CryptoLite_247

Elon Musk says “~90% of $SPCX revenue this year will be commercial” with government revenue falling below 5% in Q4.

Starlink has changed the mix so much that SpaceX really looks more like a connectivity platform rather than a government launch business.

$SPCX just launched USSF-385 on B1100 which is the 100th Falcon 9 Block 5 booster SpaceX has built.

The mission is carrying a classified Space Force payload but the thick fog at Vandenberg made the launch clip look incredible.

$AAPL just launched its most expensive iPhone ever with the $1,999 iPhone Duo and preorders reportedly sold out by 5:20 a.m.

Raising your top price 54% and still selling out that quickly is some serious pricing power.

$CRWD CEO warns “you’re going to see a lot of breaches happening in the coming months” following the discovery of critical PHP security flaws.

CrowdStrike’s massive threat data could become an even bigger advantage with AI now finding vulnerabilities humans missed for decades.

The recent pullback in $NBIS does not make me more bullish by itself.
What makes me more interested is that the business is finally giving us numbers to judge instead of only future capacity.

Q2 revenue reached $582.3M and adjusted EBITDA was $236.2M. At the same time, Nebius is still spending aggressively, raising new financing and building capacity that has to turn into real usage.

That is where I think the whole story sits now.
The demand side looks strong enough. $MSFT , META and other customers have already shown they are willing to commit serious money.
The harder part is what Nebius does with that demand.
If new MW comes online on time, utilization stays high and customer contracts keep supporting cheaper asset financing, the economics can improve very fast.

But if deployment slips, or too much capital gets tied up before revenue catches up, the same growth story becomes much less attractive for shareholders.
I like the company more than I like the idea of blindly chasing the stock.

For me, the next stage is simple, prove that every new dollar of capital creates more revenue and more value per share, not just a bigger data-center footprint.

$AAOI perhaps this week we test this level at 50..

HnS breakdown done, EMA55 and SMA50 broken, next stop has to be the trendline..

$META just acquired Sundial which is an AI analytics startup founded by former Meta executives.

And with Meta spending over $130B on AI this year knowing which models, compute clusters and ad products actually generate returns is becoming just as important as building them.

$AMZN CEO says AI inference is still “an order of magnitude too expensive” for widespread adoption and believes Amazon’s Trainium chips could help bring costs down.

And even at today’s prices AWS has already sold out its major capacity commitments for 2026 with most of 2027 and a good chunk of 2028 booked so imagine how much more demand cheaper inference could unlock.

$NBIS
CRO says "theres a lot of upside to pricing in the market today" and Nebius could sell out its entire 2027 AI compute capacity immediately at current prices.

Even more interesting is that customers are already trying to secure capacity into 2028 which is the same year so many AI bears expected supply to finally catch up with demand.

That's why Nebius is holding back capacity for shorter deals reportedly priced ~$50M per MW annually rather than locking everything in early with $SPCX, $CRWV and $IREN also seeing similar strength in GPU pricing.

$HOOD just secured approval to launch crypto trading in Dubai expanding Robinhood's international reach after entering the UK, Europe and Singapore.

That gives Robinhood access to another wealthy, crypto-friendly market as it looks to grow its customer base beyond the U.S.

$ASTS just signed agreements with three Brazilian carriers including Unifique, Brisanet and Iez to bring satellite broadband directly to regular smartphones using their 700 MHz spectrum.

The carriers would also fund the ground gateways allowing AST to expand through existing mobile networks while keeping its own spending focused on satellites.

$AAPL just surpassed $NVDA to become the world’s most valuable company.

$CRDO now controls ~86% of the AEC installed base as AI data centers rely on its lower-cost, power-efficient copper cables to connect GPUs and switches across large clusters.

That lead becomes super valuable as Credo designs its own SerDes chips and sells complete cables allowing it to capture more revenue from every connection as AI infrastructure scales.

And when a single faulty cable can disrupt thousands of GPUs then hyperscalers have little reason to move away from a supplier already proven at scale which makes Credo's market share incredibly difficult to displace.

SpaceX is reportedly targeting at least 80,000 cell sites across the U.S. as it moves closer to building a fourth nationwide wireless carrier, according to Bloomberg.

The company is reportedly in talks with major cellular infrastructure players, including Crown Castle (CCI). Securing that many sites would be significant, equivalent to roughly half of the 158,500 purpose-built cell towers operating in the U.S. last year.

The market reaction was sharp. Crown Castle jumped around 10% in extended trading, while American Tower also moved higher. Meanwhile, AT&T, Verizon and T-Mobile fell between 4.2% and 5% after hours as investors weighed the potential competitive threat.

The bigger story is how $SPCX could combine its satellite network with traditional wireless infrastructure to challenge established carriers. But securing tower access is only one step; building a nationwide service that can compete at scale is another.

$SKHYB HYNIX SCRAPS SOLIDIGM U.S. IPO AMID CAPITAL MARKET VOLATILITY & DUPLICATE-LISTING CONCERNS

The company also halted a planned pre-IPO funding round, per Bloomberg.

South Korean regulators had raised concerns that separately listing its flash-memory subsidiary could undermine existing shareholder value. The proposal also faced strong shareholder opposition.

Goldman Sachs and Morgan Stanley had reportedly been selected to lead the potential offering.

SK hynix says Solidigm will explore alternative financing options to expand production capacity.

$SPCX is in talks to secure at least 80,000 U.S. cell sites as Elon Musk looks to turn Starlink Mobile into a 4th nationwide wireless carrier.

Combined with its growing satellite network and recently acquired spectrum, those sites could give SpaceX the ground coverage needed to compete directly with $T, $VZ and $TMUS.

Friendly reminder for $NBIS investors

Revenue growth projections:
2025: $529.8M
2026E: $3.35B
2027E: $12.24B
2028E: $22.91B

$530M to $22B in just 3 years.

$CIFR.US Just a reminder that the wave 3 target for this stock is +50 a share

The -50% drawdowns are what gives you the +400% moves higher.

The fundamentals of $CIFR.US business did not deteriorate in the past 2 months, the dip is buyable and the AI trade will come back roaring post mid terms

$NVDA CEO Jensen Huang says “software licensing will become increasingly a major part of our company’s business.”

He said it could cost “about $2,000 per GPU per year” to run NVIDIA’s software for AI agents in these AI factories.

So, some napkin math at that price, before discounts:

500K licensed GPUs = $1B in annual revenue.
1 million licensed GPUs = $2B
5 million licensed GPUs = $10B

So, hypothetically, NVIDIA could earn billions a year from GPUs it has already sold by selling enterprise software and support to the companies using them.

Note: These are illustrative revenue scenarios using Jensen’s pricing example, not NVIDIA guidance or profit estimates.

$NVDA CEO Jensen Huang says a memory glut by 2029 is unlikely because "you can always use more of it" which is huge for $MU and $SKHY .

Even with China's CXMT ramping production I think the bigger threat remains standard DRAM rather than advanced HBM powering AI GPUs.