If you put $10,000 in Micron 1 year ago,

you'd have almost $40,000 today. And if

you drop that money into SanDisk, you'd

have over $100,000 right now. That's

because these companies are solving big

bottlenecks in the AI market. So, data

centers are buying everything they make.

And in this video, I'll highlight three

more AI stocks winning big for the same

reason, making them a great way to get

rich without getting lucky. Your time is

valuable, so let's get right into it.

First things first, I'm not here to hold

you hostage. This video is all about

optical networking, and here are the

stocks I'm going to cover. Lumentum,

ticker symbol LITE, which is the leading

supplier of lasers and optical switches

for AI data centers. Coherent, ticker

symbol COHR, which builds high-speed

transceivers and optical interconnects.

And Ciena, ticker symbol CIEN, which

builds long-haul optical networks

connecting multiple data centers

together. And of course, which of these

three stocks I think is the best buy

right now. I want to make the best use

of your time, so let's start with the

big picture and cover what these

companies have in common, like their

markets, their revenues, and their

risks. Right now, AI data centers have

three big bottlenecks: compute, memory,

and networking. They're also limited by

power, so every watt of compute, memory,

and networking matters even more. GPUs

used to be the main constraint, but with

each new generation, like Nvidia's

Hopper, Blackwell, and Rubin, we have

more compute capacity than most models

actually need. So, the real limit

becomes how fast those GPUs can access

data and how quickly that data can be

moved. That's why I spend so much time

covering companies like Micron,

Broadcom, and Arista Networks, and why

they've all outperformed the market. But

something big just happened in this

space, which is why I'm making this

video right now. Google DeepMind just

released a data compression method

called Turbo Quant, which optimizes how

data is stored, retrieved, and reused by

GPUs. Without getting too into the

weeds, every time an AI model generates

a token, it has to store lots of numbers

in what's called the KV cache. For

models with billions of parameters and

long context windows, that KV cache can

eat up most of a GPU's memory. Turbo

quant is a huge deal because it cuts the

KV cache size by over 80%. It speeds up

key parts of inference by up to 8x, and

it can be applied to existing models

without any retraining or fine-tuning.

This is a big breakthrough that will

affect many parts of the AI stack. But

here's the part we care about right now.

Thanks to Turbo quant, GPUs will spend a

lot less time waiting on high-bandwidth

memory, which means they can process a

lot more tokens in the same amount of

time, as long as the network can keep

up. That means the network just became

the next big bottleneck. So optical

networking just became even more

important. At a high level, optical

networking is just networking with light

instead of electricity. Traditional

routers and switches send electrical

signals over copper wire, which works

well for short distances, but breaks

down for large distributed AI data

centers. Optical networks transmit light

through glass fibers, which can carry

much more data over much longer

distances with much lower losses and

interference. So while copper is great

inside a server or inside one rack,

serious AI clusters need optics to move

data between racks, between buildings,

and even across continents using

undersea fiber. Optical links can push

400G, 800G, or even 1.6T of bandwidth

per port. G stands for gigabits per

second. Your copper internet connection

at home is probably hundreds of megabits

to 1 gigabit per second. If you Google

internet speed test, you'll see your

current internet speeds, and hopefully

they're close to what you're paying for.

1 gigabit per second is fast enough to

stream dozens of 4K videos at a time. A

400G optical connection in a data center

is 400 times faster than that. And 1.6T

means 1.6 terabits per second, or 1600G.

That's the kind of bandwidth that large

AI clusters need as GPUs start pumping

out more tokens every second. One more

important point before we get into the

stocks. Optical networking isn't a

single product. It's a full stack. Tiny

laser chips and photonic components

generate and detect the light.

Transceivers are the little plug-in

modules that sit in a switch or in a

server port to both transmit and receive

data, which is why they're called

transceivers. They turn electrical

signals from the chip into light at one

end of the fiber, and then they turn

that light back into electrical signals

at the other end. Then, full optical

systems and software stitch all those

links together into complete networks.

Lumentum, Coherent, and Ciena dominate

different layers of the stack, making

them great ways to invest in the shift

to optical now that networking is

becoming the next major bottleneck for

AI. But, while everyone's focused on AI

and chip stocks, gold has quietly

rallied by 50% in the last year alone.

