GLW: a 175-year-old glass company as an AI infra play

thewontoninvestor
11-07

update (Oct 27): $Corning(GLW)$  just reported Q3, so here's the quick scorecard against the conditions i wrote down before the print. 

Optical grew 44% (my bar was 30%), with Enterprise Networks, basically the AI fibre business, up 84%. core sales were $5.08B, up 19% and above their own guidance. solar still lost money ($7M), so that one's a miss. they signed four multibillion-dollar deals including AT&T and Verizon, but nothing new from a hyperscaler or on glass substrates. (and Verizon, the customer behind the ~80 million fibre-mile deal, reported the day before: 406,000 new phone subscribers, about double what analysts expected, plus 155,000 new fibre broadband customers.) and i still need the 10-Q to check share issuance. so: 1 hit, 1 half, 1 miss, 1 unknown, and none of my walk-away conditions triggered. the full scorecard is on my substack. everything below is the original thesis, written before the print.

my four buy conditions vs what actually showed

hey everyone, Wonton Investor here. let's talk about Corning.


quick disclosure first: i don't own GLW, it's on my watchlist. i wrote this thesis before Corning's Q3 print on purpose, because the last time i bought a stock first and wrote the thesis later (hi, $Coinbase Global, Inc.(COIN)$ ) i broke my own entry rule. so this time it's thesis first, money later


the idea behind it:

my view on AI infrastructure right now is that the hottest parts of the trade - the optical modules that connect GPUs and the power equipment that runs data centres - are heading for commoditisation. demand is real, but every company in those spaces is building new capacity at the same time, and when everyone expands at once, prices fall and margins get squeezed. it happened to solar panel makers and it happened to optical component makers in 2000.


so instead of asking "who's growing fastest", i've been asking what's still scarce after everyone finishes building their new factories, who owns it, and where else it gets used. and that's how i landed on Corning.


the recipe book


Corning doesn't really sell a product, it sells a recipe book.


think about bread. bread is a commodity, there's a bakery on every corner. but a bakery that spent a hundred years learning how to make one specific dough at huge scale, with ovens that take years to build, and that sells that dough to every restaurant in the city? that's a supplier nobody can easily replace, even when bread gets cheap. Corning is the dough supplier.


what it actually owns is glass chemistry and manufacturing know-how - glass that's ultra-pure, ultra-flat, doesn't expand when it heats up, doesn't lose light over long distances, and can be made in huge volumes without defects. you learn that by running furnaces for decades, and a new glass melting tank takes years to build and qualify. so the product can get commoditised, but the process is a lot harder to copy.

one glass and fibre know-how feeding 4 different gaps: AI fibre, the chip package, non-AI telecom, and energy

the four gaps

i see four places that recipe book gets used.

fibre for AI data centres. this is the one the market already prices in. Corning has an up-to-$6B multi-year deal with Meta, a multi-billion dollar deal with AWS, and a partnership with Nvidia to grow its US optical-connectivity capacity roughly 10x. and they sell it as co-investment, where the customer commits to years of volume before Corning builds the capacity, which is how you protect margins before a price war starts.


the chip package. this is the one i think people are sleeping on. big AI chips are warping the organic base they sit on, and the industry's answer is glass core substrates, which stay flat and expand with heat at about the same rate as silicon. the supply of that kind of glass is basically an oligopoly of Corning and a few peers. it's early though, volume production is expected around 2027-2030.


demand that has nothing to do with AI. in September Corning signed a multi-billion dollar deal with Verizon for roughly 80 million miles of fibre. add display glass, phone cover glass, cars and pharma, and AI is the growth engine but not the only one.


energy, through materials. Corning makes US polysilicon and solar wafers, and management is targeting a $3B revenue stream there. it's the same move as the fibre business: make the input, domestically, where there's policy support.


the full version on my substack goes through each gap properly, with what has to go right for every one of them, plus a diagram of exactly where Corning sits inside an AI chip package.


the numbers:

Q2 core sales were $4.74B (+17%), core EPS was $0.78 (+30%), Optical Communications did $2.07B (+32%) with Enterprise Networks up 65%, and Solar did $438M (+90%). management's Springboard plan targets a $20B sales run-rate by the end of this year, $30B by 2028 and $40B by 2030.

Corning's Q2 numbers: core sales, optical communications, enterprise networks and solar, with growth rates shown

so why haven't i bought it? 

mostly because of the price. at around $155 it's down roughly 40% from its June high near $255, but it's still about 72x trailing earnings, which is a lot for a glass company. and it trades like a pure AI stock - it fell 13% in one day in September on an AI-slowdown scare, even though a big chunk of the business has nothing to do with AI.


there's also the awkward fact that solar, the industry i keep using as my main example of commoditisation, is one of Corning's four gaps 😭 the segment grew 90% and still lost money last quarter.


so the business case is strong, and the stock case is somewhere between fine and expensive. which is the same place i ended up on AMD.


should YOU buy it? i'm not a financial advisor, so i can't tell you that, and i haven't decided myself yet. but i did write down exactly what would make me buy and what would make me walk away, before the earnings came out, so i can't negotiate with myself later.


peace out,

The Wonton Investor


not financial advice. i don't own GLW as of writing - this is just my own reasoning, and i'd rather you disagree with it and think it through yourself than copy it.

Modified in.11-14
Corning's core sales growth and performance in the third quarter of 2026 exceeded expectations.
Corning's core sales reached $5.08 billion in the third quarter of 2026, a year-over-year increase of 19%, exceeding market expectations of $5.02 billion. Sales in the optical communications business rose 44% to $2.38 billion, with core earnings per share of 91 cents, exceeding expectations. The company expects core earnings per share for the fourth quarter to be between 96 cents and $1.00, with core sales of between $5.2 billion and $5.3 billion. Following this news, Corning's stock price rose by more than 3% in pre-market trading.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment