Swollard

412 posts

Swollard banner
Swollard

Swollard

@SwollardCapital

Investor and Entrepreneur| Family Portfolio Manager| 7+ years investing+options| Not financial advice

Canada Katılım Şubat 2021
168 Takip Edilen63 Takipçiler
Sabitlenmiş Tweet
Swollard
Swollard@SwollardCapital·
This semiconductor sell-off is reminiscent of the October 2025 sell-off. However, this time, stocks have sold off amid extremely high expectations. Fundamentally speaking, nothing has changed. In some ways, beneath the surface and away from the headlines, circumstances have arguably improved or are continuing to improve. I’m confident that hyperscalers will increase Capex and provide incredibly bullish commentary supporting the AI buildout. $GOOG and $INTC will be our first read-throughs this upcoming week. A few things I’ll be watching are 2027 AI infrastructure pipelines, capacity constraints, HBM procurement, GPU deployments, and networking investments. I keep coming back to the conclusion that we are still in the early stages of this technological revolution, and I continue to invest in the companies directly benefiting from the projected $1 trillion in Capex for 2027. Are you buying, selling or waiting? What are your thoughts?
English
1
1
0
78
Swollard
Swollard@SwollardCapital·
Idk if I should listen to this guy and Larry Fink and other incredible investors or some random dumbass on X??
Fireside Alpha@firesidealpha

Gavin Baker @GavinSBaker says open source is "awesome" for AI infrastructure names because it shifts the economic value from the frontier labs to the infrastructure, and it is already 80% plus of all tokens processed. "A misconception that a lot of people have is that open source models are somehow bad for AI. They're awesome for the AI infrastructure providers. They just shift economic value from the margins of the frontier labs to the infrastructure." "But I do think there's still a role for these frontier models. And it may be true, to date, frontier tokens are capturing 90% of the economic value, and open source tokens are probably 80% plus of tokens processed, and those ratios may be here to stay." _______ Follow @firesidealpha for more highlights of the best business and technology conversations.

English
0
0
1
14
Swollard
Swollard@SwollardCapital·
To summarize, buy memory, connectivity and CXL stocks
FUNDA@FundaAI

Deep|LLM: Kimi K3's KV Cache Is Smaller - Why That May Actually Be Positive for DRAM/NAND K3 is still a large-memory model: Kimi K3 has 2.8T total parameters, activates 16 of 896 experts per token, and Moonshot recommends deployment on high-bandwidth supernodes with at least 64 accelerators. Its smaller KV cache does not mean lower overall memory and networking demand; K3 still needs HBM-scale capacity, scale-up networking, and GPUs. KV offload becomes practical only after compression: Large uncompressed KV caches are hard to place on the active decode path because PCIe, NIC, and SSD latency can erase the benefit versus recomputing prefill. KV compression reduces the data that must move across DRAM, NAND, and the network, turning offload from a theoretical capacity extension into a deployable architecture. DeepSeek V4 shows the threshold effect: DeepSeek V4-Pro cuts KV cache to 10% of DeepSeek V3.2 at 1M context, yet channel checks suggest NAND usage has increased in real deployments. The reason is that smaller KV enables higher offload ratios, more retained sessions, longer prefixes, and better cache-hit economics. Implication for DRAM/NAND: The key trend is not KV cache disappearing, but KV cache becoming a tiered data asset across HBM, DRAM, NAND, and networking. DRAM becomes more important as a warm-cache and staging layer, while NAND becomes a large-scale pool for historical sessions, shared prefixes, and inactive KV. Detailed Report fundaai.substack.com/p/deepllm-kimi…

English
0
0
0
14
Swollard
Swollard@SwollardCapital·
@jrouldz Need to make an X sentiment tracker and fade it
English
0
0
1
23
Ben Bajarin
Ben Bajarin@BenBajarin·
Remember my saying. In a compute shortage (and we are in a deep one) all viable compute will be used. I think they announce more new customers as well. Looking forward to attending AMD Advancing AI day this week to be boots on the ground with @austinsemis and @mweinbach!
SemiAnalysis@SemiAnalysis_

ANTHROPIC WILL BE AN AMD CUSTOMER, ACCORDING TO THE PUBLIC GITHUB OF AMD’S SENIOR DIRECTOR OF AI 🚨🚨 We explain the GitHub code and nuances below👇️ 1/4🧵