The problem is owning a single miner

usually means a single team digging

metal out of a single hole in the

ground. That's where Versamet Royalties

comes in, the sponsor of this video.

Instead of running mines, Versamet

finances them and collects a slice of

every ounce produced through royalties

and metal streaming contracts. The

higher gold and silver prices get, the

more each stream is worth. Versamet

produced over 9,800 gold equivalent

ounces in 2025, nearly double the year

before, and they're guiding to over

20,000 oz in 2026. They have an

impressive track record deploying over

$400 million across 28 royalty and metal

streams so far, giving investors

exposure to gold, silver, and copper

across multiple mines on multiple

continents, all with one ticker. And

this isn't their first rodeo. Their CEO,

Dan O'Flaherty, previously built a

royalty company that got acquired for

around $750 million

and he's running First Met with much of

the same team. No wonder more than 75%

of shares are held by strategic

investors and insiders. So, if you want

to invest in a diversified portfolio of

gold and precious metals, especially in

this economy, check out First Met

Royalties. They just began trading on

the NASDAQ on March 6th under the ticker

symbol VMET. This is not financial

advice and I'm not a licensed financial

advisor. So, do your own due diligence

before making any investment decision.

Investing in mining stocks involves

significant risk, including the possible

loss of your entire investment. So, make

sure to see the full disclosure in the

description below. All right, Lumentum,

Coherent, and Ciena all sell into the

same core markets, hyperscale cloud

providers building AI clusters, big

telecom operators upgrading their

infrastructure, and large enterprises

that need high-capacity connectivity.

That ties their revenues directly to the

same capex cycle that drives Nvidia,

Broadcom, and the rest of the data

center stack. They also have similar

business models, heavy upfront

investments in research, fabs, and

specialized manufacturing that turn into

multi-year design wins and supply

agreements for lasers, modules, and

end-to-end systems. That means their

margins depend on how fast hyperscalers

and telecom carriers are expanding.

Since the more they sell, the more

revenue they make against these big

upfront fixed costs. They also have

similar risks, like customer

concentration. A handful of cloud

providers and telecom companies account

for a big chunk of this market. So, any

pauses on spending can hit all three of

these companies at once. They also sit

in the middle of complicated global

supply chains for things like wafers,

lasers, and advanced packaging. So,

supply chain disruptions, export

controls, or tariffs can impact costs

and delay delivery timelines. And they

all have serious competition from tech

giants like Broadcom, Nokia, and Huawei,

which limit their long-term pricing

power once supply catches up with

demand. But on the flip side, they also

have the same big tailwinds. AI optics

alone are projected to jump from roughly

an $18 billion market 2025 to around $90

billion by 2030. And every time

companies like Nvidia and Google solve a

computer or memory bottleneck, they put

more pressure onto the network.

