$APP
Talked to a former Applovin employee over the weekend, here’s some of his thoughts:
- Applovin’s business has been ads within mobile games. It is very profitable, 1B people per day play games, and $META or $GOOGL have not been able to break their moat. But, growth rates have slowed down from 70% to now 55% which is incredible, but not good enough for the street because they don’t see the TAM expanding yet.
- Mobile gaming opportunity is large but viewed as saturated and very high growth rates likely aren’t sustainable in this one category.
- Company has guided to keeping their growth at 30%+ but the decline in YoY growth rates have hurt sentiment. Applovin can change this by showing their Ecom expansion can increase the TAM but the company has failed to meaningfully show this which is why the stock has been stuck. The Q2 call didn’t really address the TAM expansion either which is why the stock has been falling so rapidly since then.
- He said management is exceptional and understands what they are doing, it’s not a team to bet against, but the problem is the street is not convinced their growth can expand beyond gaming yet. He himself is also not fully convinced and feels ready to sell a significant portion of stock if the Q after holiday season doesn’t show the TAM expanding beyond mobile games.
- He also said the short reports are pretty much just noise. Nothing ever serious has come from them and they aren’t rooted in anything real.
Stock is down 50% YTD…obviously looks like a bargain and I wonder how much lower the street can send it. Only major software large cap that hasn’t recovered…is anyone buying it here?
@mariojmourad Usually short term (1-2 weeks) if there's strong momentum or if there's a large selloff that's running out of steam, and I find 30 delta to be a good sweet spot for strikes.
Hit a milestone today - one year ago I had 0 shares of $INTC. Now I have 10000. All this from a simple strategy of buying Intel calls, using half the profits to buy shares, and using the other half to buy more Intel calls.
I feel like AI investing is simpler than what people expect.
Because $NVDA + Jensen literally tells everyone what's coming.
But somehow. Almost every. single. time. Markets dismiss it until it actually happens?
Be Nvidia in 2025: Buys up EML and laser capacity.
Markets dismissing it: "Photonics is a bubble and like quantum! ____ company is a scam with shady management"
1 year later: $LITE +678%, $AAOI +475.12%, $COHR +260.7%, $AXTI +3,843.9%.
_
Nvidia in 2026:
Buys up CW/EML capacity with LTAs. 800V shift. Extraordinary explicit about CPO shift. States Physical AI as the next theme.
Markets now: "CW players are meme stocks! 800v, CPO is not coming anytime soon, Humanoids are not profitable!"
Yeah... We'll see what happens in 2027.
I think I'm putting my money on Jensen/Nvidia as the leading indicator.
Just to be clear, I know absolutely nothing about neoclouds. This isn’t my field at all, but I am learning… quickly.
However, unlike semis where you need to know the impact of HBM4e on WFE supply chain and when yields will ramp. This seems to be much simpler to understand, you just need to own people who have POWER and aren’t completely inept.
BURRY:
- I SHORTED MORE $MU $ORCL & $NBIS TODAY
- NEBIUS IS WHAT THE TOP OF A BOOM LOOKS LIKE
- DEPRECIATION IS THE MAIN CONCERN, THE SAME THING THAT EXTENDED THE DOT COM BUBBLE
looks like he’s doubling down
Rumors circulating in Korea as to explain why SK Hynix has been dumping so hard since July
Translation
7:19 三沙 Yeouido Rumor Mill)
Right now, SK Hynix is supplying HBM3E to Nvidia dirt-cheap. Because it's set at an annual fixed price, Samsung is aggressively hiking prices on general-purpose memory and posting great earnings, whereas Hynix's quarterly profit growth remains relatively low for this reason.
Regarding HBM4, it is true that Nvidia delayed the Vera Rubin launch due to defects from SK Hynix.
SK Hynix's HBM4 price hike for Nvidia next year will likely be lower than competitors (competitors raising by over 80%, Hynix by 50%).
In return, SK Group secured priority supply rights for Nvidia chips for its data center business.
An SK Hynix sales employee vented about this over drinks.
Chae Min-sook, an analyst formerly with SK Hynix, heard this story and leaked it to Yeouido.
In July, SK Hynix stock plunged, while SK and SK Eternix hit the daily upper limit.
After uncovering the full story, SK Hynix banned Chae Min-sook from entering their premises.
Fund managers instructed other analysts to verify the facts.
Many say the pricing deal isn't even finished yet, and wild rumors continue to circulate.
so…the S&P is basically at ATHs
Trump has said 7 different times this week that he wants a deal with Iran
Bessent said a ceasefire could be announced over the weekend
Rate hike probabilities tanked by 20% today because of the labor market data
$PLTR is up 40% in a week…which means the software stocks are finally getting the respect they deserve including $MSFT $RDDT $SHOP
earnings continue to compound aggressively
the super high-beta semi names are consolidating like $MU $NBIS which isn’t bearish at all, especially when their growth continues to be massive
hyperscalers continue to spend on capex and cloud growth rates are showing the ROI
leverage also has been wiped out significantly
feels like we could be setting up for an end of year run IF hikes are out of the picture and earnings continue strong
the main bear case felt like oil/leverage/war/hikes
my only fear is OpenAI and Anthropic’s numbers not being meaningful enough to carry out their spending commitments but those will be shown over time
if those broader bear cases are not as relevant as inflation comes down…maybe we can continue with the momentum
Not good for memory
NVIDIA CONSIDERS LESS HIGH-BANDWIDTH MEMORY FOR RUBIN ULTRA GPU: INFORMATION
NVIDIA IS WEIGHING A RADICAL STEP TO DEAL WITH A SHORTAGE OF ADVANCED HIGH-BANDWIDTH MEMORY CHIPS: INFORMATION
A TON OF THINGS HAPPENED IN THE STOCK MARKET TODAY.
