EnerTuition

14.2K posts

EnerTuition

EnerTuition

@EnerTuition

Skeptic honed in seeing through IB and management BS. Writing about AI, CPUs, GPUs, Solar, BEVs, Autonomous driving, etc. Tag me on anything related.

enertuition.bsky.social Katılım Ekim 2017
330 Takip Edilen2.1K Takipçiler
Sabitlenmiş Tweet
EnerTuition
EnerTuition@EnerTuition·
Beyond The Hype - the rare against-the-grain technology stock analysis service – is now exclusively on substack. Here is a preview of the type of content you can expect (much free content as you scroll down): enertuition.substack.com
English
0
1
2
10.3K
EnerTuition
EnerTuition@EnerTuition·
$AVGO There has been much noise for a while about Broadcom's insurmountable SerDes leadership and why MediaTek cannot succeed. This was always a terrible take. Also, it is also good to critically evaluate management exhaust. Hock has been delusional on the competitive threat.
EnerTuition@EnerTuition

$AVGO dump is not about the quarter numbers of the next quarter guide but about competitive concerns. Beyond The Hype has been a very consistent voice on this subject and lightened up on AVGO in the model portfolio a while back. The growth story is risk laden.

English
2
0
20
5.7K
EnerTuition
EnerTuition@EnerTuition·
$AVGO dump is not about the quarter numbers of the next quarter guide but about competitive concerns. Beyond The Hype has been a very consistent voice on this subject and lightened up on AVGO in the model portfolio a while back. The growth story is risk laden.
English
2
0
19
8.7K
EnerTuition
EnerTuition@EnerTuition·
One of my favorite things to say about Citron is: Citron, where the research is the lemon
Wall St Engine@wallstengine

Citron Research founder Andrew Left was found guilty of securities fraud by a federal jury in Los Angeles. The case centered on tweets about dozens of companies that prosecutors argued were used to move share prices and generate ~$20M in trading profits from 2018 to 2023. Left testified in his own defense. The trial lasted three weeks, with the jury deliberating for two days.

English
0
1
0
378
EnerTuition
EnerTuition@EnerTuition·
$NVDA Jensen announces at GTC Taipei that Vera Rubin which was in full production at CES is in full production now. It took 5 months to send the design from Santa Clara to TSMC. One has to wonder if it will be in full production at CES 2027.
English
0
1
5
922
EnerTuition
EnerTuition@EnerTuition·
@Nvidia has too much money to waste it seems. Going after a small niche with a delayed product.
𝐷𝑟. 𝐼𝑎𝑛 𝐶𝑢𝑡𝑟𝑒𝑠𝑠@IanCutress

youtu.be/JdB722MK380 It looks like NVIDIA is launching the N1X tomorrow. If you've been following my tweets, it seems like a really confusing product. It's late, the power is questionable, and how much is this thing going to cost? It's 2:30am, Jensen might prove me wrong at the keynote tomorrow, but I wanted to share my thoughts.

English
0
0
0
196
EnerTuition
EnerTuition@EnerTuition·
@ShanuMathew93 @RichardMeyerDC @JigarShahDC I have suggested Jigar to take it easy on the point, but the man is on a mission. Smaller data centers will pervade because of inference and a lot of reasons but there are many factors that are pushing companies toward larger data centers. It is very far from black and white.
English
0
0
0
12
Shanu Mathew
Shanu Mathew@ShanuMathew93·
I think if there were no obstacles, I'd be more willing to underwrite GW sites but given the issues around it + smart engineering, feel like we land somewhere in the middle. The inference site is more desirable longer-term vs. a large training site that becomes more inflexible to longer-term changes (e.g., OpenAI walking away from the additional 800MW at Abilene due to tech stack changes). Training can also be in less optimal locations whereas it's easier to underwrite longer-term inference demand closer to population centers & better infra. Lot of data centers now being built for optionality/fungibility given how fast everything is moving. You'll run training + inference, you'll network across sites 100s-1000s of KMs away, etc. I think that all probably favors the more mid-sized sites but at the end of the day we are extremely compute constrained and demand seems infinite so there will be plenty of developers lining up to take the big swings. Comes down to finance-ability + offtake. Let markets do their thing and price that risk.
English
1
0
0
48
EnerTuition
EnerTuition@EnerTuition·
$NVDA $AMD A lot of influencers have been falling for @Nvidia Vera nonsense lately. I think the truth will be a lot closer to this @AMD take than the selective information Nvidia is putting out there. linkedin.com/pulse/part-ii-…
EnerTuition tweet media
English
0
0
2
318
EnerTuition
EnerTuition@EnerTuition·
Bad news for MediaTek. Musk's projects rarely amount to anything. Most probably waste of design resources.
Jeff Pu@sssjeffpu