Lumentum, Coherent, and Ciena are all

positioned to capture revenue by

releasing that pressure. Let's start

with Lumentum, ticker symbol LTE,

because they're currently the leader in

lasers and optical switches for data

centers. Lumentum's latest quarter

really tells the story here. The revenue

came in at $665 million,

which is up 65% year-over-year. Their

non-GAAP operating margins came in at

25.2%

up from just 8.2% a year ago. Most of

that growth came from their component

segment, which sells the laser chips and

photonic parts that go inside high-speed

transceivers and co-packaged optics

modules for hyperscalers. When you hear

the term co-packaged optics, that just

means that the optical engines, which

are the parts with the lasers that turn

electric signals into light and light

back to electric again, sit right next

to the chip instead of at the edge of

the circuit board. Today, most switches

use pluggable transceivers in the front

panel, which means every bit of data

moves as electrical signals for several

centimeters before getting converted

into light. At 800 or even 1600G speeds,

that burns a lot of power and adds a lot

of losses. But co-packaged optics put

the lasers in the same package as the

switch chip or the accelerator, which

makes the electrical distance a few

millimeters. That 10x reduction in

distance cuts power per bit by 30 to 70%

it improves signal strength and it lets

hyperscalers keep scaling to 1.6 or even

3.2 terabit per second speeds without

melting their chips or their cooling

budgets. That's why massive AI clusters

are shifting more of their networks to

optics and Lumentum's ultra-high power

lasers and optical engines are being

built specifically for these

deployments. So, if co-packaged optics

are the future of AI networks, then

Lumentum is the company selling the

light. Lumentum's revenue from

components hit $444 million last

quarter, representing about 2/3 of their

total sales. And management is now

guiding for around $805 million

in total revenues for next quarter,

which would imply 21% growth quarter

over quarter. At OFC 2026, which is the

Optical Fiber Communication Conference

held earlier this month, Lumentum's

management laid out a path to $2 billion

in quarterly revenues within the next 2

years, roughly a 3x from here. That

number is backed by real orders. The

backlog for their R300 optical circuit

switch is now over $400 million

and they're already booking massive

orders for co-packaged optics shipments

starting in 2027. And of course, Nvidia

announced a multi-year agreement with

Lumentum earlier this month, including a

$2 billion cash investment and a

multi-billion purchase agreement for

lasers and optical engines. Nvidia gets

priority access to future capacity for

optics and Lumentum gets the funding and

demand visibility to build out a new US

fab in North Carolina to support up to

$5 billion in revenue capacity once it's

fully ramped up. For investors, the

takeaway is simple. Lumentum is becoming

a core part of Nvidia's plan to scale AI

factories. Revenues and margins are

growing fast. Their backlog points to

multi-year growth and Nvidia's $2

billion investment and purchase

commitments help lower their overall

risk. If networking really is the next

bottleneck for AI, then Lumentum is one

of the few companies positioned to solve

it, especially with Nvidia in their

corner. All right, let's cover Coherent

next, ticker symbol COHR. Coherent

reported $1.7 billion in revenue last

quarter. That's up 17% year-over-year

with non-GAAP gross margins of roughly

39% and non-GAAP earnings per share of

$1.29.

That's up roughly 35% from a year ago.

That growth comes from their data center

and communication segment, which

generated about $1.2 billion in revenue

for the quarter. That's up 34%

year-over-year and now makes up over 70%

of their total revenues. While their

legacy industrial segment stayed roughly

flat, that shift from industrial to AI

matters because of how Coherent makes

their money. Unlike Lumentum, which

focuses mostly on components, Coherent

is vertically integrated across the

whole photonics stack. They design and

manufacture their own laser chips,

package them into optical engines, and

build complete 800G and 1.6T

transceivers and other systems around

them. On their latest earnings call,

Coherent's management team said that

most of 2026 is effectively booked with

orders extending into 2027. And for

every dollar of optics they delivered,

they booked over $4 of new demand.

That's a very clear signal that demand

is far ahead of supply, giving Coherent

a very healthy revenue backlog to work

through. But a big part of Coherent's

competitive edge is their 6-in indium

phosphide wafer fabs. Let me break that

down for you. Silicon is great for

compute, but it's terrible for making

light. Indium phosphide, or InP, is

great for making and detecting the laser

light used in fiber optic networks. So,

it's the go-to semiconductor for lasers

in optical transceivers, modulators that

encode that data onto light, and

photodetectors that convert it back to

electrical signals on the other end.

Moving from 3-in to 6-in wafers lets

Coherent make more than four times the

chips per wafer while cutting die costs

by over 60%. That's a massive cost,

capacity, and margin advantage in a

supply-constrained market. At OFC 2026,

that same conference, Coherent showed

off their 1.6 terabit transceivers. They

showed off advanced optical engines and

end-to-end optical systems all built on

the 6-in InP process. And just like with

Lumentum, Nvidia signed a multi-year

strategic partnership with Coherent,

including a $2 billion investment and a

multi-billion dollar purchase commitment

for advanced lasers and optical

networking products. As an investor, I

really like Coherent's full-stack

approach to optics. While Lumentum is

leaning into lasers, optical engines,

and switches, Coherent is building the

entire chain from their 6-in wafers all

the way to high-speed transceivers.

