$GOOGL CEO @sundarpichai said it himself…
“We are compute constrained ... our cloud revenue would have been higher if we had more compute"
So bullish for energy and power plays..
English
Shachi Merchant
1.3K posts
Shachi Merchant
@AIGuruKidd
🚀 Fmr Head of the 1st Robotics ROBO ETF. Speaking on AI & Identifying the bottlenecks of the AI revolution: GPU/CPU, Photonics, Energy, & more🛠️🤖🦾💻
Dallas, TX Katılım Mart 2017
142 Takip Edilen90 Takipçiler
It’s absolutely wild to see all the downstream companies start to get action, especially Bloom energy. That’s obviously a huge sign based on earnings reports and data centers not being able to be built that it’s not because of hyper skaters not having demand. In fact it’s the exact opposite their compute constraint, and they don’t have the power and energy to build. $BE
Lin@Speculator_io
The best earnings reactions today: $COCO Vita Coco +29.67% $BE Bloom Energy +27.21% $NXPI NXP +25.55% $VIAV Viavi Solutions +20.36% $PI Impinj +17.80% $FVRR Fiverr +16.99% $QCOM Qualcomm +15.66% $EAT Brinker +14.45% $FORM FormFactor +13.53% $STX Seagate +11.10% $ETSY Etsy +10.11% $UMC UMC +9.35% $SBUX Starbucks +8.45% $V Visa +8.26% $FFIV F5 Networks +8.02% $GD General Dynamics +7.99% $VRNS Varonis +7.27% $VRSK Verisk Analytics +6.53% $GOOGL Alphabet +6.44% $TMUS T-Mobile +6.13% $AMZN Amazon +4.60% $VIRT Virtu Financial +3.91%
English
Based on the latest earnings , we can obviously tell that there’s massive demand in the hyper-scalers are completely compute constrained. The order book for cpu, gpu , and memory companies will continue to ramp up. Power and energy constraints are also limiting the construction of data centers, which explains why reports of their inability to be built are emerging.. So expect to see massive infrastructure energy investment moving forward $VRT $VST $ETN $BE $GOOGL $AMZN $META $AAPL
Gaurab Chakrabarti@Gaurab
You cannot buy a new gas turbine until 2030. Order books at GE, Siemens, and Mitsubishi stretch to 2029. Turbine prices have nearly tripled since 2019. Every AI data center needs power and every gas plant needs a turbine. And every turbine has one part that bottlenecks the entire industry: The blade. It has to survive in gas 500°C above the melting point of the metal it's made from and spin at up to 20,000 RPM under 10,000 g of centrifugal force. Each blade is grown as a single crystal of nickel superalloy, pulled through a vacuum furnace at 3 mm per minute. A set of blades costs $600,000 and takes 90 weeks to grow. The same metallurgy powers modern jet engines. Only 3 companies on Earth can build one. China spent $42 billion trying to catch up. They bought a Russian fighter engine, took it apart, and copied every part. Their copy ran 30 hours between overhauls versus 400 for the original. Modern Western engines run 4,000. You can reverse engineer the shape of a turbine blade. You cannot reverse engineer 60 years of metallurgy.
English
I mean right off the bat , if the hyper scalers are compute constraint, they’re gonna do two things 1) they’re gonna buy as many GPU CPUs and memory as they can to build out supply so they can serve their customers demand and 2) they’re going to continue to build out their own chips .. ** Also, continue to look for constraints in the industry , as those where will prob see the most growth in the next year
Shay Boloor@StockSavvyShay
$AMZN, $MSFT, $GOOGL and $META just updated guidance and now plan to spend a combined ~$710B on CapEx this year to build the AI economy: • Amazon ~$200B • Microsoft ~$190B • Google ~$185B • Meta ~$135B
English
🚨 WSJ just dropped a wild claim that could affect memory!
Majestic Labs’ Prometheus server claims 1,000x the memory of NVIDIA GPUs, 128TB high-speed RAM in ONE box.
Could this be the reason for $MU’s ~4% pre-market dip today.
