Pinned Tweet
A year ago, Michael Burry (who has called 20 of the last 2 crashes) said $NVDA chips run on a 2-3 year cycle.
$CRWV , yesterday:
Ampere (2020) and Hopper (2022) pricing up 30-50% QoQ, above original contracts
Multi-year deals signed beyond the GPUs' depreciable lives
~300K GPUs come off contract in 2027-28. At today's prices, re-renting them is worth nearly $10B a year, with zero additional capex
A chip launched six years ago now rents above its original contract.
And there are people who call this a bubble?
Impossible not to be bullish with these numbers.
The depreciation bear case just met the order book.
$IREN $NBIS
COREWEAVE $CRWV Q3’26 EARNINGS HIGHLIGHTS
🔹 Revenue: $3.66B (Est. $3.55B) 🟢; +168% YoY
🔹 Adj. Operating Income: $379M (Est. $234M) 🟢
🔹 Adj. EBITDA: $2.26B (Est. $2.04B) 🟢; +169% YoY
🔹 Adj. Net Loss: -$529M (Est. -$677M) 🟢
🔹 GAAP EPS: -$1.27
Raises FY26 Guide:
🔹 Revenue: $13.0B-$13.3B
🔹 Adj. Operating Income: $1.3B-$1.4B
🔹 CapEx: $37B-$39B; from $35B-$39B
🔹 Active Power: >1.85 GW
🔹 Run-Rate Revenue: $20B-$21B
Other Q3 Metrics:
🔹 Revenue Backlog: ~$132.5B
🔹 Active Power: >1.7 GW; +200 MW added
🔹 Contracted Power: >4.6 GW; +900 MW added
🔹 International Power: >1.3 GW
Announcements:
🔹 CoreWeave Forge: AI development environment launched
🔹 Vera Rubin NVL72: 4.8x greater token throughput
🔹 Convertible Notes: ~$4.2B raised
🔹 DDTL 5.5 Facility: $2.6B closed
Comments:
🔸 “This was a breakout quarter for CoreWeave, with revenue increasing by more than $1 billion sequentially and adjusted operating margins reaching double digits a quarter ahead of plan.”
🔸 “approximately $9 billion in customer prepayments and deposits expected in the next three months alone.”
$NVDA daily RSI just hit 22.2, LOWEST SINCE JULY 2008
Back then, Nvidia had just slashed guidance and taken a charge for defective chips. The stock fell 31% in a day.
This week it reported both revenue and earnings growing above 100% YoY..
INSANE
2
133
$NVDA back to its main trendline (since beg. 2023). Would be highly surprised if this level doesn't hold, given current P/E
1
3
426
$IREN at $14.. But look at balance sheet figures:
- Assets: $20.6B
- Liabilities: $15.5B
- Equity: $5.1B
- Market cap: $6.2B
The market pays $14.70 per share for $12 of book value Barely 1.2x
Meanwhile, the company targets >$4B ARR by end of Jan as Horizons 2-4 are delivered to $MSFT
Add net debt and EV/ARR is < 2.7x
The current multiple is VERY HARD to justify. The last time $IREN traded near book value was April 2025. It became my best purchase ever. I’m buying again at these levels.
This is a generational opportunity! Bookmark this
$QQQ The Nasdaq 100 has been testing the trend line that has held it since April 2025
That trendline started on Liberation Day lows, when Trump announced the tariffs
It was tested twice in the last few days. That’s good, but not necessarily bullish.
To reverse the trend, the index needs to break above the short-term downtrend (in red) that started in October, preferably with strong volume!
Level to watch: ~28,500
One of the most underrated risks I see in $NVDA is talent retention.
Interim CEO Colette Kress promised employees raises and more equity in December, months ahead of schedule (per WIRED).
Why? Well. Talent retention, especially now..
But fun fact: A survey found that nearly 80% of Nvidia's employees are millionaires, with half having +$25 million in net worth.
The classic retention tool is stock that vests over years. But that doesn't hold people who are already rich.
The signals are already there: a director wrote to the board that "irreplaceable talent" is contemplating leaving (per WSJ). And Nvidia has already ruled out Huang's return.
How do you think the company will handle this forward?
What a great time to be alive.
Hedge funds have severe PTSD and are trading 2026 like it's 2000.
Their tech long/short ratio just hit 1.5, the lowest in Goldman Sachs' data, which goes back to 2016.
The numbers say otherwise:
- S&P 500 EPS: +32% in Q3, the third straight quarter above 25%.
- Tech EPS: +69%.
- $NVDA and $GOOGL Cloud revenue: +103% YoY.
