AI Infrastructure • Compute Economics • Semiconductors • Models • Power • $NVDA since 2016
Los Angeles, CA
Joined April 2009
- Tweets35.4K
- Following1.3K
- Followers26.2K
- Likes77.1K
Pinned Tweet
I’ve been an NVIDIA shareholder for nearly a decade, and I’ve seen plenty of reasons to sell along the way.
I think this selloff is BS, and I see it as a buying opportunity.
The company’s products are still in demand. Developers are still building on its platform. And the buildout of AI infrastructure isn’t going to stop because of a chaotic news cycle.
I believe Jensen will continue to lead NVIDIA. But this company is also bigger than any one person. It’s thousands of incredibly talented people building products that have taken decades to develop.
Look at Apple. It went through enormous uncertainty around Steve Jobs, yet the business continued to grow and return capital to shareholders. The circumstances are different, but there’s a lesson in looking beyond the headlines.
There could be more volatility ahead. That doesn’t change how I’m thinking about the long-term opportunity.
I’m long NVIDIA, and I’m buying more. $NVDA
NVIDIA shareholders can JOIN FORCES.
The bylaws allow multiple holders to combine their stakes to request a special meeting if they collectively control 15% of voting power and have held those shares continuously for at least a year.
That’s not a veto over the CEO decision. But the board answers to shareholders, too.
$NVDA
Anthropic just disclosed its revenue run rate jumped from $71B to $81B in a month. That’s 14% growth.
But sure, let’s call the AI trade broken because a non-representative slice of Ramp’s token-spend customers grew 0.5%.
The methodology literally says it excludes subscriptions.
Maybe check the actual company filings before declaring the boom over.
NEW Ramp AI Index: In October, AI spend rose 2%. OpenAI and Anthropic growth fell to low single-digits.
Engineering teams drove 72% of spend, followed by customer service and leadership.
Given this relatively modest growth in spend, and new competitive dynamics that may stick around, the markt should look to new fronts in the AI trade.
Love you Marcelo but why do people keep saying “CUDA is dead” because Anthropic got part of a model running on AMD in three days.
That’s impressive. But declaring NVIDIA’s moat gone from that is bullshit.
CUDA isn’t just a few lines of code you port over. It’s a massive ecosystem of libraries, tools and years of optimization that developers use across AI, science, robotics and more.
As John Linford and I discussed in our video (link below) getting something to run on another chip is one thing. Getting the whole workload running reliably at scale, with competitive performance and cost, is another.
Competition is real, and AI will make it easier to use other hardware. That doesn’t mean the entire ecosystem NVIDIA built over decades just disappeared.
youtu.be/yunV1tlr3bg
Follow the timeline and intentions it’s all starting to make sense. Nice job Mo
$NVDA
Would you hand a new hire the keys to your data center on day one?
NVIDIA’s Adel El Hallak explains why AI agents need clear boundaries, including OpenShell’s deny-by-default runtime and explicit permissions.
Full interview Friday at 5am PST.
AI creates more bytes. That doesn’t mean every byte belongs on flash.
During inference, weights are mostly read; a conversation’s KV cache keeps being written.
Reuse, reload time and wear determine the economics.
Our free report with Rick Xie:
newsletter.bepresearch.com/p…
The 2x agent-CPU headlines shrink once you check what they're measured against.
NVIDIA's Vera ($NVDA) vs AMD's current EPYC: about 1.2x under load. The bigger multiples are against Intel.
Arm's 2x per rack: about 1.2x per chip. The rest comes from fitting more servers in the rack.
I put Vera, EPYC, Xeon, Graviton and Arm's chip through five tests. None passes all five yet.
Part 3 of my CPU series:
newsletter.bepresearch.com/p…
3
15
2,746
A little timeline cleanse: let’s talk about what NVIDIA is actually building.
Everyone’s talking about AI designing chips. What about testing them?
In this clip, NVIDIA’s Adel El Hallak shows how teams of AI agents can help generate tests, find bugs and run regression testing, turning a verification process that can take weeks into hours.
The tools used to build the next generation of chips are changing too.
Our full interview with Adel drops Friday. $NVDA
2
13
2,401
I’ve been thinking about Apple during the iPhone years. When Steve Jobs was diagnosed with cancer, investors had to confront a difficult question: what happens to Apple without its founder?
The circumstances around Jensen today are very different, but they’ve brought up a similar question about NVIDIA. The board’s review needs to run its course, and I don’t expect Jensen to leave.
What’s interesting is the financial parallel: Apple’s earnings grew while its multiple compressed, and buybacks later became a big part of the story. NVIDIA has its own products, people and opportunities ahead.
History doesn’t repeat exactly, but sometimes it rhymes. The mission is the boss. $AAPL $NVDA
Welcome to X @AminVahdat
Excited to have you here!
Thrilled to share a big milestone from our AI & Infrastructure team at @Google.
Today we’re expanding our 8th generation TPU portfolio with:
• TPU 8i Max: Built for massive context windows and the highest compute density per rack
• TPU 8tx: Purpose-built for secure, confidential inference
This marks the first time we’ve introduced four distinct TPU systems in a single generation, in one year.
