The same algorithm that drove an 85% cost cut on Falcon is now pointed at compute.
Starship V3 puts 100 tons in orbit. V4 doubles it. Colossus 2 is already running Anthropic workloads on the ground.
The orbital layer is built from parts that already exist.
*** Thank you @DavidCarbutt_ for the edit.
Most people think AI gives you the answers. An ex-Citadel PM uses it to tell the humans what to go find out:
@FundamentEdge, ex-Citadel and D.E. Shaw PM:
"One of my favorite questions with AI is: create me a research plan. Who should I talk to? What work should I do to validate this hypothesis?"
"The AI systems I'm building become an orchestration tool of the humans."
"These are the 12 analyses we should do. These are the eight people we should speak with. AI can't go have a soul-to-soul conversation — so go and have those conversations."
"In a way, the LLM can orchestrate the human."
"These are chatbots trained on a web scrape. Best case, they'll give you a consensus view — and consensus will lead to a sub-mediocre result in investing."
"But it can give you the range of outcomes. It can stimulate ideas on which path to follow."
"Which nodes on that probability tree can I support with research to build investible conviction? That's the piece I don't think AI can do — and arguably never will."
No Merger For Tesla and SpaceX
Then how will they fund Terafab?
The secret has been hidden in plain sight the whole time.
Robotaxi and Optimus updates
AI Stocks Report including Marvell MRVL
@EthanHe_42 on our Pod Watch
0:00 Tesla SpaceX No Merger
1:28 Terafab Joint Venture
4:14 Terafab Chip Cost
8:05 Tesla SpaceX Chip Savings
12:20 Tesla Robotaxi Optimus
14:38 AI Stocks Marvell
@JacobEdwardInc@MB_Trening I really appreciate what you’re saying I have 5 kids and couldn’t agree more. I’m having a really hard time getting myself motivated to get back in the gym
The LARGEST human ivermectin cancer study EVER conducted found 84% of cancer patients declared COMPLETE REMISSION, TUMOR SHRINKAGE, or HALTED TUMOR GROWTH.
Our study is now PEER-REVIEWED and PUBLISHED by the International Institute of Anticancer Research.
The tide is turning.
Interviewed my almost 95 year old dad about his experience with his one month old Tesla Y. Here’s what he said both the good and the bad. Please show anyone on the fence about getting a Tesla Full a self Driving car or truck. It may change their mind.
I've run Local SEO for 700+ home service companies across 40 states. Exposed every strategy that actually moves the needle. Exposed every one that's a complete waste of money.
Here is the full 2026 playbook I'd use to dominate any suburban market in America from day one. Google Maps. Website. Reviews. GBP. Citations. Content. All of it. Step by step.
I'm posting the entire system for free. 🧵🧵
Get this book. Keep the link in your phone. Then anytime one of your friends is having a baby or taking their child to the doctor for " well baby checks" ( nothing more than a chance to induct your child into a lifetime of pharmaceutical control) then send them the article. One parent at a time. This is how we finally kill this sacred cow.
I absolutely love this.
Have one in college and one just graduated HS.. but I have 4 still in the house and we’re definitely implementing!!!
Wished I have done this sooner
If I were in my 30s or 40s right now and wanted to leverage AI to retire within 10 years, here's what I'd do:
1. Immediately form an LLC company. Not next month. Not once you're 'ready.' This week.
$SITM @SiTimeCorp
Today we do our third earnings review of SiTime Corporation (SITM) as Top Pick.
We added the company on 8-25-2025 for $234.
The stock is trading at $750 in the after hours session off exceptionally strong guidance.
The quarter makes the point plainly.
SITM is scaling faster than expected, but the more important development is qualitative: the company is turning precision timing into a platform layer.
This is no longer a narrow AI data center story, even though AI infrastructure remains the obvious accelerant.
Management is now connecting the same timing architecture to optical modules, switches, telecom, LEO satellites, drones, smart munitions, satellite communications, positioning navigation timing systems, autonomy, industrial automation, and physical AI.
That broadens the thesis materially.
The common thread is not “AI” by itself.
The common thread is that faster, denser, more distributed, more autonomous, and more mission critical systems need better synchronization.
Timing used to be a background component.
SITM is making the case that it is becoming a system level requirement.
And this quarter, the numbers continued saying that out loud.
Why We View SITM as One-of-One
When we first added SITM as a Top Pick, certainly we were bullish.
But as time has passed, we have grown more bullish, after expert calls and more time to dig into the nuts and bolts of its business and that of the timing business in general.
We use this section as a sort of “standing thesis,” knowing full well that what stands can also fall.
But for now, between AI, data centers, defense, and satellites, we just see several green shoot business lines for this company, each large enough to support substantial growth on their own.
The History of Timing
Long before 2019, cloud servers and networking gear already used clock chips and oscillators because computers have always needed an internal “beat” to coordinate processors, memory, storage, and network traffic; Intel’s server clock-device documentation and AWS’s hardware time docs both reflect that timing hardware is a normal part of server and cloud infrastructure, not something invented for AI.
A Word on the Literal Function
It is conducting timing by sending a steady beat, basically a stream of ultra regular pulses, so the rest of the machine knows exactly when to do each step.
Think of it like a drummer counting 1, 2, 3, 4 for an orchestra: the timer is not carrying the music itself, it is carrying the beat that keeps everyone lined up.