And bears argue AI capex doesn't pay off? Google Cloud runs a 36% operating margin. Alphabet as a whole: 32%. The AI business is already above the company average. This is adding value!
Some AI infrastructure names are down 50%. To get back to where they were, they need +100%. When sentiment turns, these recoveries will be VERY violent IMO.
I believe some will lose A LOT of money in this phase of the cycle. But it won't be those who looked at the numbers...
$IREN $NBIS $CRDO $NUAI $INOD
Take a look at sentiment right now.
Because of a personal event involving Jensen Huang, there are now widespread calls that the "AI bubble has popped."
Sentiment is as though AI has failed, the investment wave is over, and we are in a global depression.
In reality? The S&P 500 is up +7% YTD, +60% in 5 years, and ~7% away from a record high, with US AI CapEx running at over $1 trillion per year.
Meanwhile, AI-associated sectors and tech investments have accounted for ~44% of total US GDP growth on a rolling four-quarter basis.
And, Nvidia, $NVDA, is guiding nearly $750 billion in revenue for next year, even without Jensen Huang.
The underlying fundamentals of the AI Revolution have not suddenly deteriorated because of a leadership crisis at one company.
The unfortunate result of this drawdown, as with nearly every other correction since the AI Revolution began, will be countless investors getting crushed on the way back up.
Today's fearmongers become tomorrow's short squeeze as soon as sentiment shifts again.
The comments on this post will be a perfect example of just how polarized sentiment has become in the "doom" direction.
Asset owners will win and the AI Revolution is far from over.
$IREN $NBIS
$NVDA just explained who wins the next phase of AI.
- Data center revenue: +110% YoY.
- Neoclouds are seeing “extraordinary” demand, in Nvidia's own words.
- The next critical phase is securing land, power and shells for data centers.
- The risk is the balance sheet: many neoclouds are growing faster than their balance sheets can support, and Nvidia's commitments only kick in if they secure financing.
The winners are those who already have power and financing locked in.
$IREN owns its sites with secured grid power and raised $19B in 12 months. On its latest contract, the customer committed to prepay 90% of the GPU capex. At $NBIS, customers already prepay over 100% of capex on new contracts.
Nvidia CFO Colette Kress on neoclouds:
“Neoclouds and frontier labs are seeing extraordinary demand for AI infrastructure, yet many are growing faster than their balance sheets and long-term credit profiles can support.”
Nvidia says securing land, power, and data center capacity has become the next critical phase of the AI buildout.
To address this, Nvidia is helping select neoclouds secure that infrastructure. Its commitments are conditional on customers obtaining financing and getting the compute capacity deployed and operational. Commitments can be reduced or eliminated if those conditions aren't met.
Under these arrangements, Nvidia earns revenue from upfront infrastructure sales and could also receive a share of the revenue neoclouds generate from third-party customers.
bitcoin:native Bitcoin has 1 week to repeat history.
Previous bear markets lasted 363 to 410 days:
2014: 410 days
2018: 363 days
2022: 376 days
The current one began on Oct 6, 2025: day 410 is Friday, Nov 20.
The low so far this cycle came on Jul 1, at $57.7k (day 268).
For history to repeat, Bitcoin would need a new low below $57.7k by Nov 20.
If not, either this was the shortest bear since 2011, or it will be the longest ever.
ethereum:native $MSTR $BMNR
$NVDA Board is preparing to oust Jensen Huang.
Short term, this may hurt neoclouds. Jensen was their biggest backer: Nvidia invested in $CRWV and $NBIS and treated them as partners. A more conservative new CEO may pull back from those deals.
Long term, the logic stays. $GOOGL $AMZN $MSFT and $META design their own chips. $CRWV, $NBIS and $IREN don't. They don't compete with Nvidia. They buy from it.
A CEO is replaceable. Customers that don't build rival chips are not.
BREAKING: Market expectations for Jensen Huang to be removed or resign from his role as CEO of Nvidia, $NVDA, surge to a record 85% chance.
The stock has now erased -$650 billion in market cap since the news of Jensen Huang's first affair arose.
Goldman Sachs projects $ANT Anthropic's revenue to top $600B+ by 2030, over 130x its 2025 revenue level.
That implies a staggering 165% revenue CAGR through 2030!
But the most critical metric for the market lies in capex: $900B already committed in compute capacity to support customer demand.
This signals a massive, aggressive pipeline flowing straight into the infrastructure layer (neo-clouds & energy).
The real bottleneck in AI isn't demand... It's the global scarcity of compute capacity.