Abstract of MediaTek (2454 TT) - Note - New ASIC likely SpaceX 🚀 •Target Price raised to NT$5,520 •Massive ASIC revenue ramp projected:
$2.2bn (2026) → $18bn (2027) → $48bn (2028) •Likely SpaceX Win for new datacenter ASIC. We expect MTK to be the design service provider for xAI’s datacenter ASIC based on Intel 14A + EMIB-T. Volume production late-2028, ramp in 2029 with robust volume. •Zebrafish CoWoS upgraded:
Capacity raised to 180k (from 160k) in 2027. ASP hiked to $5.5k (from $4.5k) → implies $17.8bn datacenter ASIC revenue in 2027E (well above guidance & consensus) •Strong TPU momentum:
Competitor’s PumaFish cancelled → MTK first mover to 2nm (Humufish). High confidence in successor chip (Helios Next) with strong EMIB yields (in out note, we highlighted EMIB has started trials for testing for major current AI chips. #Mediatek #TPU #SpaceX #Intel

English
1
0
0
2.5K
Muhammad Zuhair
Muhammad Zuhair@mzuhair123·
$NVDA Vera CPU Has Been Benchmarked For the First Time, and It Absolutely 'Obliterates' x86 Competitors @phoronix, a popular Linux-focused media outlet, managed to access NVIDIA's Vera CPU, and gladly, ran some tests on it, focusing on benchmarks like Geometric Mean, and many others, and interestingly, the results are absolutely surprising. "Going into this, I didn't really know what to expect of NVIDIA's Vera with the new Olympus cores. But in the end, I was left realizing this is the most formidable competition to Intel and AMD x86_64 processors ever realized. NVIDIA Vera is much more performant than what we have seen out of the likes of Ampere Computing or the custom in-house ARM solutions at public cloud providers like Google Compute Engine and Microsoft Azure. NVIDIA Vera is actually competitive to the latest AMD EPYC and Intel Xeon processors across a swath of workloads."
Muhammad Zuhair tweet mediaMuhammad Zuhair tweet media
English
10
5
55
11.8K
EnerTuition
EnerTuition@EnerTuition·
@JigarShahDC @Tesla Once Musk delivers the Solar Roof, the truck will come. And he will get to Solar Roofs after he does the alien dreadnaught factories. /s Quit amazing to see the world constantly paint a fraud as a savior.
English
0
0
1
64
EnerTuition
EnerTuition@EnerTuition·
@magicsilicon The leather jacket messiah was called it the dying product not that long back :-) People keep underestimating the importance of linear code in the world.
English
0
0
2
611
Pushkar Ranade
Pushkar Ranade@magicsilicon·
the killer product is still the cpu
English
24
23
198
18.5K
EnerTuition
EnerTuition@EnerTuition·
@EdLudlow It comes down to if journalist like you have the courage to ask tough questions instead of softballs. Diamands says the dumbest things and goes unquestioned. Feel free to point to some of the things raised here and see how Peter responds: enertuition.substack.com/p/elon-reveals…
English
0
0
0
45
Jeff Pu
Jeff Pu@sssjeffpu·
CPU to be Computex 2026 Highlights 🚀 •Nvidia Vera CPU focus: June 1 keynote to spotlight Agentic AI CPU with up to 1.5x faster, 2x perf & 4x density vs x86. •Rubin Pod layout: Early 1:1 GPU:CPU ratio (2 Vera racks + 16 Rubin racks). Potential to reach 1:4 with agentic workloads. •Intel Wildcat Lake: Entry-level 18A CPU (2P + 4LP-E cores, 15W) targeting MacBook Neo competitiveness. Nova Lake (52-core) ES started. •Server CPU TAM: Explosive 49% CAGR to $211bn by 2030E, driven by Agentic AI shifting power ratios toward higher CPU density. •Stock implications: INTC, AMD, Arm, Kinsus (3189 TT), IBIDEN (4062 JP), Unimicron (3037 TT), FIT (06088 HK) #Computex2026 #AI
English
5
36
288
317.8K
The Information
The Information@theinformation·
SpaceX’s prospectus shows a company still anchored in Starlink and rocket launches, not yet the moonshot cloud-computing business implied by a $1.75 trillion IPO valuation. On one back-of-the-envelope analysis, its current businesses add up to closer to $700 billion. Read more: thein.fo/3PCxjMj
English
2
1
16
3.6K

Keşfet

@MicronTech @Samsung @SKhynix @nvidia @ShanuMathew93 @RichardMeyerDC @JigarShahDC @AMD