This is the key take-away: $AAOI is NOT a CPO play. AAOI is a pluggable transceiver play. If CPO gets delayed, AAOI benefits by earning longer from its transceiver sales. This is bullish $AAOI.
For the normies:
Imagine AAOI is a car manufacturer, and the market expects that internal combustion engine cars will become obsolete in one year due to flying cars. Then it gets announced that flying cars will be delayed. That means longer sales of normal cars > bullish for AAOI.
But since the market is irrational > AAOI still sells off because it gets lumped into the “car manufacturer” basket.
Take my money here.
BREAKING NEWS: Archimedes has just disclosed a new long position in $IREN (again). This is noteworthy because he was an early investor in IREN, several quarters before the AI power crunch thesis went mainstream and investors like @leopoldasch, @ericjackson, & others went long
@P_Remarks Second derivative peaking now means first derivative peaks in 2027 and Capex potentially flatlines in 2028. You know what second derivative is bro?
Is $IREN going to announce a new deal tomorrow?
$IREN announced its deal with $MSFT a few days before its previous earnings call. Naturally, a lot of investors expected the same pattern this time around. Yet here we are, the day before earnings, and nothing has been announced.
Personally, I never expected new deals to be announced the same way they were last time. Given $IREN's price action following the previous earnings call, I was fairly sure management would change up their strategy.
Either they announce a new deal independently of earnings, meaning in the weeks before or after the call, or they announce it on earnings day itself.
Keep in mind that $CIFR, $CORZ, $WULF, $APLD, $GLXY, $CRWV, and $NBIS have all announced major deals and partnerships during earnings events. Timing deals with earnings releases seems to be the more common IR approach across this industry.
We should also remember that $IREN moved its Q4 earnings date forward by several weeks, which could be a hint on its own.
Generally speaking, public companies often shift earnings dates to better align with important disclosures. Under SEC rules, companies have 4 business days to disclose material information.
I wouldn’t rule out the possibility that $IREN moved its earnings date forward to align with the week they might sign a new deal, similar to what other industry players have done in the past.
That said, I don’t have a crystal ball, and I won’t speculate on what management may reveal tomorrow. Still, I’m confident they’ll deliver.
Very much looking forward to tomorrow’s call!
PS: Brutal price action today. But as long-term investors, we shouldn’t get distracted by the noise.
Fundamentally, nothing has changed today...
Bullish for $IREN $NBIS
1. “Demand is unsolvable.” This isn’t cyclical demand. This is a structural scarcity of AI compute.
2. Contract duration moving from 3 to 5–6 years, so customers are locking capacity like it’s energy or real estate, not cloud SaaS, which implies long-lived, capital-intensive infrastructure wins.
3. Contracts signed 12+ months in advance show AI buyers are terrified of not having compute. The supply owners gain pricing power + visibility.
4. 6-year GPU depreciation is becoming acceptable, which signals confidence that GPUs won’t be obsolete as fast as feared, and it dramatically improves project IRRs for anyone owning the metal + power.
This is infrastructure scarcity, not software exuberance.
Now we have an earnings date, how about an update from Oklahoma.
"Expedite connection for certain data centers" 👀
Before earnings, we have a big week ahead for $IREN in Oklahoma.
Stay tuned.
$IREN: The Drawdown
I've been in six >50% pullbacks:
1) 2020 TSLA $700 pre-split covid crash to $300 - I actually successfully sold out near the top in Feb to roll SPY puts but I rolled too long and got crushed. I knew how bad Covid was from my relatives in China so I sold TSLA and rolled SPY puts from Feb - May 2023. I made lots of money until the market stopped going down in May then I lost all my gains as my PUTS expired worthless. My lowest price TSLA from this batch was $15 split adjusted from 2019.
2) 2023 TSLA $400 to $100. Brutal 75%+ pullback over two years as I was all in. The bottom was also during Christmas Holidays. My 2023 batch average price was $150-$190.
3) 2023 EOSE $5 to $2. Small allocation but seeing where EOSE is now buying at below $1 was the right move.
4) 2024 IREN $15->5 Culper scare. I made what seemed to be a mistake at the time of rotating large amounts of TSLA when it was $200 to IREN at $15. TSLA ran up first to $400 in end of $2024 so I looked like an idiot for an while.
