Offshore
Photo
App Economy Insights
🎉 Welcome to 2026!
Did you keep track during the holidays?
📈 Sandisk Absolutely Crushed 2025
🧠 Meta Bets on AI Startup Manus
🤖 NVIDIA + Groq: The $20B Deal
🔎 Gemini’s Very Good Year
$SNDK $META $NVDA $GOOG
https://t.co/WTFLZRkimA
tweet
🎉 Welcome to 2026!
Did you keep track during the holidays?
📈 Sandisk Absolutely Crushed 2025
🧠 Meta Bets on AI Startup Manus
🤖 NVIDIA + Groq: The $20B Deal
🔎 Gemini’s Very Good Year
$SNDK $META $NVDA $GOOG
https://t.co/WTFLZRkimA
tweet
Offshore
Photo
Quiver Quantitative
Wow.
We've been reporting on Representative Tim Moore's recent purchases of Intel stock.
Moore sits on the House Subcommittee on AI.
$INTC has now risen 103% since his August 1st purchase. https://t.co/LyTnMs1nXS
tweet
Wow.
We've been reporting on Representative Tim Moore's recent purchases of Intel stock.
Moore sits on the House Subcommittee on AI.
$INTC has now risen 103% since his August 1st purchase. https://t.co/LyTnMs1nXS
tweet
Offshore
Photo
The Few Bets That Matter
Remember the $NVDA “inventory bear case”? The one built on 2 digit IQ that generated millions of views?
I broke down why it was deadly wrong in my repost.
Turns out the so-called “inventory” included ~700,000 H200 units; over $15B of produced but unsold hardware due to export curbs.
Not weak demand. Not forged accounts. Not unpaid bills.
That narrative created an so much fear, pushed by people who look at numbers without context, and did slow down the market a little bit.
Fast forward weeks later, to facts, as they always take time to surface. A recent article made the situation clear.
“ $NVDA is scrambling to meet strong demand for its H200 artificial intelligence chips from Chinese technology companies and has approached contract manufacturer $TSM to ramp up production.
Chinese technology companies have placed orders for more than 2 million H200 chips for 2026, while Nvidia currently holds just 700,000 units in stock.”
$NVDA sees a path to ~$500B in sales in the West alone. With renewed access to China, where demand from players like $BABA is massive, the supply chain will be stressed once again.
You’re just not bullish enough. The AI trade isn't over.
https://t.co/mm3A4EB6ze
🚨BREAKING: The $610 Billion AI Ponzi Scheme Is Not A Ponzi Scheme
Here’s why $NVDA isn’t the disaster the algorithms - and the bears, want you to think it is. Far from it.
Shanaka’s argument claims that Nvidia’s rising inventory, receivables, and DSO suggest demand is slowing and the company is pushing more product than customers can absorb, in terms of need and payment.
In brief: no more demand nor cash to pay for their GPUs.
1. Rising Inventory ≠ Red Flag
Shanaka says rising inventory is evidence of weak demand, but ignores $NVDA pricing - and many other factors we'll talk about.
When unit prices double or triple, the same volume of hardware shows up as a larger dollar value in inventories.
You'll have more bananas for $1M that airplanes, right? Just like you'll have more H100 than GB200.
When we normalize inventory by revenue - or by units shipped, the trend is stable, suggesting this is a pricing effect, not a demand problem and rising inventory in volume.
This can also be illustrated with accounts receivable per revenue, which make the same point: when product prices increase, dollar-denominated metrics rise, so metrics taken individually may look bad but within context, the story looks normal.
That being said, many could point that even then, inventory is rising. To which we need to add context, something algorythms are incapable of.
2. Higher DSO & Supply Chain Constraints
DSO - which represents the time before being paid, rising slightly is consistent with real-world constraints.
$NVDA doesn’t just ship GPUs anymore; they ship racks, custom configurations, integrated systems… These use third-party components, which require more coordination, harder logistics, and can temporarily increase time before revenue recognition and therefore inventory.
Add to this the fact that foundries, as proven many times these quarters during $TSM & co earnings, run at full capacity, and you get even more delays.
More customization + constrained supply chains = longer installation cycles before revenue can be recognized and rising inventories until then.
This is an operational bottleneck, not a credit problem.
A move from 46 to 53 days is marginal especially considering this value has been roughly stable for three quarters.
3. Circular Economy
As for the claims about a circular economy and the same dollars being used across multiple companies, I have no counters but this: circular economies are normal, that’s how economies work.
