Offshore
Video
The Few Bets That Matter
OpenAI with another video of poor taste...
Begging the markets not to let the AI trade fade and using the famous Chinese competition to do so...
This isn't what tech is about.
Markets reward execution and cash.
So execute & generate cash.
https://t.co/2ImqKWYyq2
tweet
OpenAI with another video of poor taste...
Begging the markets not to let the AI trade fade and using the famous Chinese competition to do so...
This isn't what tech is about.
Markets reward execution and cash.
So execute & generate cash.
https://t.co/2ImqKWYyq2
tweet
Offshore
Photo
The Few Bets That Matter
$PATH bull case rests on one metric only: net new ARR.
Q3 delivered the first sequential growth.
Q4 guidance points to acceleration.
That’s it.
I break down the thesis and how to capitalize on it in my latest article. Link's in bio. https://t.co/yFT9gOq9Wo
tweet
$PATH bull case rests on one metric only: net new ARR.
Q3 delivered the first sequential growth.
Q4 guidance points to acceleration.
That’s it.
I break down the thesis and how to capitalize on it in my latest article. Link's in bio. https://t.co/yFT9gOq9Wo
tweet
Offshore
Photo
EndGame Macro
This feels like market mechanics doing what they do when liquidity thins out. You get a fast break lower once price slips through obvious intraday levels, stops get tripped, and systematic sellers lean in. That selling feeds on itself for a bit, not because everyone suddenly got bearish, but because that’s how these systems are wired to react.
Then you hit a level where selling pressure runs out, shorts start covering, and the same machines that sold on the way down flip into buy mode as price stabilizes and reclaims key averages. The snapback is a reflex. This is less about fundamentals changing in an hour and more about how a market dominated by algorithms, options hedging, and shallow liquidity can swing hundreds of billions on autopilot before most humans even react.
tweet
This feels like market mechanics doing what they do when liquidity thins out. You get a fast break lower once price slips through obvious intraday levels, stops get tripped, and systematic sellers lean in. That selling feeds on itself for a bit, not because everyone suddenly got bearish, but because that’s how these systems are wired to react.
Then you hit a level where selling pressure runs out, shorts start covering, and the same machines that sold on the way down flip into buy mode as price stabilizes and reclaims key averages. The snapback is a reflex. This is less about fundamentals changing in an hour and more about how a market dominated by algorithms, options hedging, and shallow liquidity can swing hundreds of billions on autopilot before most humans even react.
We are seeing MASSIVE swings in markets right now:
Between 11:55 AM ET and 12:25 PM ET, the S&P 500 erased -$450 billion in market cap.
23 minutes later and $320 billion in market cap has been added back.
That's a $720 BILLION swing in market cap in under 1 hour.
Capitalize on volatility. tweet
Offshore
Photo
The Few Bets That Matter
$NBIS once again focusing on solving client friction and delivering the best experience, combined with the highest compute quality at the best pricing.
$NBIS introduced two new tools designed to track reserved & available capacity so customers aren't paying for unused compute.
🔹Capacity Blocks - Full visibility into when reserved GPUs start, expansions and how capacity is allocated across projects - eliminating paid-but-unused compute.
🔹Capacity Dashboard & Public API - Real-time visibility into onGPU availability across regions so clients can be alerted directly instead of blin testing compute availability.
On its own, this might seem whatever. But combined with $NBIS other advantages, a constant focus on usage quality makes the difference in a crowded market.
tweet
$NBIS once again focusing on solving client friction and delivering the best experience, combined with the highest compute quality at the best pricing.
$NBIS introduced two new tools designed to track reserved & available capacity so customers aren't paying for unused compute.
🔹Capacity Blocks - Full visibility into when reserved GPUs start, expansions and how capacity is allocated across projects - eliminating paid-but-unused compute.
🔹Capacity Dashboard & Public API - Real-time visibility into onGPU availability across regions so clients can be alerted directly instead of blin testing compute availability.
On its own, this might seem whatever. But combined with $NBIS other advantages, a constant focus on usage quality makes the difference in a crowded market.
tweet
Offshore
Photo
The Few Bets That Matter
$ORCL has ~$110B of debt.
~$25B to be paid over the next three years.
Management expects ~$26B in operating cash flow in FY26 - up +19% YoY, driven by AI cloud ramp.
Yes, FCF will stay negative due to heavy capex. Possibly until FY28.
But project OpCF out 2–3 years, knowing that their FY26 guidance is already enough to cover their next 3Y of repayments.
tweet
$ORCL has ~$110B of debt.
~$25B to be paid over the next three years.
Management expects ~$26B in operating cash flow in FY26 - up +19% YoY, driven by AI cloud ramp.
Yes, FCF will stay negative due to heavy capex. Possibly until FY28.
But project OpCF out 2–3 years, knowing that their FY26 guidance is already enough to cover their next 3Y of repayments.
tweet
Offshore
Video
The Few Bets That Matter
This is smart. It's not how the tool will be used. But it is smart.
Although we knew Vlad was smart.