English
4
6
103
25.6K
Swollard
Swollard@SwollardCapital·
This is just human psychology playing out over and over again
English
0
0
0
14
Swollard
Swollard@SwollardCapital·
So now that stocks might go up again, is everyone going to turn bullish again? Lmao price action shouldn't determine if you're bullish or bearish. Follow the fundamentals, ignore the noise fam. Stay. Bullish. $dram $skhy $mu $intc
English
1
0
1
288
Swollard retweetledi
Swollard
Swollard@SwollardCapital·
So wtf are u bears going to do when memory companies become valued as custom silicon companies rather than commodities? ItS cYclICaL 🤡 the only thing cyclical around here is ur IQ
English
0
1
0
19
Swollard retweetledi
Swollard
Swollard@SwollardCapital·
Since we're all getting cooked in the markets, I thought I'd change the topic and share my work station. Just finished up cable management, what a difference. Share your setups fam 🖥️💰 $dram $skhy
Swollard tweet media
English
1
1
1
208
Swollard
Swollard@SwollardCapital·
Isn't $spcx like $10-15 bucks off its IPO price? Tell me u bought the top without telling me u bought the top
English
0
0
1
27
Parmita Mishra
Parmita Mishra@parmita·
me when I get banned from Anthropic for life for asking “how do you synthesize acetaminophen?”
Parmita Mishra tweet media
English
19
7
221
38.1K
Gregory Blotnick
Gregory Blotnick@gregoryblotnick·
PTJ - on time stops; "When I trade, I don't just use a price stop, I also use a time stop. If I think a market should break, and it doesn't, I will often get out even if I am not losing any money." Very very tool/skill to add to ur repertoire Because u anticipate, “market will do X over next 2 weeks.” Forgetting P&L, if the market isn’t acting how u thought it would, why are u still in the trade??? Exit To do otherwise is to operate w no plan or no discipline If the tape feels off relative to ur expectation, exit early, exit immediately Lay the plan out and if the mkt doesn’t conform to the plan, exit asap, draw up new plan Hanging around in a busted idea bc ur like, lost, or have no other ideas….ur discipline is fucked. It’s awful. U cant even execute on little plans, don’t expect to be trade calmly in a crash tape. Lots of trading is making little promises to urself, repeatedly, and being ironclad on every single one U break one, u are fucking worthless Ironclad or nothing, no good trader is a “lots of exceptions” thinker Add time stops, get used to them, both PTJ and Platt support them, and are very good for keeping u honest and adhering to the plan u PREVIOUSLY laid out Changing plans on the fly….NGMI Plan it out, full range of outcomes, execute when path arrives, exit/stop out, new plan, repeat Be a robot. Be professional. Guys like PTJ, the trade entry, take profit, and stop out…lot of times guys just put in all 3 at once. Then u chill Nothin to think about Experiment w all this shit but the “fast trigger finger —> revert to neutral/flat/cash”, having that shit lightning quick, is what will ensure u always live to fight another day.
English
1
1
23
2K
Swollard
Swollard@SwollardCapital·
@TomLeeTracker Fkn larp, I thought ETH was supposed to be $10k+ by now
English
0
0
0
6
Tom Lee Tracker (Not actually Tom)
TOM LEE: MARGIN DEBT IS UP 54%, THE SIXTH HIGHEST RISE IN 60 YEARS A cohort of traders borrowing to buy has piled in, and history says the market reacts: - Margin debt is up 54% year over year, the sixth highest rise in about 60 years of monthly data - The other five times, the market consolidated over the following six months, because once that borrowed-up cohort is tapped out, it's vulnerable to a drawdown - Korea just showed what that looks like: 1.2 million brokerage accounts hit with a margin call, possibly 10% of all adult accounts, followed by a severe correction
English
7
29
248
106.5K
Gregory Blotnick
Gregory Blotnick@gregoryblotnick·
@SwollardCapital u got it brotha a lot of the articles in my pinned tweet go deeper on many of these same concepts, run thru those and after a day when I tweet a lot I usually take 3-4 days off to go read books + “refill” but always reply/dm with q’s etc and I’ll hit them on the next go around
English
1
0
1
53
Ren
Ren@ren_stocks·
If anyone wants to know what my bottleneck is for tomorrow, it’s this one right here. You think you have it bad after losing 30-50% of your port in two weeks? That’s nothing but a scratch if you’re in the right names for the long term. Seeing my 3-month-old in the ER because he got sick this week and couldn’t breathe? 100x worse. Don’t invest more than you can afford to lose. Have your priorities straight. You don’t lose when the market goes down. You lose when you lose the conviction that made you buy in the first place, just because the price changed. A good reason to sell is when the underlying changes. When new evidence forms that presents not just a challenge but a major uphill climb, one that most likely won’t be resolved. The biggest lesson time in the market has taught me is that time in the market, in companies building the future, is the best return on both my investment and my sleep quality. As for my baby? He’s already doing so much better. Just like most of the stocks building the future. Watching closely as margins compress, but that day is not today. Not yet. Not soon. Happy weekend ⚡️
Ren tweet media
English
40
6
311
18.1K