$IREN $NBIS $META $ORCL
$MSFT is scaling back the AI capacity it rents from $ORCL
The reason? Older-generation $NVDA chips that no longer meet Microsoft’s specs.
That’s not a demand problem. That’s Microsoft upgrading.
And old GPUs don’t sit idle: Oracle says some 4-year-old GPU contracts renewed at a 20% premium. $IREN re-contracted H100s at 2x.
For $IREN and $NBIS holders: their Microsoft contracts runs on GB300, among Nvidia’s latest. And nothing was cancelled.
Microsoft Winds Down Oracle AI Compute Deal $ORCL $MSFT
Microsoft is scaling back rented AI computing capacity from Oracle after choosing not to renew parts of the agreement. The capacity had become less useful as it relied on older-generation Nvidia chips that no longer matched Microsoft’s specifications, prompting the company to turn off the rented infrastructure over recent months.
The move forms part of a broader reassessment of Microsoft’s cloud partnerships and AI infrastructure spending. Oracle maintains that demand for its AI infrastructure remains strong, including for older GPUs, and said in September that some four-year-old GPU contracts were being renewed at a 20% premium to original pricing.
B. Angelo | βα retweeted
🚨Panic Selling in Markets
S&P 500 -4% | Nasdaq 100 -6%
What changed? NOTHING
Q3 earnings for S&P 500:
- EPS +30.2% | Revenue +13.4%
- 82% beat estimates
Fwd P/E: 18.7x | 10-year avg: 19.1x
Earnings up, price down.
Panic selling is how you lose money.
More on my profile.
$INOD’s biggest opportunity isn’t even in the guidance.
It’s negotiating a program with a leading frontier lab at a run rate of “millions of dollars per month”
Guidance excludes large opportunities until they’re awarded. This one is out, for now.
Why that much money? AI is running out of internet (literally) to learn from. As a matter of fact, Epoch AI estimates that public human text runs out between 2026 and 2032. The next data has to be produced..
Robotics is the clearest case: $47.4B of VC in H1 2026, and no internet of robot data to train on.
$INOD captures it with camera headsets in homes and across hundreds of businesses, plus its own motion-capture lab. One new humanoid customer alone could be a multi-million opportunity next year.
That’s where $INOD sits.
The bear case on $INOD was always customer concentration. The Q3 call shows why it's fading.
The traditionally largest customer ended a program in October: -$12M in Q4, but is having replacement expected by Q1
But it's no longer #1. The former #2 took the lead: a new $29M program, plus ~$18M already signed in Q4
A hyperscaler did $3.1M over the last 12 months and could pass $20M in 2027
They added 5 new "logos", and are talkimf with a leading frontier AI lab on a major expansion, and with a major social media company
If all you see is the $12M, you're looking at a program, not the customer book.
The bear case on $INOD was always customer concentration. The Q3 call shows why it's fading.
The traditionally largest customer ended a program in October: -$12M in Q4, but is having replacement expected by Q1
But it's no longer #1. The former #2 took the lead: a new $29M program, plus ~$18M already signed in Q4
A hyperscaler did $3.1M over the last 12 months and could pass $20M in 2027
They added 5 new "logos", and are talkimf with a leading frontier AI lab on a major expansion, and with a major social media company
If all you see is the $12M, you're looking at a program, not the customer book.
$INOD just showed why the October selloff made no sense.
- Revenue $95.6M, +53% YoY, 15% above consensus
- EPS $0.29, ~70% above consensus
- FY26 revenue growth guide raised to 42%+ (from 40%+)
- A customer under 5% of revenue a year ago is now the largest: less concentration
- With a large hyperscaler: $20M+ of revenue forecast for 2027, vs $3.1M over the last 12 months
Watch: adj. EBITDA margin fell to 18.5%, and a program at a large customer is winding down, which the company expects to replace in Q1.
$IREN Q1 FY27. What matters:
- Revenue $228M vs $221M expected. AI Cloud already $200M, at an 86% gross margin
- The $4B+ ARR target is now FULLY contracted and funded
- It moves from year-end to January, possibly end of February, after a customer-requested design change and supplier timing (We all know it's $MSFT )
- New 5-year contract at ~$22M per MW, with 90% of GPU capex prepaid by the customer. Plus a short-term deal at $40M per MW 🔥
- $20.3B backlog, about two-thirds investment-grade
The delay is 1 month, if so. Demand and pricing per MW tell the rest.
