Posted my first substack on Lemonade Insurance - breaking down the company, vision, and direction. Check it out: https://t.co/YiRBA3abOx $lmnd cc:@shai_wininger@daschreiber
New substack on the network bottleneck within the AI trade.
$LITE Lumentum is building the high bandwidth networking technology used in data centers by Google and Nvidia.
The world is currently facing a 25-30% supply gap for Indium Phosphide (InP) lasers (specifically EMLs). $LITE is vertically integrated in supplying InP lasers, owning their own fabs.
Let me know what you think!
https://t.co/nhIDLNwAUf
It’s pretty clear with codex 5.5 that incremental model improvements are trending model providers closer to commodities.
Value lies in the infra and customer experience.
$HOOD TAKE FLIGHT NY HIGHLIGHTS:
- New platinum credit credit card, higher annual fees, more benefits
- Robinhood Families: view a households portfolios from a single space
- Early Dividend Payouts, ~1 month before dividend date
The rate of execution is unmatched.
Slower dcas into new positions that are falling knives is key.
Learned this with $AMD last year as I kept buying the dip for 4 months. This becomes especially important when ur trying to convert a large cash position to a stock a position. Important to realize cash is also a valid position.
Has software gotten its final flush out?
Market seems to have moved focus to geopolitical tensions in the Middle East. Could see a final flush in software if $SPY and $QQQ correct 5% from here. That seems most likely as market behaves similar to the initial days of govt shutdown in October.
$ADBE $CRM $NOW
Feels like the reset Netflix faced in 2022.
Investors were convinced that user growth was capped and there was no new verticals to be explored.
Everyone is collectively convinced the company will go out of business with $1B in cash and ~0 debt. Even if DAUs and MAUs are growing at a slower clip, there’s still growth… and the pivot from management should reaccelerate that.
There are $52M DAUs, why haven’t those users just switched to AI or dropped of cus of google translate?
Learning is one of the most high friction activities. You could have the information readily available and you still wouldn’t hit the books. You could learn stuff with chat or Gemini, but you probably won’t stick to it.
Understanding what makes users stay and gamification is the moat. They have all the data and A/B testing on what works.
There are 52M DAUs and that’s growing, hard for them to fall off the grid completely imo.
New substack on the network bottleneck within the AI trade.
$LITE Lumentum is building the high bandwidth networking technology used in data centers by Google and Nvidia.
The world is currently facing a 25-30% supply gap for Indium Phosphide (InP) lasers (specifically EMLs). $LITE is vertically integrated in supplying InP lasers, owning their own fabs.
Let me know what you think!
https://t.co/nhIDLNwAUf
The $META deal with $AMD is a great hedge for the capex spending they’re planning.
Suddenly all that capex they’re spending doesn’t seem so bad when u factor in that they will be sharing the fruits of the spending with the benefiting party…
Not only does the spending help them fill their growing AI workload demands, but they added a new vector of growth: AMD equity ownership.
The same way startups have always disrupted legacy businesses.
It’s not going to be an overnight process, but if they’re able to offer restaurants better margins/deals, what’s the harm in restaurants expanding their delivery network presence?
While I don’t think restaurants are gonna vibe code their own agents and delivery services, there’s definitely more room for growing the N optimal providers in the space. Even then, how many folks will stick to it and have the same level of performance and reliability as DoorDash? not as many as people are assuming, but the competitive landscape has definitely shifted and the second and third order impacts will be interesting to see unfold
“And who wins then? The provider of these optimal deals”.
There can now be N different providers, whilst today it’s concentrated to a few companies with network effect.
An interesting correlation is the Indian delivery market - the cost of app development is so low already that there are new competing delivery startups undercutting each others margins every 3 months. DoorDash is just an example that’s more prone to facing such competition which has previously been blocked by barriers to entry.
Creativity will become increasingly scarce in a world that’s trained on datasets leading to similar content generation.
In this scenario, consumers will seek uniqueness that a model can’t quite capture.
How is creative uniqueness consumed? Art, pictures, and videos.
$ADBE
Assume the scenario where you have agents that are able to hit N delivery services that are reliable and trusted within an agentic network. Same way delivery drivers have an agent which finds optimal deals from them across different apps.
Disruption won’t happen overnight, but a thought experiment to what could be possible. At the end of the day, the responsibility is to connect buyers and sellers and there can now be infinitely more sellers.