$HOOD
tweet
This is smart. It's not how the tool will be used. But it is smart.
Although we knew Vlad was smart.
$HOOD
“Some people have already started to realize that using prediction markets can be cheaper than conventional fire, flood, and hurricane insurance.”
- Vlad, Robinhood https://t.co/mSSDvJqUVs tweet
Offshore
Photo
Dimitry Nakhla | Babylon Capital®
$MELI trades at an attractive PEG
NTM P/E ~39x
2026 EPS➡️ $59.74 (+48%)
2027 EPS ➡️ $83.48 (+39%)
2028 EPS ➡️ $115.61 (+38%)
CAGR at various multiples assuming 2028 EPS of $100 (~15% below est)
34x | 20%
32x | 17%
30x | 15%
28x | 12% https://t.co/dGLuxGgjwO
tweet
$MELI trades at an attractive PEG
NTM P/E ~39x
2026 EPS➡️ $59.74 (+48%)
2027 EPS ➡️ $83.48 (+39%)
2028 EPS ➡️ $115.61 (+38%)
CAGR at various multiples assuming 2028 EPS of $100 (~15% below est)
34x | 20%
32x | 17%
30x | 15%
28x | 12% https://t.co/dGLuxGgjwO
tweet
Offshore
Photo
App Economy Insights
We're excited to partner with @range_finance.
Range is your ultimate centralized wealth hub.
📈 Investment strategy
⚖️ Tax optimization
🏖️ Retirement planning
📊 Cash flow & budgeting
🏠 Real estate
View all your financial data at a glance. https://t.co/SZ6WuWRHgM
tweet
We're excited to partner with @range_finance.
Range is your ultimate centralized wealth hub.
📈 Investment strategy
⚖️ Tax optimization
🏖️ Retirement planning
📊 Cash flow & budgeting
🏠 Real estate
View all your financial data at a glance. https://t.co/SZ6WuWRHgM
tweet
EndGame Macro
In my opinion it wasn’t some accident in the data it was useful timing, and it lines up with what was already happening underneath. Inflation has likely cooled more than the headlines suggest, especially in goods and other rate sensitive areas. The sharp drop in CPI shelter looks dramatic because of a data quirk where October prices were carried forward from September during the shutdown, effectively imputing near zero inflation for rent and OER that month. That compresses what would normally be a slow, lagged adjustment into a tighter window, making the cooling look abrupt. It doesn’t change the trend it just reveals it sooner. Private data like Zillow had already been pointing this way; CPI is just catching up, mechanically.
That matters for policy. Without this quirk, inflation would still be cooling in reality, but the official data would have looked stickier for longer because shelter lags by design. The carry forward accelerates recognition, not disinflation itself. And that recognition gives the Fed a defensible narrative to keep cutting or cut faster without appearing reactive to economic stress. They’ve already decided to ease; the pivot now is about how fast, how far, and with what tools. Shelter finally rolling in CPI provides the public justification to shift from normalizing policy to managing risk, moving earlier rather than waiting for a labor or credit problem to force their hand, all while protecting inflation credibility.
tweet
In my opinion it wasn’t some accident in the data it was useful timing, and it lines up with what was already happening underneath. Inflation has likely cooled more than the headlines suggest, especially in goods and other rate sensitive areas. The sharp drop in CPI shelter looks dramatic because of a data quirk where October prices were carried forward from September during the shutdown, effectively imputing near zero inflation for rent and OER that month. That compresses what would normally be a slow, lagged adjustment into a tighter window, making the cooling look abrupt. It doesn’t change the trend it just reveals it sooner. Private data like Zillow had already been pointing this way; CPI is just catching up, mechanically.
That matters for policy. Without this quirk, inflation would still be cooling in reality, but the official data would have looked stickier for longer because shelter lags by design. The carry forward accelerates recognition, not disinflation itself. And that recognition gives the Fed a defensible narrative to keep cutting or cut faster without appearing reactive to economic stress. They’ve already decided to ease; the pivot now is about how fast, how far, and with what tools. Shelter finally rolling in CPI provides the public justification to shift from normalizing policy to managing risk, moving earlier rather than waiting for a labor or credit problem to force their hand, all while protecting inflation credibility.
From @fcastofthemonth
"This is totally inexcusable. The BLS just assumed rent/OER were zero for October. I am sure they have a good technical explanation for this, but the only way you get a two-month average for rent of 0.06% and OER at 0.135% is assuming October was zero. There is just no world in which this was a good idea, but here we are." tweet
X (formerly Twitter)
Nick Timiraos (@NickTimiraos) on X
From @fcastofthemonth
"This is totally inexcusable. The BLS just assumed rent/OER were zero for October. I am sure they have a good technical explanation for this, but the only way you get a two-month average for rent of 0.06% and OER at 0.135% is assuming…
"This is totally inexcusable. The BLS just assumed rent/OER were zero for October. I am sure they have a good technical explanation for this, but the only way you get a two-month average for rent of 0.06% and OER at 0.135% is assuming…