Reality check: Micron just hit an all-time high yesterday. This dip is classic profit-taking + some 2028 growth questions. Majestic is still early-stage (stealth until Nov 2025). Real AI memory demand (HBM) is sold out through 2026.
Not selling my $MU $DRAM $SNDK $SKHYNIX
You? Bullish or worried? 👇
Follow for no-BS AI stock breakdowns 🔥 #MU #Micron #AI #Semiconductors #Stocks
English
I actually think $MRVL used this 100 % as a way to get out their contract with $POET . It def begs the question of “is the tech itself flawed or did $MRVL find out greater troubles with their partnership”…
Daniel Romero@HyperTechInvest
Did $MRVL use this NDA breach as a way out? Would $MRVL reject a top-tier optical interposer in such a tight supply chain if the tech were actually good? It makes you wonder Many supply contracts are disclosed in this industry, so why the concern for secrecy?
English
🚨 $SANM JUST CRUSHED EARNINGS LIKE A ROCKET SHIP! 📈💥
Q2 FY2026:
• Revenue: $4.01 BILLION (+102% YoY) 🔥
• Non-GAAP EPS: $3.16 (massive beat!) 💰
AMD AI racks + ZT Systems integration = $1.88B firepower 🧠
Full-year guidance RAISED sharply + $600M buyback approved!
Stock +14% in after-hours already!
The hidden AI infrastructure + photonics beast is AWAKE 🚀
$SANM $AMD
Buying the dip or FOMOing in? Drop a 🔥 if you’re bullish! 👇
#SANM #AI #EarningsSeason #Stocks #AMD #Photonics #Investing
English
Well you called it !! Great call $SANM
Painting a Simple Picture of Sanmina’s Photonics/Optics Involvement:
Think of today’s AI data centers like a massive city of super-intelligent brains (thousands of GPUs). These brains need to talk to each other at insane speeds—trillions of operations per second—while keeping power and heat under control. Traditional copper wires are like old telephone lines: they get too slow, too hot, and too power-hungry over the distances inside huge racks and across the data center.
Photonics = the fiber-optic superhighway solution. Data travels as pulses of light (lasers) instead of electricity. It’s faster, uses far less power, generates less heat, and scales better for AI workloads.
Where Sanmina fits in (their Advanced Microsystems Technologies / AMT division has 30+ years of optical expertise):
1. They build the “on-ramps and off-ramps” — high-speed optical transceivers (like 800G and cutting-edge 1.6T modules). These tiny devices convert electrical signals from chips into light (and back) using lasers, modulators, photodetectors, etc. Sanmina does the precision assembly of Transmitter Optical Subassemblies (TOSA), Receiver Optical Subassemblies (ROSA), and full pluggable modules.
2. They do the advanced packaging — This is the hard part. Sanmina has made major 2025 breakthroughs in silicon photonics (SiPh) packaging, heterogeneous integration, and co-packaged optics (CPO). They integrate tiny photonic integrated circuits (PICs) with electronics, chiplets, and waveguides so light can move inside the package with minimal loss. This enables higher-density, lower-power optical interconnects right next to the GPUs/CPUs.
3. They handle the full fiber infrastructure — fiber cables, connectors, active optical cables (AOCs), line cards, and fiber trays that carry the light across the rack or data center.
4. They assemble the whole AI “city block” — Thanks to the ZT Systems acquisition, Sanmina now also builds the complete liquid-cooled AI racks and clusters (with AMD chips inside) that include all this photonics hardware already integrated and ready to deploy.
Bottom line (easy analogy):
Copper is like sending messages by shouting across a noisy room. Photonics (what Sanmina specializes in) is like using laser pointers and fiber-optic cables for crystal-clear, lightning-fast communication. Sanmina doesn’t just sell the laser pointers—they design, package, manufacture, and integrate the entire high-speed optical “highway system” into massive, ready-to-run AI data center racks. Their new deep partnership with AMD supercharges this for the next wave of AI infrastructure.
