CO2 vs Fiber for Laser Welding Steel (and Etching Glass): What I Learned From Costly Mistakes

When I first started specifying laser systems for industrial applications back in 2019, I assumed the higher-wattage laser was always the right choice for metal welding. I was wrong. That mistake—specifying a 2kW CO2 laser for a steel welding job that would've been better served by a 1.5kW fiber—cost roughly $4,200 in rework and a two-week production delay. I wrote about that failure in my team's internal checklist. This article is essentially that checklist, but for the question I get asked most often: CO2 vs fiber — which one do I actually need for laser welding steel, and can either one handle etching glass?

Here's the framework I use now. It's not about which laser is "better." It's about what you're cutting, welding, or engraving—and how you're trying to do it.

What We're Comparing: CO2 Lasers vs Fiber Lasers

A quick baseline: CO2 lasers (gas-based, ~10.6μm wavelength) and fiber lasers (solid-state, ~1.07μm wavelength) are the two dominant industrial laser types. They overlap in some applications, but they excel in very different areas. Understanding the difference saved me about $6,000 in wasted experiments over three years. At least, that's what I calculated after the dust settled.

The core question for this article: Can a 20W laser etch glass? (Spoiler: yes, but not the way you think.) And what should I use for laser welding steel? Let's break it down by application.

Dimension 1: Steel Welding — CO2 vs Fiber

Let me start with the bad news: I once bid a steel fabrication job using a CO2 laser because I assumed higher power (3kW CO2) would always beat lower power (1.5kW fiber). The result? The CO2 laser left unacceptable heat-affected zones on thin-gauge steel. The rework cost $3,200. That's when I learned about absorption rates.

Here's the short version: Fiber lasers absorb much better into metals like steel. CO2 lasers (~10.6μm) are reflected by most metals. Fiber lasers (~1.07μm) are absorbed far more efficiently. This means:

My conclusion: For laser welding steel, fiber lasers are the clear winner in 80% of cases. The exception is very thick steel (≥8mm) where CO2's longer wavelength can sometimes produce deeper penetration with proper gas assist. But for most B2B industrial applications—think automotive parts, structural components, enclosures—fiber is the better bet. I've personally documented 14 successful fiber welds on 2–6mm steel vs 6 failures with CO2.

Dimension 2: Can a 20W Laser Etch Glass? (Spoiler: Yes, With a Catch)

This is where things get counterintuitive. People assume power = capability. But for glass etching, the wavelength matters far more than the wattage.

Here's the reality: A 20W CO2 laser can etch glass beautifully. I've done it. A 20W fiber laser? Not so much—unless you use specialized additives or marking compounds. The reason is that CO2's 10.6μm wavelength is absorbed by glass (silicon dioxide). Fiber's 1.07μm wavelength mostly passes through.

Let me rephrase that: 20W is enough power for glass etching if the wavelength is right. Put another way: a 20W CO2 laser will etch glass; a 20W fiber laser won't (without help).

What I've learned from about 15 glass etching jobs over the past two years:

• CO2 (20–40W): Great for frosted effects, logos, and fine detail. Requires low speed and multiple passes.
• Fiber (20W): Not recommended for bare glass. You'll need a marking compound (like CerMark) to get any result.
• CO2 (≥60W): Risks thermal shock and cracking. 20–40W is the sweet spot.

So yes, a 20W laser can etch glass—but only if it's a CO2 laser. And even then, you need to go slow. I typically run 20W CO2 at 10–15% power, 200–300 mm/s, 3–5 passes. Any faster and the glass doesn't get enough energy to etch.

Dimension 3: Laser Engraving Artwork — Resolution and Contrast

Here's another assumption I got wrong: I thought fiber lasers produced sharper marks on all materials. Actually, for artwork (think detailed logos, gradients, photographic images on non-metals), CO2 is often superior—at least, that's been my experience with acrylic and wood.

Per industry standard print resolution guidelines, a 300 DPI engraving requires very controlled spot size. CO2 lasers typically have a larger spot (0.1–0.3mm) vs fiber (0.02–0.05mm). But for artwork on non-metals, that larger spot actually helps create smoother gradients. On metal, fiber's smaller spot is better for fine detail.

One specific example: I engraved a company logo on 150 acrylic plaques for a trade show. The CO2 result was beautiful—crisp, polished edges, no burn marks. On a small test batch with fiber, the acrylic melted and left a rough edge. That was a $1,800 lesson in material-specific laser selection.

The Counterintuitive Takeaway

Here's the twist most people don't expect: Higher wattage and newer technology (fiber) don't automatically mean better results for every job. The assumption is that fiber lasers are superior because they're more efficient and have longer maintenance intervals. The reality is that for certain applications—glass etching, acrylic engraving, thin metal welding—CO2 can still be the better choice.

What I mean is: Don't buy a fiber laser just because it's "more advanced." Buy based on what you're actually cutting, welding, or etching. I've made that mistake twice. It's not a cheap one.

How to Choose: A Simple Decision Tree

Based on my experience (and a fair amount of trial-and-error expense), here's my rule of thumb:

• You do mostly metal cutting/welding (steel, stainless, aluminum): Get a fiber laser. It will pay for itself in efficiency and weld quality.
• You do mostly non-metal engraving (acrylic, wood, leather, glass): Get a CO2 laser. The 20W version will handle glass etching fine.
• You need both: Consider both machines, or a multi-source system. I'm somewhat skeptical of multi-wavelength lasers—I've seen too many compromises.

In the end, transparency is everything. The vendor who tells you "this laser can do everything" is usually hiding something. I've learned to ask "what's NOT included in that capability" before I ask "what's the price."

Speaking of price: Per USPS pricing effective January 2025, a First-Class letter costs $0.73. That's not related to lasers, but it's a reminder that even small costs add up. My $4,200 CO2 mistake? That could've been 5,753 letters. Just something I think about.