$IREN Q1’27 EARNINGS HIGHLIGHTS
🔹 Revenue: $228M (Est. $221M) 🟢; -5% YoY
Segment Net Revenue:
🔹 AI Cloud: $200M
🔹 Bitcoin Mining: 28
Announcements:
🔹 ARR Target: More than $4B fully contracted and funded; targeted for January 2027
🔹 Five-Year Contract: ~$22M per IT-MW; 90% of GPU capex prepaid
🔹 Short-Term Contract: $40M per IT-MW
Other Q1 Metrics:
🔹 AI Cloud Segment Gross Margin: 86%
🔹 Adjusted EBITDA: 66
🔹 Net Income (Loss): (263)
Key Updates:
🔹 Unrealized Loss on Financial Instruments: (28)
🔹 Revenue Backlog: $20.3B; approximately two-thirds investment-grade
🔹 FY27 Capex: $15B remaining planned
Comments:
🔸 Daniel Roberts, Co-Founder and Co-CEO: “The more than $4 billion of contracted ARR is funded and now targeted for January, after a customer-requested design change and supplier timing moved it from year-end; it may extend to the end of February.”
🔸 Daniel Roberts, Co-Founder and Co-CEO: “Since quarter-end we have signed a five-year contract at about $22 million per megawatt with 90% of the GPU capex prepaid by the customer, and a short-term contract at $40 million per megawatt.”
$INOD just showed why the October selloff made no sense.
- Revenue $95.6M, +53% YoY, 15% above consensus
- EPS $0.29, ~70% above consensus
- FY26 revenue growth guide raised to 42%+ (from 40%+)
- A customer under 5% of revenue a year ago is now the largest: less concentration
- With a large hyperscaler: $20M+ of revenue forecast for 2027, vs $3.1M over the last 12 months
Watch: adj. EBITDA margin fell to 18.5%, and a program at a large customer is winding down, which the company expects to replace in Q1.
$INOD is down roughly 27% to $44 on reports it laid off about 550 people in India, with more cuts in the Philippines. No official statement yet.
Some perspective:
- Headcount went from4.3k in 2023 to 10.1k in 2025
- The India cuts are just 5% of the company
- On LinkedIn, almost half of its India and Philippines profiles are in business development and operations. IT and engineering are about a quarter
The market's fear:
The termination email cites a ramp-down of the work they were hired for. With the top two customers at 71% of Q2 revenue, the fear is that one of them is pulling back.
My perspective:
A project winding down is different from a client walking away. On the last call, management said growth is "increasingly research-driven," and that the largest customer had just awarded "a significant new program" in agentic AI, "now scaling."
Older, more manual work winding down while newer programs scale is exactly the kind of shift that shows up in margins, and adjusted gross margin went from 43% to 49% in a year. With a new CEO since September, a reshuffle makes even more sense.
And it's not an isolated issue: yesterday $MSFT offered voluntary exits to roughly 60,000 employees, over a quarter of its workforce. In April, $META announced a 10% cut.
$INOD Metrics: Q2 revenue +58%, Adj. EBITDA doubled, FY26 +40% growth guidance reiterated.
A 27% drop for a 5% trimming looks a massive overreaction to me. Nov 5 earnings will provide clarity.
I am buying $INOD here.
159
$IREN This week the market sold Texas data centers as if they were all about to stop. Here's the part few people read.
Gina Hinojosa said that data centers that already have permits WILL BE ALLOWED TO FINISH CONSTRUCTION. That covers essentially all of 2027 and well into 2028.
The new rules ONLY apply to newly permitted projects, for 2028-2029 delivery: pay taxes, no power price hikes for households, no hit to water, union labor and a seat at the table for communities.
In other words: permitted, powered and under construction is out of the line of fire. Building new gets slower and pricier.
For $IREN , with Childress and Sweetwater 1 already energized and under construction, that's exactly the scenario I laid out this week.
$CRWV $ORCL $NBIS
The governor-elect of Texas, Gina Hinojosa, has explicitly said that all currently permitted datacenters in Texas will be allowed to finish construction. This essentially covers all of the 2027 and likely well into 2028.
She is going to pass laws that ensure that all *newly* permitted datacenters going forward for delivery in 2028/2029 will pay taxes, commit to no increase in power prices for consumers, don’t impact water availability for people, be constructed with union labor and give “Texans a seat at the table.”
Sounds reasonable. She doesn’t want to lose a lot of good blue collar jobs and property taxes for small towns in rural Texas.
Datacenter commentary starts at 6:30 in this interview.
Also attaching her official campaign position from her website in the next post.
Sidenote: After watching a bunch of her speeches and interviews, I do think she is a political talent and probably has a future in national politics.
youtu.be/D-T6H3b8ZeE?is=mEWu…
1
242

















