beyond mental models retweeted
BREAKING: A deal between Iran and Oman declaring the Strait of Hormuz "transit corridor" is expected to be announced today.
Iran has said talks with Oman are strictly bilateral and the agreement is not necessarily a guarantee that the waterway has become safe for vessels.
$TSM raising capex guidance due to N3 node capacity expansion and not due to bleeding edge is rather abnormal. Shows that there are a lot of GPUs/ASICs that need to migrate from N4/N5 node to N3 and repurposing N4/N5 tools to N3 is not along going to meet the demand.
1
266
$NVDA just raised its revenue guide by ~20%+ and stock is up barely 2.5%.
The fear of a peak just keeps hitting the stock.
Street estimates for 2026 and 2027 is ~$750bn and Jensen just said $1tn in revenues.
3
313
Buying $NVDA at current valuation has given 150% return with no instances of drawdown.
Only thing that seems different is that buyside has already modelled in $500bn of backlogs. So numbers are not moving higher and the fear we may hit peak in 2027/28 has caused this derating.
1
341
Fail to understand what $CSGP is trying to achieve here. There are concerns around Homes.com gowth + AI disruption risk and market share loss for Apartments.com. By combining these two business segments, you are not helping investors to gauge health of biz
1
1
441
$BKNG also acts as merchant of record. $GOOGL never wanted to act as merchant of record even when it launched its travel offerings in search. Same will be the case with other LLMs as it brings you under the preview of regulators.
AI will take the path of least resistance.
"Morgan Stanley upgraded Booking Holdings to Overweight from Equal Weight and lifted its price target to $5,500 from $6,150, arguing BKNG remains central to online travel even as agentic tools evolve. The firm contends OTAs will continue to own the customer relationship, data, and merchant-of-record economics, with early agentic workflows largely redirecting traffic back to OTA apps/websites rather than disintermediating them. MS expects the industry structure to remain closer to paid search dynamics, supporting BKNG’s direct mix and long-term margin profile."
1
12
1,340
Continue to believe that market is not giving any credit to data moats and just focusing on what can be automated.
Case in point, #Anthropic calling out $TRI as the only place to get Westlaw data.
If you have data, then you will be more insulated.
$MDB, $VEEVA, $TRI, $SNOW
Anthropic hosting a presentation on Enterprise Agents and messaging coming out of #Anthropic is that they are working with software providers. Not replacing them.
Not sure if the moves in software names stick.
$CRM, $MDB, $TEAM, $WDAY
2
461
We are in a market that shoots first and reasons later. 6% decline in $DASH is an indication of market raising odds of disruption risk.
@benthompson put out a compelling case on $DASH not being an AI loser.
An expensive stock can derate even if biz is going to be fine.
$GOOGL's TPU and AI advantage:
Gemini 3.1 is 50% more efficient relative to GPT 5.2 and Opus 4.6 and has 2x more reasoning capabilities than Gemini Pro.
Pretty amazing, many builders are +15% YTD. Were big beneficiaries of rotation out of tech and the hope that Trump Admin. will do something to get rates down.
Fundamentals are marginally better with lower new home inventory, but not so great that $LEN, $DHI and $PHM to rip 15%.
Housing channel checks and hard data releases so far in 2026 look increasingly like a rinse and repeat of early 2025 (which wasn’t great). Mortage rates are ~6% now vs ~7% then, so it’s concerning no real momentum we’re consistently seeing yet (new homes, resale, R&R, etc.).
677
Frontier labs are on a treadmill to build ever bigger AI models so that they don't get leapfrogged. Only way to do this is by raising billions. Billions can be raised by talking about trillions in revenues.
Not an AI sceptic, I am sceptic on rev numbers talked here.
Lipstick on Frontier AI "Pigs"
read the full piece here: mbi-deepdives.com/lipstick-o…
404

