What Changed
What changed with AI is not that timing suddenly appeared. What changed is that AI boxes are denser, faster, and more interconnected, so the quality and amount of timing control matter more than they used to, which is why companies like SiTime are seeing a bigger opportunity now.
Quartz to MEMS Timing Transition
Quartz did suffice for a long time, and it still does in many products. The issue is that in tougher environments, heat changes, vibration, shock, and very dense electronics can make old style quartz parts less ideal, while MEMS (Micro-Electro-Mechanical Systems) can be built to stay more stable and rugged under those conditions.
(MEMS timing devices are a modern, silicon-based alternative to traditional quartz crystal oscillators, utilizing microscopic mechanical structures to generate highly stable frequency references.)
So the real answer is not “quartz failed.” It is that as systems got faster, hotter, smaller, and more demanding, some customers started wanting a tougher and more controllable clock than traditional quartz could always provide.
This was a change, but it was NOT the change that made SITM special.
Many companies, moved to MEMS. The change was on the back of this.
SITM had foresight, and even as many companies started to create MEMS chips, SITM did something VERY different.
SITM did not treat MEMS as a simple quartz substitute as much of the world, maybe all of it, did.
SITM treated MEMS as the foundation for an entire timing platform and no other firm really conceived of this because historically the timing segment was boring and commoditized.
That is the key difference.
SITM went further on purpose. It built not just the vibrating element, but the finished clock products, the control circuitry, the signal-cleaning layer, the synchronization layer, the distribution layer, and even the software that helps customers manage timing across real systems.
That is why SITM’s platform matters.
At the low end, it sells basic resonators and oscillators, which are the parts that create the machine’s beat in the first place – no big deal. But it did not stop there.
In fact, SITM created the industry’s first “clock-system-on-a-chip,” integrating multiple timing functions into a single chip to simplify design and reduce component count
With Chorus, SITM built a clock generator family that integrates the MEMS resonator and the clock generation function into one device, removing the need for a separate external quartz reference.
Then it kept climbing the stack.
Its jitter cleaners do not just make time; they clean noisy time so very fast systems can run more reliably. Its network synchronizers help keep whole systems aligned, including applications that depend on IEEE 1588 and SyncE.
Its clock buffers distribute timing signals across more destinations. Its clock generators can take in a reference and create multiple precisely shaped outputs for different parts of the machine.
That is the difference between selling a component and solving a timing architecture problem.
SITM also built products for the harder corners of the market where timing quality matters more and customers pay for it. This is key, because until this world opened up, timing was just a low cost commodity.
Elite X Super-TCXO is aimed at environments like data centers and edge networks where heat, airflow, vibration, and other real-world stressors can degrade traditional timing.
Its programmable OCXO and Super-TCXO products push further into high-stability timing jobs that go well beyond the original “replace quartz with MEMS” idea.
And then there is the software layer, which is easy to overlook but strategically important. I think analysts miss this too.
TimeFabric is SITM’s system level timing software suite for things like holdover and PTP synchronization in AI data centers and networks.
That matters because once you are selling software and system behavior on top of hardware, you are no longer just a parts vendor.
You are helping customers manage timing as a platform problem. Many MEMS players addressed the vibrating element.
SITM addressed the full system around it.
This is why SITM did not end up as just another MEMS timing company. It became the company trying to own timing at the system level.
SiTime is no longer optional infrastructure.
It is priced into the physics of scale, and that is a very different business than the one investors ignored for decades.
And now it has a multi year lead, a data center business growing 158% year over year.
Those are cuckoo numbers.
And yes, this sounds like the evolution of Nvidia, from a sort of “oh yeah, we need GPUs” company, too the engine of AI.
SITM will never be that. SITM will never be what NVDA is. Timing will never be what GPUs are.
But the similarity between that two is the dominance of an entire market that is a “must have,” that went from niche to booming.
1/
...
Leopold Aschenbrenner just increased his position in one stock by 648% in a single quarter.
While everyone focuses on his $8.4 billion in semiconductor puts, this conviction add is hiding in plain sight.
He’s accumulating aggressively.
Here’s exactly what he is buying and why:
The stock is CLSK. CleanSpark.
One of the largest publicly traded Bitcoin miners in North America.
Mining facilities across Georgia, Mississippi, New York, Tennessee, and Wyoming with massive existing power infrastructure.
CleanSpark is no longer just a Bitcoin miner though.
The company is actively pivoting its existing infrastructure into AI and HPC compute hosting.
Cheap power, cooling, real estate, grid interconnects.
The pieces every AI lab needs are already in place. The pivot is the trade.
This is also consistent with the pattern across his entire filing.
While shorting the chips with $8.4 billion in puts, he loaded up on the Bitcoin miners pivoting to AI.
Applied Digital, Bitfarms, CleanSpark, IREN, Riot.
He’s long the power infrastructure of AI.
What makes CleanSpark different from the others:
– One of the highest hashrates among public Bitcoin miners.
– One of the deepest power positions in the industry.
– A real Bitcoin treasury that revalues every time BTC moves.
Three layers of upside in one stock.
We are a team of 8 with one goal: help you find the right stocks early.
But you need to follow us with notifications so you don’t miss our alerts. This is EXTREMELY important.
A lot of people will regret not following us.