-
Who this checklist is for (and who it isn't)
-
Step 1: List the materials you'll cut in month one, not year three
-
Step 2: Measure your exhaust path before you measure the machine
-
Step 3: Treat "works with" claims as claims, not specs
-
Step 4: Demand a sample project, not a demo
-
Step 5: Reality-check the project library against your actual catalog
-
Step 6: Read what happens in month 13
-
Step 7: Admit what the machine can't do (the step most people skip)
-
Notes and common errors
If you're a small shop owner, a maker, or an educator about to drop $3,000-$6,000 on a desktop laser cutter, this checklist is for you. Seven steps, in order. Each one maps to a mistake I've either made myself or watched someone make since 2017.
I've burned roughly $4,200 on wrong turns in that time — wrong machine spec, wrong material assumption, wrong warranty read. This is the checklist I now keep taped inside my shop binder. It works whether you end up with a Glowforge or something else.
Who this checklist is for (and who it isn't)
Use this if you're cutting or engraving wood, acrylic, leather, and glass on a desktop-scale CO2 laser and you expect to run 20-200 projects a month. Don't use it if you're shopping for an industrial laser welding machine for sale — that's a different machine class with a different safety profile, and I'll come back to that in Step 7.
Seven steps. Fifteen minutes of reading. Maybe two weeks to actually execute.
Step 1: List the materials you'll cut in month one, not year three
Open a spreadsheet. Column A: material. Column B: thickness in millimeters. Column C: how many pieces per week. That's it.
The mistake I made in 2017 was listing "wood, acrylic, leather, glass, maybe metal" because those were the categories I was excited about. The machine I bought could engrave all of them. But my first 30 orders were 3mm birch plywood and 1/8" acrylic — the two materials I didn't give thickness numbers for. I ended up buying a second set of lenses and learning, the hard way, that "wood" isn't a spec.
Checkpoint: If you can't fill column B for at least four materials, you're not ready to buy. Go measure samples.
Step 2: Measure your exhaust path before you measure the machine
Almost everyone does this backwards. They pick the machine, then figure out where the hose goes.
Here's the rule I now follow: if the distance from your machine's exhaust port to an outside wall exceeds about 8 feet, or if the run includes more than two 90-degree turns, budget for a booster fan and a second filter. On a 40W desktop unit, a long or bendy run cuts effective throughput noticeably. I learned this on a $3,800 order back in 2019 where the customer's shop had a 14-foot run with three elbows. We lost the job.
Checkpoint: Take a photo of your intended exhaust path. Count the elbows. Measure the run. If it's over 8 feet, add $200-$400 to your budget for airflow.
Step 3: Treat "works with" claims as claims, not specs
Every manufacturer has a materials page. Read it like a lawyer.
From the outside, a "compatible materials" list looks like a green light. The reality is that most lists describe materials the machine can process under ideal conditions — not materials it will process well on your setup with your ventilation and your lens.
As of January 2025, the FTC still requires advertising claims to be truthful, substantiated, and not misleading (Source: FTC Business Guidance on Advertising, ftc.gov). A vendor saying "works with leather" is not the same as saying "works with chrome-tanned leather at 4mm." Ask which specific leather. Ask about PVC — and if they say "yes, all vinyl," walk away, because CO2 lasers and chlorinated plastics are a bad combination.
Checkpoint: For your top three materials from Step 1, ask the vendor for a specific test result or a sample. If they can't produce one, drop them one slot in your shortlist.
Step 4: Demand a sample project, not a demo
A demo is designed to look good. A sample project is designed to test the machine.
Send your file. Not their file. Your file, with your material, your thickness, and the detail level you actually sell. I once watched a sales rep run a beautiful 12-minute engraving on 3mm bamboo — a material I didn't use — then hand me a brochure. Beautiful machine, wrong test.
If you're comparing Glowforge laser cutters against something else on your shortlist, send both vendors the same file on the same day. Compare the edge quality under a loupe, not on a phone photo.
Checkpoint: One file, three vendors, one week. Score edge quality, kerf consistency, and engraving depth on a 1-5 scale.
Step 5: Reality-check the project library against your actual catalog
Community libraries are genuinely useful — this is one of the places Glowforge has a real advantage, with an active catalog and a lot of shared files. But "large library" doesn't mean "library that helps you."
Filter the library to your top three product categories from Step 1. How many files are close enough to adapt? If it's fewer than about fifteen, the library is a bonus, not a feature — don't let it carry the buying decision.
Checkpoint: Spend 30 minutes in the library. Save five files you'd actually produce. If you can't find five, discount the library from your decision.
Step 6: Read what happens in month 13
Warranty boundaries are where purchase decisions quietly reverse. I approved a machine in 2022 partly on price and didn't read the service terms until month 14. When the tube started fading, I learned the replacement timeline was six weeks and the shipping was on me. Roughly $600 out of pocket, and a month of missed orders.
Even after I signed off, I kept second-guessing. What if the price saving wasn't worth it? I didn't relax until month three, when the first tube swap went smoothly. By month 14, I knew I should have read the document first.
Questions to ask, in writing:
- What's covered in year one, and what changes in year two?
- What's the replacement lead time for the laser tube and the lens?
- Is on-site service available, or is it ship-back?
- If you're outside the US, who's the local service partner?
Checkpoint: Get the warranty document in writing before you pay. Not the sales page — the actual document.
Step 7: Admit what the machine can't do (the step most people skip)
This is the one I want you to take seriously, because it saves the most money.
People think the expensive mistake is buying too little machine. Actually, the expensive mistake is buying a machine outside its lane and pretending otherwise. A desktop CO2 cutter is not a welder. It is not a fiber marker. It is not a metal fabrication tool. If your roadmap includes "and eventually we'll weld aluminum frames," buy the cutter for cutting and budget for a welder separately — or find a shop that does the welding better than you ever will.
Here's the boundary I now draw for myself: if a job requires an industrial laser welding machine, I subcontract it. I'm not a welding shop. I'm a cutting and engraving shop. Telling a customer "this isn't our strength — here's who does it better" has cost me exactly zero orders and earned me repeat work on everything else. A vendor who says "we do everything" has usually never measured how well they do anything.
On terminology: European spec sheets sometimes list a fraktionierter CO2 laser as if it's a distinct category. It isn't — it's marketing for a fractional-power CO2 source. Same family, different label. And if you search for something like a "CO2 laser Savannah" listing, treat the place name as a seller location, not a technology. The spec sheet is what matters, not the title.
Checkpoint: Write down one sentence describing what you will not do with this machine. Keep it taped to the machine.
Notes and common errors
A few things I'll flag because I've watched them go wrong repeatedly:
The safest-looking claim is often the least tested. Be skeptical of any vendor promising 100% safe unsupervised use. No desktop laser is that. Ventilation, interlocks, and material discipline are on you.
The first project always takes longer than the demo. If your business case requires producing sellable product in week one, add a week. If it requires producing in week two, add two. The learning curve on speed and power settings is real, and it's not in the brochure.
Community advice is useful but not uniform. You'll find people running materials through machines that technically shouldn't run them. "It worked for me" is not a safety standard, and it's not a warranty standard either.
Prices move. Budget ranges in this article reflect early-2025 conditions; verify current pricing and warranty terms before you sign anything.
Okay — those are the seven. If you run them in order, you'll probably avoid the $4,200 I didn't.