Well, I did give you a heads up. China EV booming in short term. Give a ❤️ and follow @amootpoint $NIO, $XPEV, $TSLA, $GOOG

Very well said

Very interesting interview with a Former $NBIS employee explaining the relationship between $NVDA and the neoclouds and the views on $NVDA's new revenue-share partnership: 1. In his view, the recently announced revenue-share model of $NVDA with neoclouds like SharonAI and Firmus was more a result of Firmus, according to him, having financial trouble and SharonAI having problems regarding customers and investors. 2. He thinks the neoclouds hate this new model by $NVDA, as $NVDA has all the leverage over them; at the same time, they are paying a higher price for GPUs because they don't get the high discounts on buying in bulk. 3. He thinks that $NBIS and $CRWV would do these types of deals if they needed capital. He thinks $CRWV is in a worse position, as they have a significant financial gap given their big deals. 4. He mentions the relationship between $CRWV and $NVDA getting worse as $CRWV was upset that $NVDA didn't deliver the GPUs that $CRWV needed. According to him, they even threatened to go with another ASIC provider. At the same time, $NVDA is, according to him, upset with $CRWV as they are not respecting the roadmap that $NVDA wants. $CRWV wants to go after big clients for large deals, while $NVDA wants $CRWV to adopt a more sophisticated enterprise approach to provide more professional services. 5. According to him, $META has around 400k GPUS (H100, H200, and even some Blackwells) they want to sell as compute on the market, and that is going to put extreme pressure on the neoclouds, especially the lower-end ones that are deploying the older GPUs. found on @AlphaSenseInc

$MSFT and $META refuse to participate in this market rally. … but why?

Ouch @Srasgon. $SMCI

The earnings report I was waiting for all day: it wasn’t Amazon. It is Atlassian: which we don’t own but this is a really clean read on the state of software. They touch everything. • Developers • Non technical teams • Simple daily workflows • Deep, complex systems across the enterprise If software were “dead,” Atlassian would be feeling it first. The questions I had going in: • Is AI creating pricing pressure? • Are seats getting consolidated? • Is net new demand slowing? Here’s what actually showed up. Pricing pressure? Revenue +23% Y/Y. Cloud +26% Y/Y. Expanding margins. That’s higher ARPU and upsell, not discounting. Seat pressure? Cloud NRR at 120%, improving again. Customers using AI are adding users faster, not cutting them. Net new demand? RPO +44% Y/Y and $1M+ ACV deals nearly 2x Y/Y. Enterprises are committing more, not less. Then guidance. ~22% growth in FY26 with margin expansion. This doesn’t look like AI eating software. It looks like AI expanding usage inside software. More seats. More apps. More upgrades. AI means more software on top of these software stacks. $TEAM $IGV $AMZN $CRM $ZM

SaaS is not going anywhere. Nobody, and I mean absolutely nobody, wants to code and support every piece of software they use. This is a total and complete waste of time. If the market already built what you want and it's good, you are wasting time and money rebuilding it for nothing. And nobody wants to do every job either. You don't want to vibe code the damn AI accounting software and then support it and verify its output. You want the accountants using the AI accounting software. People are really losing their minds in the distortion field right now.

BREAKING: Talks over a $100 billion deal between OpenAI and Nvidia, $NVDA, have stalled. Nvidia's Jensen Huang has reportedly "privately criticized" OpenAI's business strategy.

$MU “The bottleneck is shifting from compute to context management. To scale, storage can no longer be an afterthought. As inference scales to giga-scale, context becomes a first-class data type.” -NVIDIA Senior Director of AI Infrastructure, Dion Harris

$MU The average price of DDR4 8GB DRAM surged from $1.40 in January to $9.30 in December (2025) Profit margin is expected to reach 70%, and DDR5 may even surpass the margin of HBM3E. Wild

$MU #Micron still is on track on my previous PT of ~82, checking back at the trend line. #Sept is a great tailwind for it to reach there. Patiently waiting for the trendline hit to add .... $NVDA $MU $SMCI $DELL $AMD $INTC