It only becomes a problem if AI services do not generate enough cash to honor commitments.
Because that’s what those are: commitments, not booked revenues. If those commitments can be honored, then what is the problem?
4. Algorithms Don’t Understand Conte[...]
Remember the $NVDA “inventory bear case”? The one built on 2 digit IQ that generated millions of views?
I broke down why it was deadly wrong in my repost.
Turns out the so-called “inventory” included ~700,000 H200 units; over $15B of produced but unsold hardware due to export curbs.
Not weak demand. Not forged accounts. Not unpaid bills.
That narrative created an so much fear, pushed by people who look at numbers without context, and did slow down the market a little bit.
Fast forward weeks later, to facts, as they always take time to surface. A recent article made the situation clear.
“ $NVDA is scrambling to meet strong demand for its H200 artificial intelligence chips from Chinese technology companies and has approached contract manufacturer $TSM to ramp up production.
Chinese technology companies have placed orders for more than 2 million H200 chips for 2026, while Nvidia currently holds just 700,000 units in stock.”
$NVDA sees a path to ~$500B in sales in the West alone. With renewed access to China, where demand from players like $BABA is massive, the supply chain will be stressed once again.
You’re just not bullish enough. The AI trade isn't over.
https://t.co/mm3A4EB6ze
🚨BREAKING: The $610 Billion AI Ponzi Scheme Is Not A Ponzi Scheme
Here’s why $NVDA isn’t the disaster the algorithms - and the bears, want you to think it is. Far from it.
Shanaka’s argument claims that Nvidia’s rising inventory, receivables, and DSO suggest demand is slowing and the company is pushing more product than customers can absorb, in terms of need and payment.
In brief: no more demand nor cash to pay for their GPUs.
1. Rising Inventory ≠ Red Flag
Shanaka says rising inventory is evidence of weak demand, but ignores $NVDA pricing - and many other factors we'll talk about.
When unit prices double or triple, the same volume of hardware shows up as a larger dollar value in inventories.
You'll have more bananas for $1M that airplanes, right? Just like you'll have more H100 than GB200.
When we normalize inventory by revenue - or by units shipped, the trend is stable, suggesting this is a pricing effect, not a demand problem and rising inventory in volume.
This can also be illustrated with accounts receivable per revenue, which make the same point: when product prices increase, dollar-denominated metrics rise, so metrics taken individually may look bad but within context, the story looks normal.
That being said, many could point that even then, inventory is rising. To which we need to add context, something algorythms are incapable of.
2. Higher DSO & Supply Chain Constraints
DSO - which represents the time before being paid, rising slightly is consistent with real-world constraints.
$NVDA doesn’t just ship GPUs anymore; they ship racks, custom configurations, integrated systems… These use third-party components, which require more coordination, harder logistics, and can temporarily increase time before revenue recognition and therefore inventory.
Add to this the fact that foundries, as proven many times these quarters during $TSM & co earnings, run at full capacity, and you get even more delays.
More customization + constrained supply chains = longer installation cycles before revenue can be recognized and rising inventories until then.
This is an operational bottleneck, not a credit problem.
A move from 46 to 53 days is marginal especially considering this value has been roughly stable for three quarters.
3. Circular Economy
As for the claims about a circular economy and the same dollars being used across multiple companies, I have no counters but this: circular economies are normal, that’s how economies work.
It only becomes a problem if AI services do not generate enough cash to honor commitments.
Because that’s what those are: commitments, not booked revenues. If those commitments can be honored, then what is the problem?
4. Algorithms Don’t Understand Conte[...]
Offshore
The Few Bets That Matter Remember the $NVDA “inventory bear case”? The one built on 2 digit IQ that generated millions of views? I broke down why it was deadly wrong in my repost. Turns out the so-called “inventory” included ~700,000 H200 units; over $15B…
xt
Shanaka claims that this was thankfully found by algorithm - and I can agree with him based on the market's behaviour and violence. But he forgets that algorythm are built to find fraud in 99% of cases.
But $NVDA is the 1%.
When revenue grows 60–80% YoY, it's normal for inventories, receivables, and payables to grow at least comparably in dollar terms. Maybe even slightly higher when added real-world constraints.
What matters is whether these metrics grow disproportionately relative to revenue.
And once normalized, $NVDA ratios are stable, which is consistent with a rapid ongoing expansion, not accounting games or demand collapse.