Here's a full recap:
1. Fed Governor Kevin Warsh would be open to a September rate hike if upcoming inflation data comes in hot and markets begin pricing in higher borrowing costs, according to people familiar with his thinking. The report adds to the renewed focus on whether the Fed may have to turn more hawkish if inflation pressure reaccelerates.
2. Michael Burry reportedly opened large short positions in Nebius $NBIS at $211.77 and Oracle $ORCL at $144.63, marking the first time he has taken a position in $NBIS. Burry framed the trade around AI infrastructure leverage and off-balance-sheet obligations, saying: “The fish have gorged themselves on off-balance sheet liabilities. Backstops. Uncommenced leases. Purchase commitments. The fish have gotten very fat, very large, easy to shoot. Also, so large that it shan’t be long before every last one keels over for lack of oxygen.”
3. Applied Optoelectronics $AAOI reported Q2’26 revenue of $191.9M, slightly ahead of estimates of $190M and up 86% YoY. Adjusted EPS came in at $0.06 versus $0.01 expected, while non-GAAP net income reached $5.5M, above estimates of $1.7M. For Q3, the company guided revenue to $255M–$290M versus $278M expected, with EPS of $0.11–$0.26 and non-GAAP gross margin of 29%–30.5%. Management said AAOI delivered record revenue for the 5th consecutive quarter and returned to non-GAAP profitability, while noting that demand is expected to outpace production capacity through mid-2027.
4. The June JOLTS report showed job openings easing by 178,000 to 7.359M, missing estimates of 7.454M, though openings remain above the January 2020 level of 7.124M. The openings-to-unemployed ratio edged up to 1.04, the highest since January 2025. Hiring improved, with hires rising 96,000 to 5.348M and the hiring rate increasing to 3.4%, led by health care and construction. Quits, a key measure of labor market confidence, rose 79,000 to 3.232M, the highest in nearly a year, while layoffs were essentially flat at 1.766M with the layoff rate unchanged at 1.1%.
5. The top 10 most active options today by contracts traded were $NVDA with 2.9M contracts, $SPCX with 1.7M contracts, $AAPL with 1.3M contracts, $TSLA with 1.2M contracts, $MU with 820K contracts, $MSFT with 690K contracts, $INTC with 586K contracts, $HTZ with 584K contracts, $PLTR with 477K contracts, and $ET with 473K contracts.
6. Private business investment in AI-related categories jumped $300B YoY in Q2 2026, up 25% to a record $1.5T annualized rate. The increase was led by spending on computers and peripheral equipment, followed by communication equipment, software, and data centers. Over the last two years, AI-related business investment has surged $500B, or 50%, with investment in computers and peripherals more than doubling. Direct AI investment is now estimated to account for 25%–33% of recent U.S. GDP growth.
7. Google $GOOGL is reportedly planning to raise money through a U.S. investment-grade bond offering. The company has started marketing notes in as many as 10 parts, with maturities ranging from 2 years to 40 years, according to Bloomberg. Initial price talk for the longest-dated tranche is around 1.55 percentage points above Treasuries.
8. Tether purchased 14 tonnes of gold in Q2 2026, bringing total holdings to a record 146 tonnes, now worth roughly $18.8B. The company previously bought 53 tonnes between Q3 2025 and Q1 2026, with its gold holdings more than doubling since Q1 2025 and their value nearly tripling over the same period. Tether is now the largest known private holder of gold outside of central banks and governments. In the first half of 2026, only 4 central banks bought more gold than Tether: Poland, Uzbekistan, China, and Kazakhstan.
9. Amazon $AMZN founder Jeff Bezos filed for the sale of 1,209,649 Amazon shares at $286.41 per share. The transaction brought in roughly $346.5M before taxes.
10. U.S. data center construction spending jumped 46% YoY in June to a record $68B annualized rate, the largest annual increase in 12 months. Since January 2024, spending has surged 158% and is now more than 3x higher than 2022 levels. At the same time, office construction spending has fallen by more than $25B since 2022 to roughly $43B, the lowest since 2016. Data center construction now exceeds office construction by $25B, the widest gap on record, a massive reversal from 2022 when office construction was $57B higher than data centers.
11. Trading activity in the Memory ETF $DRAM has surged to extreme levels, with daily volume reaching as high as roughly $8B, surpassing the $5B peak that ARKK hit during its 2020–2021 mania. Cumulative flows into $DRAM have climbed to around $27B, already above ARKK’s peak of roughly $18B, despite DRAM only launching in April 2026. The comparison is not perfect since DRAM tracks profitable memory chip companies rather than the mostly unprofitable growth names that dominated ARKK, and today’s rate environment is very different from 2020. Still, ARKK’s flows eventually peaked and reversed for years, while $DRAM is already down nearly 40% from its June high.
12. Alibaba $BABA reportedly plans to seek revenue sharing for the next version of its open-source Qwen AI model, while Moonshot is asking partners for up to a 30% revenue share for its Kimi K3 model, according to Reuters. The move suggests China’s leading AI labs are starting to push harder on monetization as open-source model adoption scales.
WALL STREET IS THE GREATEST SHOW ON EARTH.
And no one is addressing the elephant in the room of linear attention or the thermal and reliability issues with HBF on package.
When it comes to increasing reliability I haven’t learned anything new and no one on controller side is talking about codesign just about PCIe 6.0.
I’m so far ahead of the industry right now it’s scary. These guys are going to be doing this conference in the lobby of a best western if they keep this crap up.