5) 2025 IREN $15->$6 Tariff Bottom. This was the setup for the big IREN run to $75.
So the reason why I don't trade in and out stem from my lesson in pullback 1. But now looking at this run on IREN, trading out at $75 and in around now would have been optimal but I'm not sure I could have timed that as it always looks easy on retrospective. The sign was then IREN didn't move up materially on MSFT news.
Forward Analysis
Ironically this draw down is what is going to instill discipline in the Neocloud market to not overbuild yet. No more endless debt for $CRWV, $ORCL. No more easy funding for Nscale. No more handouts for Aker, $APLD, $CORZ. $IREN will tighten up to and only sign commitments at higher toplines. I've always said the IREN deal was passable cashflow wise and needed improvement on subsequent deals. $NBIS should follow suit since they are power constrainted anyways. This is how the market enforces all the Neoclouds to only be able to sign deals at stronger economics.
Now the GPU ecosystem has two paths:
1) OpenAI/Anthropic and end SaaS application command higher margins like OpenAI raising their token pricing on ChatGPT 5.2. The end SaaS application will be able to pay more to incentivize their build out. It's entirely possible that OpenAI/Anthropic can build out their own DCs but there's no guarantee they get better economics because they have no experience operationally and they will use either behind the meter gas turbines or nuclear as grid.
The utility of coding agents insane for SDEs and coding is actually harder than most other white collar jobs.
2) Nvidia has no where to plug in their GPUs as Microsoft already has GPUs sitting in inventory. Nvidia will need to lower margins to enable more build out of DCs and buying of their GPUs.
Remember that AWS, Meta, Microsoft, Elon have max Capex budgets too and Nvidia wants to sell all the GPUs it can and not be beholden to negotiating with 3 hyperscalers+Elon while Google competing with their TPUs.
Fighting the FUD
I am holding. For those that are holding too. I have written a series to help people out with the FUD:
1) Edge Inference Layman's Terms https://t.co/nghBqLkuS1
2) Edge Inference More Technical https://t.co/DWidjBNOTF
3) SpaceDC Economics https://t.co/9XytMOrOGL
4) SpaceDC Radiation Damage https://t.co/JeKtVeaYc4,
5) Cashflow FUD https://t.co/KhqLs9YoMs
6) What To Look For in Next Deal Economics https://t.co/uaop7qbEUX
7) AI Breakthrough that Consumes Less Power https://t.co/Wz75GFCOgV
Bought some more $IREN today ✅
Ironically, I'm more bullish on $IREN today than I was when I first entered at $4.8 in May 2024.
The main difference now is that:
1) Over the past 18 months I've spent thousands of hours analyzing and covering this ticker (literally thousands). My confidence in the management team is at all time highs.
2) $IREN has vastly exceeded my expectations as a company, especially with its AI ambitions.
Back in 2024, I treated the move into AI/HPC as the cherry on top, with the "cake" being earnings in $BTC mining. I believed $IREN's edge in mining, plus strong $BTC appreciation, would do most of the heavy lifting.
Most new $IREN investors today don't realize what sentiment was like a year ago toward the company's cloud ambitions. Pretty much nobody took it seriously in the short to mid term. Even some of the biggest $IREN bulls thought management was wasting time and capital on its "little AI cloud project" and should focus entirely on colocation instead.
Little did we know $IREN was playing 4D chess from day one, moving up the value chain, prioritizing cloud over colocation, and actually succeeding in building a fully vertically integrated AI/HPC cloud platform with hyperscalers as clients.
I'm telling you: NO ONE was expecting this a year ago.
This company is on a generational run, and I'm grateful for the kind of buying opportunities we saw today.
Is this the bottom? I don't know. Do I care? Nope.
NFA of course, but I think this is easy money.
Cheers! ✌️
Big picture macro trend is higher, the short-medium term trend is choppy or lower.
Both can be true.
Clearly we’re in a macro AI bull market, we’ve had a few choppy-down periods whether it was tariffs or DeepSeek… neither ended the macro trend, this won’t either in my opinion.
@BagelC47@GlobalCollapse@GlobalCollapse can you help me understand this, can they sign a client for 1gw for ~ 10b arr per year any day now which would be live April 2026? So they would go from 3b arr to 13b arr?