I like this thread below from @mccdmatthew which compare's $PYPL to $APPL. I was heavily invested in $AAPL back in 2013 and the stock traded at 10X Earnings. I know many of you won't believe this but that's where $APPL used to trade because Tim Cook was not Steve Jobs and it was just a dumb and slow growing phone company. Then they ignited the App Store growth and investors began to understand $AAPL was more than just a phone company. I see similarities with $PYPL today. Similar valuation and also the loudest online are now saying the company is "dead" while ignoring long-term growth opportunities such as Agentic Commerce, Ads, BNPL, and Venmo. Like $AAPL, $PYPL is also buying back a ton of stock here. Let's see how this plays out. Not investment advice, but I feel like I've seen this movie before. Good luck.

CoreWeave’s Moment of Truth: Building the Future on Borrowed Time What Just Happened CoreWeave’s latest earnings were impressive on the surface, revenue more than doubled year over year to $1.36 billion, and its AI computing backlog ballooned past $55 billion. The company is still growing faster than almost anyone else in the space. But beneath the headline beat, the story gets more complicated. A delay in one of its data centers forced them to trim full year guidance, and despite the huge growth, they’re still posting losses and carrying a mountain of debt. The market noticed. Shares dropped about 6% in extended trading after the report, closing at $105.61 down from intraday highs but still up 164% since the March IPO . The message from investors was clear: growth alone isn’t enough when guidance weakens. Under the Hood The company’s core business is clearly booming, but its cost structure tells a different story. Interest expenses alone hit over $300 million just last quarter. That’s the cost of financing one of the fastest infrastructure expansions in the industry. CoreWeave is spending billions upfront, $3billion in capex last quarter and funding most of it through high interest loans. They’re basically racing to build out capacity before the next wave of AI demand hits, hoping that scale will make the math work later. The GPU Reality Check Here’s where it gets tricky. CoreWeave extended the “useful life” of its GPUs from four to six years, spreading costs over a longer period. On paper, that boosts profits. In reality, the economics are moving the other way. H100 rentals that peaked around $8–$16 an hour in 2023 are now down to roughly $2–$4, with some spot prices even lower, a drop of nearly 70–80%. That tells you the market has already repriced the value of that hardware. The chips still work, but their economic yield falls fast as new generations like Nvidia’s H200, B200, and now the Blackwell GB300 come online. Stretching depreciation schedules makes earnings look cleaner, but it also widens the gap between accounting and reality. Why It Matters This is the same dynamic Michael Burry just warned about with companies extending the life of hardware to make profits look smoother. The problem is that AI infrastructure runs on an arms race cycle, not a steady one. Every new GPU leap instantly devalues the old one. If pricing continues to slide while book values assume a slow fade, CoreWeave risks being caught with overstated assets and underreported costs. Eventually, that gap shows up through impairments, tighter margins, or both. The Bigger Picture At a macro level, CoreWeave sits at the intersection of U.S. industrial policy and the AI boom. The U.S. government wants domestic control over compute power, both for national security and technological leadership and CoreWeave fits neatly into that agenda. That strategic positioning gives them tailwinds that pure commercial players don’t have. But it also ties their fate to policy. Power allocation, environmental limits, and export rules could all start shaping how and where they grow. Where This Heads Next CoreWeave’s growth is real, but it’s being built on expensive debt, falling chip economics, and accounting optimism. The company can make this work if it pivots fast, moving into newer GPU cycles, securing more long term, take or pay contracts, and managing its borrowing costs. But until they show they can turn that massive revenue backlog into durable free cash flow, they’ll remain caught between story and substance. CoreWeave is building real infrastructure in a strategic place at a critical moment. But the next chapter will depend on how long the market tolerates growth first, cash later in an environment where capital isn’t free and technology refreshes faster than balance sheets can keep up.

$TLT. Pivotal move for TLT and rates today. Everything pointing towards $96 as the first stop. Follow @amootpoint, if you like. $XBI $IWM $IJR $ARKG $TAN $ENPH $SEDG