Their revenues and margins are expanding

thanks to explosive demand from AI data

centers. While Nvidia's $2 billion

investment plus their purchase

commitments lower the risk of all that

upfront spending to ramp up production

capacity over the next few years. Talk

about a great way to get rich without

getting lucky. And that leaves us with

Ciena, ticker symbol CIEN. Ciena sits

one layer above Coherent and Lumentum,

turning all of those advanced optics

into full AI networks. Ciena just had a

record quarter, over $1.4 billion in

revenue, which is up roughly 33%

year-over-year. And their adjusted

earnings per share more than doubled

from a year ago. They ended the quarter

with a $7 billion backlog, which grew by

$2 billion in just the last 3 months.

But what's really changing for Ciena is

their customer mix. Revenue from cloud

providers grew by 76% year-over-year and

now represents roughly 40% of their

total revenue. A big shift from their

traditional telecom carrier base to

hyperscalers that are buying optical

systems and data center interconnects.

Ciena specializes in high-end optical

signaling to squeeze more data into each

fiber over longer distances. For

example, their WaveLogic 6 is the

industry's first 1.6 terabyte per second

solution that runs over a single

wavelength. That lets operators double

capacity on existing fibers while

cutting power per bit, which effectively

gives more power back to the GPUs.

They're also working on AI-ready

networks for neoscalers, which are AI

and cloud companies that want simpler,

higher capacity networks that use less

space, less power, and less cooling. The

goal here is to automate more network

operations, keep latency low, and

utilization high across AI clusters that

span multiple data centers. As an

investor, I think of Ciena as the

systems in silicon that sits one layer

above Coherent and Lumentum. Those

companies sell the lasers, the engines,

and the transceivers, while Ciena

packages them into full networks. So,

they win no matter which laser company

comes out on top. And now that we have

all that context, we can answer the big

question, which optical networking

company is the best investment right

now? And if you feel I've earned it,

consider hitting the like button,

subscribing to the channel, and sharing

this video. That really helps me out,

and it lets me know to make more content

like this. Thanks. Now, let's talk about

Lumentum, Coherent, and Ciena stock. I

think Lumentum is the most focused way

to play the component side of AI optics.

They're doing $665 million in quarterly

revenue. They're growing 66% year over

year with 42% gross margins and 25%

operating margins while guiding to even

higher margins over time. On top of

that, they're building the lasers, the

optical engines, and the switches that

are already designed into next-gen AI

networks. And they have Nvidia as a

strategic partner that de-risks a lot of

that roadmap. So, if you want exposure

to AI optical demand without paying for

full systems, Lumentum is for you.

Coherent is all about vertical

integration. They own the 6-in indium

phosphide wafer fabs, design their own

laser chips, and ship complete 800G and

1.6T transceivers. Their data center

revenue is growing fast. Their data

center backlog is growing four times

faster than their shipments. And they

have a multi-billion dollar investment

and purchase agreement with Nvidia. So,

if you're the kind of investor that

looks for vertically integrated

companies that are also funded by

Nvidia, Coherent could be for you. Ciena

sits one layer higher, turning lasers

and optical engines into full networks

that stitch AI data centers together.

They just printed $1.4 billion in

revenue, which is up 33% year-over-year.

They have adjusted gross margins in the

mid-40s and a $7 billion backlog that

gives them better revenue visibility

than most hardware companies can dream

of. With cloud and hyperscaler customers

now driving a big chunk of their

business, Ciena makes sense for

investors who want a more diversified

systems-level way to play the optical

networking buildout. Personally, I'd be

happy owning all three since that would

give me broad exposure to the whole

industry from the laser chips and the

engines all the way up to the full AI

networks. But if I had to pick a single

winner based on today's growth, backlog,

positioning, and

pick Coherent because I think their 6-in

wafer process plus Nvidia's investment

give them the deepest moat and the best

ratio of risk to reward. I'm also a huge

fan of vertical integration. But let me

know which stock you're buying below. Or

if you want me to make a deep dive video

on any of these companies. And if you

want to see more science behind the

stocks, check out this video next.

Either way, thanks for watching. And

until next time, this is ticker symbol

U. My name is Alex, reminding you that

the best investment you can make

is in you.