Founder, CEO, Chief Analyst @MoorInsStrat. Co-founder of @TheSixFiveMedia and @Signal_65. Ex-AMD CVP & Fellow, Compaq and AltaVista dot-com bomb participant.
Austin, TX
Joined May 2008
- Tweets73.4K
- Following4.8K
- Followers70.3K
- Likes74.7K
Pinned Tweet
It's The Week Ahead In Tech for November 9, 2026: @nvidia's board is investigating whether Jensen Huang was the target of a foreign intelligence effort, right before NVIDIA reports.
The board sees no reason to believe NVIDIA secrets were compromised. Three House committee chairs reportedly want a closed-door briefing.
Let the facts come in.
Tonight I'm on BBC News talking $NVDA. Friday I'm on @CNBC Closing Bell Overtime previewing its earnings.
$CSCO booked $9.3B of AI orders from the cloud giants last year. Is the gap to revenue supply or timing?
$CRWV: its biggest customer went from 71% of revenue to 36 in a year. Do its borrowing costs reflect that yet?
It’s The Week Ahead In Technology for November 2, 2026. $MRVL agreed to buy $LITE in a deal valuing it at 115 billion dollars. A chip company is paying up to own the lasers, because Marvell says copper wires can’t keep up with most of AI. Thursday I’m on @schwabnetwork the morning after $QCOM reports.
It’s all stock. Lumentum shareholders end up with about a third of the combined company. Marvell says it sees more than 1 billion dollars a year in cost savings, with close expected in the first half of 2028.
@intel says the PC slowdown lasts at least through the first half of next year. If $AMD shows the same pain, it’s the market’s problem. I don’t know if it will, though.
Also: $AAPL $ARM $LSCC $COHR $ALAB $IONQ.
Should chip companies own the lasers?
The reason I left out many questions below is that there will be USG *and* BoD investigations.
The Senate Banking, Intelligence Committee and House Foreign Affairs Committees will investigate this too. This *won't* be a sloppy R versus D lines. And many of these will be public.
My DC sources tell me this will be "China Hawks vs. Uber China Hawks." Also, that this will also make it likely that some form of export control legislation will pass if not in this year’s NDAA, it will pass next year.
Like I said: teapaw.org/JohnCornyn/status/2119…
Headed on @BBCBusiness tonight to discuss H200s and China. Was scheduled Friday. Something tells me we may shift the topic a bit. 😑
The Nvidia board will run an SoP board investigation. What I expect it to look for:
- Was anything sensitive shared?
- Did he follow Nvidia's code of conduct, including disclosing conflicts?
- Did any Megaspeed business decisions get touched?
- Were Nvidia's public statements accurate?
BREAKING: Nvidia CEO Jensen Huang admits affair, board to launch investigation: ""I was unfaithful to my wife. My actions hurt her and our children. There is no excuse for what I did." axios.com/2026/11/08/nvidia-…
When X started to get negative, I mute words and it works well. Problem is, I can’t mute the words “datacenter” or “Elon”. Maybe I mute “governor” and “senate” for a while.
Max is amazing. Even if he started 2nd I don’t know if he would have caught Norris. Maybe but at that pace, Lando would have kept it.
This was a fun exercise. The best was ChatGPT, the worst was Claude.
“Fill this room with what you know about me.”
Early pod post elections and of course earnings breakdown.
Data centers were on the midterm ballot, and AI stocks went from euphoria to a sharp selloff in just two days.
On Ep. 326 of The Six Five Pod, @PatrickMoorhead and @danielnewmanUV unpack the political implications for AI infrastructure and what earnings from $AMD, $AAPL, $QCOM, $ARM, $LSCC, and $COHR reveal about demand.
Tune in:
00:00:00 - Intro
00:01:40 - An Important Time In History
00:05:27 - Neocloud Stocks
00:07:11 - US Midterm Results
00:10:19 - Voter Surge In Texas
00:11:52 - Data Center Developments & Political Shifts
00:20:00 - Global Data Center Opportunities
00:26:35 - Data Center and AI Guardrails
00:32:00 - AMD's Strong Performance & Future Outlook
00:37:30 - Apple's Market Position and Challenges
00:41:20 - Qualcomm's Diversification + Growth Potential
00:48:50 - Arm's Continued Success in the Market
00:52:35 - Lattice Semiconductor's Impressive Growth
00:55:04 - Coherent Q1 2027 Results
The knock on DGX Spark was that it's a developer box, not an inference box. This open-source DeepSeek V4.1 Flash custom build on four of them says otherwise:
- 80 tok/s chat, 138 tok/s code, single user
- 438 tok/s at 16 concurrent users, ~27 tok/s each
- 10K tok/s prefill, 1M context
- vs. SGLang TP4 on the same boxes: 18-25% faster single user, ~1.9x the prefill
Four desktop boxes and a 200G switch. No rack, no liquid cooling, no cloud bill.
The hardware was never the limiter. The software was. A purpose-built engine for one model on one topology beats general-purpose serving stacks, and that is the template for departmental AI: open weights from @deepseek_ai, small boxes from @NVIDIAAI, and a community writing the inference stack.
438 tok/s at 16 users DeepSeek V4.1 Flash on 4x DGX Spark, now open source:
80 chat / 138 code single stream
438 tok/s at 16 users
1M context, 10k tok/s prefill
image support soon
github.com/Picom-code/pico-d…
Best take on the Jensen-China story so far.
Let me add something from my time in DC this year: the intelligence agencies are 100% opposed to shipping ANY GPUs to China regardless of performance or BIS guidelines. And they’re also the ones leaking this.
Let the facts come in before jumping to conclusions.
A few things about this Jensen/China story that don’t (yet) make sense:
1. Why would China brazenly recruit an executive to get Jensen to advocate for a policy … that he was already advocating for – open global access to American AI chips?
2. If there’s such an unusual relationship between Alice and Jensen, why are there no significant photos or anecdotes so far, beyond showing ?
3. We’re in an era where the public is inclined to believe a story about a philandering tech billionaire. I tend to pause when there’s a vague story circulating that people are predisposed to believe. #nvidia #jensen $NVDA
nytimes.com/2026/11/06/us/po…
I guess there’s always local AI if this whole datacenter thing doesn’t work out.
438 tok/s at 16 users DeepSeek V4.1 Flash on 4x DGX Spark, now open source:
80 chat / 138 code single stream
438 tok/s at 16 users
1M context, 10k tok/s prefill
image support soon
github.com/Picom-code/pico-d…
Discussing $QCOM Earnings on Schwab Network, November 5, 2026 — Broadcast Analysis moorinsightsstrategy.com/res…
Another tech CEO gets it on datacenter locality. This time $ARM CEO @renehaas237. He nails it.
The US doesn’t have a labor or even time to build advantage and can we built outside the US. Most companies planned to build here to appease the administration, the promise to remove barriers, and in a few cases, latency.
Put the barriers up and these will get built outside the US. And all the manufacturing jobs and taxes with it.
FAFO.
Interesting.
