The 'Photonics' ETF That Isn't - and Why the Label Should Make You Pause

> Summary > - Corgi's new Lithography & Semiconductor Photonics ETF (EUV) crossed $150 million in two weeks - a strong debut for a fund launched May 6, but nothing like the $6.5 billion haul Roundhill's DRAM pulled in over the same period > - Despite the "photonics" branding, the top four holdings are TSMC, ASML, Applied Materials, and KLA - established chip equipment names, not optical interconnect specialists > - The silicon photonics market is roughly $2 billion globally; even at aggressive growth rates, it won't move the needle on companies earning $30–$40 billion a year > - The 0.35% expense ratio on an actively managed fund that largely replicates existing semiconductor ETF exposure is hard to justify > - Real photonics exposure exists through names like Coherent and Lumentum - but you can get it directly, not through a rebranded equipment basket

I've been very surprised that the market's embraced this new photonics ETF with so little skepticism about what it actually holds. The narrative is seductive: DRAM proved that narrow thematic ETFs can attract billions in days, so why not replicate the play with the next AI infrastructure bottleneck - optical interconnects? The problem is that EUV doesn't do what the marketing says it does.

Let's start with the factual spine. Corgi launched the Lithography & Semiconductor Photonics ETF (EUV) on May 6, 2026. Within two weeks it surpassed $150 million in assets under management. That's decent - but let me put it in perspective. Roundhill's DRAM ETF, which launched just a few weeks earlier in April, hit $6.5 billion in 36 days. The memory chip fund's launch was a phenomenon. EUV's is a warm reception. In the ETF industry, $150 million is the floor for survival, not proof of conviction.

Now here's where the false narrative gets interesting. The fund's name promises photonics - the technology of using light instead of electricity to transmit data within data centers, which many analysts believe is the next AI infrastructure bottleneck after memory and compute. If you actually read the holdings, that promise unravels quickly. The top positions are TSMC at 9.7%, ASML at 7.9%, Corning at 4.9%, Applied Materials at 4.9%, Lumentum at 4.2%, and KLA at 4.2%. Of the top six holdings, only Lumentum is a genuine photonics company. Corning makes fiber optics, which is related but operates in an entirely different market with different dynamics. The rest are semiconductor equipment and foundry names.

What you're really getting is a chip equipment ETF wearing a photonics costume.

That being the case, the investment thesis collapses under basic scrutiny. TSMC earns approximately $30 billion annually - how much of that revenue comes from silicon photonics? The answer is virtually none. TSMC makes logic chips; it doesn't build optical transceivers. ASML's $3.6 billion in annual revenue comes from extreme ultraviolet lithography machines, which happen to use light, but in the context of etching circuit patterns on silicon wafers, not optical data transmission. ASML is a lithography company, not a photonics company. The fund's full name even admits this - "Lithography & Semiconductor Photonics" - but the branding and the pitch lean entirely on the AI optical narrative.

The silicon photonics market itself is tiny. Industry estimates put it at roughly $2 billion globally in 2025, growing to perhaps $7 billion by 2035. That's impressive growth mathematically, but on an absolute basis, it's a rounding error for the companies this fund actually holds. Even if the market triples in five years, $7 billion is nothing compared to the tens of billions in revenue flowing through TSMC, ASML, and Applied Materials from unrelated business lines.

However, the underlying photonics thesis is not wrong. The companies that actually do it are posting extraordinary numbers. Coherent reported $1.69 billion in quarterly revenue with its data center segment up 33.6%, and a book-to-bill ratio exceeding 4-to-1 - meaning for every dollar of revenue, they're booking over four dollars in new orders. They raised their full-year 2026 guidance to $5.7 billion to $6.1 billion, representing roughly 24% growth. Lumentum posted record quarterly revenue of $665.5 million, up 65% year-over-year, driven by AI data center demand. These are the companies that validate the optical interconnect thesis. Both are in EUV - but combined, they represent roughly 8% of the fund.

The real question is: why pay 0.35% in annual fees to an active manager from a company founded in 2024 to hold a basket that's mostly semiconductor equipment you can already get in a dozen other funds? There are passive semiconductor ETFs with expense ratios below 0.30% that give you the same TSMC, ASML, and Applied Materials exposure. If you want pure photonics, you buy Coherent and Lumentum directly. If you want broad AI infrastructure, the S&P 500 Information Technology sector already covers it.

I've seen this pattern before. ETF issuers chase the previous quarter's hot theme, stretch the definition until it covers their desired holdings, and market it as the next big thing. DRAM worked because memory chips are a distinct, recognizable category with clear cyclicality - and the AI memory crunch was real. Photonics is different. The market is too small, the supply chain is too dispersed, and the pure-play names are too few to support a 40-position fund that doesn't look like what it claims to be.

Of the companies in this space, I favor Coherent and Lumentum for their demonstrated execution, accelerating AI revenue trajectories, and the structural fact that hyperscalers are now mandating optical upgrades as part of every data center buildout. If you want photonics exposure, buy the actual photonics companies, not a lithography equipment fund with a rebrand.

EUV is a Hold for investors who already want broad semiconductor equipment exposure and don't mind a slightly elevated fee for active management. It is not a photonics play, despite the marketing. The real photonics opportunity is in the names the fund underweights, not the ones it uses as its anchor.

In my opinion, the false narrative here is that thematic ETF branding equals thematic purity. It doesn't. Read the holdings.