle cost, divide by inspected parts per year, and add any subscription. On a typical two-shift line at 60 parts per minute, Year-1 output is around 25 million parts. A classic €12,000 bundle lands at roughly 0.05 euro cents per inspection. A smartphone bundle at €1,000 hardware plus an Enao subscription still lands lower per inspection, often under 0.02 euro cents. Scale that across 20 inspection stations on a large line and the difference is material.
What to do with this information
Before you sign a machine vision quote, ask for the full bill of materials including lens, lighting, compute, cabling, enclosure and integration. A quote that only mentions the camera is incomplete. Ask whether the integrator's hours are fixed-price or time-and-materials. Ask how a changeover between SKUs affects the model and whether additional training is a separate billable item. Those four questions move 80% of the cost onto the table where it belongs.
If you want to see a smartphone-based inspection station costed end-to-end against your current machine-vision project, join the Enao community and we will share the reference bill of materials and a side-by-side template you can drop into your own business case.