This positions Sanmina as a key “behind-the-scenes” player in the AI buildout: not just a contract manufacturer, but a critical enabler of the optical interconnects that make hyperscale AI clusters possible.
Gaetano@crux_capital_
There is a company that... > Has Exposure To Optics > Has Exposure to CPU / Inference boom > Has Exposure to Transformers > Grew Revenue > 50% YoY > Trades at less than 1x rev And practically no one is talking about it Guesses?
English
@pepemoonboy Great if it wasn’t a paid sponsorship , I’m
Sure this could be done easily through open-claw, perplexity, codex , Claude chrome, etc . Cool concept tho!
English
Just automated what a $200k/year Goldman analyst does...
Built a Xynth workflow that pings me every Sunday:
>Sector rotation data
> Mid-caps with institutional accumulation
> Setups with unusual options flow
> 5 options spreads w/ full PnL
This is retail's AI edge in 2026.
Xynth@xynth_m
English
Oracle just made a massive power move for the AI era 🔥
Switching Project Jupiter in New Mexico to up to 2.45 GW of Bloom Energy fuel cells instead of gas turbines & diesel generators? That’s a 92% NOx emissions cut while keeping the entire AI campus on a single microgrid.
Clean, reliable power at hyperscale — this is exactly how big tech should be scaling.
Who’s next? $BE $ORCL
#AI #CleanEnergy #DataCenters #SustainableTech #BloomEnergy
Wall St Engine@wallstengine
Oracle said it will now power its Project Jupiter campus in New Mexico with up to 2.45 GW of Bloom Energy $BE fuel cells instead of gas turbines and diesel generators. The switch cuts NOx emissions by about 92% and keeps the AI data center on a single microgrid campus.
English
Oracle said it will now power its Project Jupiter campus in New Mexico with up to 2.45 GW of Bloom Energy $BE fuel cells instead of gas turbines and diesel generators. The switch cuts NOx emissions by about 92% and keeps the AI data center on a single microgrid campus.
English
$XE closed above $36!! I was spooked when I saw $arkk invested but patience pays off. Took a 2500 share position and closed out before market close today. Could run some more but this isn’t a day trade. This is thematic shift to SMRs in the next 3-5 years ! $AMZN $XE $OKLO $SMR
Shachi Merchant@AIGuruKidd
Holy smokes $XE up big at $36 in after trading !! $AMZN holds a 20% stake in the SMR startup and is partnering with X-Energy to deploy carbon-free nuclear power for its AWS data centers. This comes at a time when hyperscalers are racing to secure 24/7 baseload energy to fuel the massive power requirements of the AI revolution. $GOOGL $META $MSFT
English
Talk about a "loose lips sink ships" moment. 🚢 $POET crashing 45% today after $MRVL canceled their orders due to a confidentiality breach.
Hard lesson for small-cap tech: don't celebrate the win until the ink is dry and the partner is ready to talk. #Stocks #POET #MRVL
Is $POET a buy at these levels, or is the trust gone?
English
@ivanofde @StockSavvyShay @FuturumEquities Yes, most likely. At least short term till we confirmation of $AAPL , $AMD or other downstream involvement @aleabitoreddit
English
@StockSavvyShay @FuturumEquities What about $SIVE ? Do you think the thesis is in danger due to the news?
English
I said on last Friday’s Futurum Equities Podcast that $POET was one photonics name I would not touch due to management red flags and paid influencer promotion.
Today it’s down 45% after $MRVL canceled the Celestial AI orders due to a breach of confidentiality.
Shay Boloor@StockSavvyShay
$POET down over 40% after $MRVL canceled all purchase orders tied to Celestial AI including initial production orders first disclosed in 2023. POET said the cancellation was tied to alleged confidentiality breaches while it continues serving other customers.
English
@StockSavvyShay That is extremely rough news… for not only $POET but also $SIVE
English
$POET down over 40% after $MRVL canceled all purchase orders tied to Celestial AI including initial production orders first disclosed in 2023.
POET said the cancellation was tied to alleged confidentiality breaches while it continues serving other customers.
English