For developers, hardware is only as good as the software stack. And we want you to be able to build with TPUs as easily and efficiently as we do at Google. You can now:
↳ Run native PyTorch & Hugging Face models via TorchTPU (zero code rewrites)
↳ Cut KV transfer latency >90% with open-source TPU Sync
↳ Eliminate idle compute waste with Agent Compute Engine
Read more about our newest TPUs on our blog: bit.ly/3VfETPR
I look forward to connecting with builders, researchers, engineers and organizations pushing the boundaries of AI and computing!
Wow, didn’t expect my Sunday to go like this.
There’s a lot of hysteria on X right now. Stacy Rasgon @Srasgon has covered NVIDIA for years and understands the company. He’s a voice of reason amid a lot of armchair commentary from people who don’t.
In his new note from today, “NVIDIA: Stirring the (honey)pot?”, he says the situation could create political and regulatory noise, and the overhang may stick around. But he doesn’t expect it to materially change demand for NVIDIA’s products.
A headline can create uncertainty without changing what customers are buying or the long-term demand for compute.
Bernstein maintained its Outperform rating and $400 price target.
On Friday, I defended Jensen and said he was faithful to his wife. I was wrong, and I own that. I spoke with certainty about something I didn’t know.
As a husband and father of three, I believe infidelity hurts families. I’m disappointed.
At the same time, NVIDIA is bigger than any one person. Thousands of hardworking employees have poured their talent into building something extraordinary. I’m proud to support their work.
BREAKING
Lots of smoke no fire. Nvida BoD will investigate. Carry on
$NVDA
Love Jon Fortt. Real Journalism in action. Asking the right questions and piecing the puzzle together. This is all by design
More details now from the NYT’s leading reporting about the Jensen/China story. The TYPE of detail is interesting: Marco Rubio, long considered a China hawk, oversaw the drafting of the memo so many are buzzing about.
We don’t get much more detail here about the allegations themselves, except that they are from “electronic surveillance” of Alice Huang. That could mean a lot of different methods.
A line that caught my eye: “a number of people inside the administration were concerned about Mr. Huang’s influence on the president’s policy of only lightly regulating A.I., people briefed on the matter said.”
So, China hawks in the administration oversaw the development of this memo ahead of a trip to China. They distributed the memo. Jensen initially wasn’t invited on the trip. Then President Trump invited Jensen anyway. The hawks’ efforts to distance Trump from Jensen’s influence failed.
Now, days before the president is set to meet with the Chinese again, the allegations surface publicly. The first effort rolls off the news cycle, overtaken by midterms. The second, more salacious pass, lands right at the end of the week as the post-midterm buzz is ebbing. And the Rubio detail lands on a Sunday, fresh for the new week.
The administration China hawks who didn’t win the argument behind the scenes sure are lucky this came out somehow. #nvidia #jensen #china $NVDA
nytimes.com/2026/11/08/us/po…
3
23
14,679
Timeline cleanse
I’ve been writing about this: the next inflection is the lab.
I lost both my parents to cancer and my father in law to Alzheimer’s
Jensen lays out the vision here. Connect AI agents to instruments and robots, run experiments, and feed the results back into models.
I’ve seen a piece of this firsthand at Terray, where I’m an investor. The moat is the measurement.
More experiments. More data. More tokens. 🧬🤖
We need faster cures
People are worried politics will slow down AI infrastructure.Meanwhile, @FundaAI reports SpaceX has ordered 6,000 racks for orbital compute.
Elon says the computers are being co-designed with NVIDIA and will “blow your mind.” 🤯
The next frontier of AI could be above us. 🚀
Weekends are for philosophy
I've known and followed Jensen Huang for over a decade. He's devoted to his wife, and he's a true American patriot.
China wants what we have in AI. It has tried to close the gap through talent, technology and model distillation. But how much could anyone learn about NVIDIA from one conversation with Jensen?
Look at lithography. China is working on its own EUV machine, while ASML has spent decades building an entire ecosystem around it. China has a commercial jet, too, but it still uses Western engines and hasn't secured European certification.
NVIDIA is no different. CUDA, the chip designs, the software, and the people behind them took decades to build. You can't just copy and paste all of that.
America's greatest advantage is the diverse talent that comes here to build. That's a lot harder to copy than a piece of technology.
New: The spy who targeted Jensen Huang was recruited by MSS, China's top intelligence agency. @trippmickle @paulmozur @maggieNYT nytimes.com/2026/11/06/us/po…
Sorry I know I’m going to piss off a lot of the memory and storage guys but chips without software is just expensive sand
Know this. Kagan is an OG
$MU $SNDK $SKHY $WDC $STX
Chip without software is just sand!
Michael Kagan, NVIDIA CTO and Mellanox founder, on his Intel days. They built a vector processor "about 50 times faster on the vector arithmetic." But "Intel didn't invest in software," and it "didn't have any significant market success."
His lesson: "chips without software are just expensive sand, not more than that."
That's the NVIDIA playbook. CUDA turned the GPU into programmable intelligence. Mellanox and NVLink took it further, so many GPUs show up as "a single GPU exposed to the software."
Everyone is trying to copy the chip. Almost nobody can copy the stack.
@midnight_captl 🔥 $NVDA