That being said, everything isn’t necessarily perfect. But again: algorithms are configured to gauge 99% of the market, so of course the 1% will raise red flags.
Add some organic grey cells, context and reality, and the picture is very different, even if the stock continues to fall.
The market is about emotions, not rationality. And X is great at sharing emotions, less for rationality.
Conclusion.
I might be proven wrong in time and $NVDA might be an accounting fraud. I personally continue to believe in the AI revolution, have my own concerns about the circular economy but did not find any indications that AI won't yield cash flow and that commitments can't be honored as of today.
I continue to be bullish. And shared all my moves and reasoning with subscribers yesterday.
The future is bright for those with a system. - The Few Bets That Matter tweet
Shanaka claims that this was thankfully found by algorithm - and I can agree with him based on the market's behaviour and violence. But he forgets that algorythm are built to find fraud in 99% of cases.
But $NVDA is the 1%.
When revenue grows 60–80% YoY, it's normal for inventories, receivables, and payables to grow at least comparably in dollar terms. Maybe even slightly higher when added real-world constraints.
What matters is whether these metrics grow disproportionately relative to revenue.
And once normalized, $NVDA ratios are stable, which is consistent with a rapid ongoing expansion, not accounting games or demand collapse.
That being said, everything isn’t necessarily perfect. But again: algorithms are configured to gauge 99% of the market, so of course the 1% will raise red flags.
Add some organic grey cells, context and reality, and the picture is very different, even if the stock continues to fall.
The market is about emotions, not rationality. And X is great at sharing emotions, less for rationality.
Conclusion.
I might be proven wrong in time and $NVDA might be an accounting fraud. I personally continue to believe in the AI revolution, have my own concerns about the circular economy but did not find any indications that AI won't yield cash flow and that commitments can't be honored as of today.
I continue to be bullish. And shared all my moves and reasoning with subscribers yesterday.
The future is bright for those with a system. - The Few Bets That Matter tweet
Offshore
Photo
Fiscal.ai
Constellation Software is still in its largest drawdown ever.
Current EV/FCF: 21.3x
Why wouldn't this beat the market over the next 5 years?
$CSU https://t.co/XxMdVkbe6F
tweet
Constellation Software is still in its largest drawdown ever.
Current EV/FCF: 21.3x
Why wouldn't this beat the market over the next 5 years?
$CSU https://t.co/XxMdVkbe6F
tweet
App Economy Insights
🎉 Welcome to 2026!
Did you keep track during the holidays?
📈 Sandisk absolutely crushed 2025
🧠 Meta bets on AI startup Manus
🤖 NVIDIA + Groq: The $20B deal
🔎 Gemini’s very good year
$SNDK $META $NVDA $GOOG
https://t.co/WTFLZRkimA
tweet
🎉 Welcome to 2026!
Did you keep track during the holidays?
📈 Sandisk absolutely crushed 2025
🧠 Meta bets on AI startup Manus
🤖 NVIDIA + Groq: The $20B deal
🔎 Gemini’s very good year
$SNDK $META $NVDA $GOOG
https://t.co/WTFLZRkimA
tweet
Appeconomyinsights
📈 The Top Stock in 2025
The S&P 500’s best performer wasn’t on your bingo card
Offshore
Video
memenodes
the real-time training https://t.co/QMfikXDbii
tweet
the real-time training https://t.co/QMfikXDbii
USERS AREN’T JUST LIKING GROK, THEY’RE TRAINING IT IN REAL TIME
What’s spreading right now isn’t just praise for Grok.
It’s people actively helping it get better.
Grok isn’t a sealed black box.
It’s designed to learn directly from users.
Every thumbs up tells it what worked.
Every thumbs down sharpens what didn’t.
People get this, and they’re engaging with it like a tool they actually want to improve, not just consume.
Why?
Because Grok feels worth the effort.
The answers are direct.
No corporate hedging.
No censor-heavy tone.
Take two two seconds to hit feedback because they know it counts.
Source: @grok tweet
Offshore
Photo
The Few Bets That Matter
$LYFT is the equivalent of price action porn.
Don't follow the company and wouldn't buy against $UBER. But every stock has a price.
And this one is going higher. https://t.co/uBcds1IDPN
tweet
$LYFT is the equivalent of price action porn.
Don't follow the company and wouldn't buy against $UBER. But every stock has a price.
And this one is going higher. https://t.co/uBcds1IDPN
tweet