Bingoto (@Bing0to) header
Bingoto (@Bing0to) profile picture

AI investor // Memory Bull Local barista @interndao // Hyperliquidated turned Evangelist

Joined October 2014
Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)Photo by Bingoto (@Bing0to)
Unpopular opinion: Assymetric upside in crypto has mostly been captured Don't think it's necessarily a 24/7 kinda market anymore, unless a skilled trader Like 5 investable assets out there, tradfi won't save yr bags Tokenpickers world, on par with equities ROI going forward
6
2
69
8,727
Now I feel a lot better getting only filled for 4% at the IPO
If you ever feel stupid, just know that I didn’t invest in Anthropic at $183B because I wanted to continue supporting my ex’s lifestyle (Luckily my seed investment did well and I landed 9 figures of Anthropic stock though)
2
128
Bingoto retweeted
Ashton Kutcher after his $435m investment in Anthropic will be $7B at IPO (15x) and his $30m early check into OpenAI now worth $1.2B+ (40x)
Ashton Kutcher and Guy Oseary VC firm first invested in Anthropic in 2023 and since put in a total of $435 million. It now owns 0.5% of Anthropic. At an IPO valuation of $1.4 trillion, that stake is worth $7 billion.
217
591
130
9,101
1,684,202
Imagine getting sued for underbuilding with a p/e below 3 and the market expecting a supply glut by 2028
A federal court in California found that the plaintiffs’ allegations plausibly suggest Samsung, SK hynix, and Micron agreed to restrict DRAM supply. The court denied the three companies’ motion to dismiss, allowing the antitrust lawsuit to proceed.
3
133
God forbid these guys finally are making some money after decades of brutal bearmarkets where 90% died because of overbuilding Rest of the semi industry: “why are you not overbuilding again and giving all your margin to me?” The stocks are even priced as if an overbuild is in the cards by as soon as 2028 and they wont make any money anymore
A federal court in California found that the plaintiffs’ allegations plausibly suggest Samsung, SK hynix, and Micron agreed to restrict DRAM supply. The court denied the three companies’ motion to dismiss, allowing the antitrust lawsuit to proceed.
1
109
Jensen getting Jack Ma’d
74
Bingoto retweeted
Replying to @fchollet
So what? Speed and scale are most definitely scalars, and 10⁶ Von Neumann in a datacenter running at 10⁴ the speed, with 10⁹ better communication bandwidth than teams of humans will let you reach the same conclusions.
2
2
47
6,546
Bingoto retweeted
100% of dogs want a bone.
66% want Medicare for All 61% want a billionaire wealth tax 57% want a $20/hour minimum wage 63% want a data center moratorium 58% want to end military aid to Israel 73% want to ban super PACs 79% want a 10% credit card interest rate cap No more excuses. Democrats must deliver.
42
21
2
818
49,531
Bingoto retweeted
I respect Jukan, but I feel a little frustrated whenever I see this kind of argument. You can build any scenario around plausible assumptions, and yes, of course, the outcome will follow from those assumptions. If you assume demand collapses, customers accumulate inventory, and contract terms get renegotiated, then naturally, memory companies’ revenue and profits will decline. But honestly, this feels like the perspective of analysts who treat memory as just a commodity you manufacture and sell, without properly understanding the technological changes taking place in custom memory. Who exactly are these customers we are talking about when we say they could cancel LTAs or defer deliveries? They are the very customers working with memory companies to design the next generation of HBM and 3D stacked DRAM. After jointly designing and validating these products, can customers really cancel those volumes so easily or push deliveries out by one or two years? How would they manufacture their own products, which were designed around that memory? What happens to the launch and system deployment schedules they have already coordinated? Switching suppliers also raises questions of design compatibility and requalification. If you are going to argue that contracts can be adjusted, you also need to explain what that decision would mean for the very customers canceling them. This is one of the biggest reasons I keep saying that memory is no longer the same cyclical industry it used to be. The parties to these LTAs are already tied together from the design stage. But if you still view memory simply as a commodity where rising prices lead to more production, and inventory buildups lead to price cuts or canceled orders, you will inevitably miss this shift in the technical and business relationships. One more thing. These arguments about a collapse in memory demand seem to assume that planned AI data center investments will be canceled on a massive scale, or that AI infrastructure expansion will slow so substantially that memory demand falls sharply. What interests me more than whether contracts can be adjusted is why you expect demand to weaken that dramatically in the first place. I think it would be more realistic to discuss directly whether the AI era we expect will actually fail to materialize, and whether much of the AI infrastructure investment currently planned will disappear. If that is your outlook, I can at least understand why you would expect memory demand to collapse and customers to cut volumes even for products they have jointly developed.
I'd like to believe LTAs are durable too. But take one example. Analog chipmaker onsemi said this in 2023. On January 5, 2023, at JPMorgan's CES forum, CEO Hassane El-Khoury was asked how much LTSAs would still hold once the shortage eased. His answer: “They are a legally binding agreement ’cause they’re a liability for the customer because they’re take or pay.” He also said he would not accept a customer reopening price talks just because it could get the same parts 10% cheaper elsewhere. On October 30, 2023, on the Q3 earnings call, onsemi cut its SiC shipment forecast for the year from $1 billion to more than $800 million after one automotive OEM reduced demand. An analyst asked how this could happen when the company had been saying its LTSAs were legally binding. The same CEO reaffirmed that they were binding, but said he did not want to enforce the contracts just to push inventory onto a customer whose demand had dropped. If it worked for both sides, terms could be adjusted customer by customer, whether by a quarter, a year, or longer. In its 10-K filings, onsemi also states that it amended LTSAs with some strategic customers by mutual agreement in 2024, and changed delivery schedules and committed volumes on some LTSAs in 2025 as demand shifted. Microchip is another case. Customers in its Preferred Supply Program (PSP) originally had to place 12 months of orders and could not cancel or reschedule them unless Microchip raised prices. On its November 2025 earnings call, however, the CEO said the company had accepted some cancellations and pushouts so customers could adjust their backlog. On the same call, the CFO separately addressed the multiyear LTSAs the company had signed. Some of those contracts were still in effect, he said, but customers had been allowed to defer taking delivery of committed volumes by one to two years if they needed to. Every time I see cases like these, I doubt how much memory makers' LTAs will protect earnings in a downcycle. Right now the most reliable thing is the prepayments and deposits themselves. That cash can go into capacity expansion today. I can picture hyperscalers with no DRAM inventory continuing to buy at current prices. But once inventory has piled up, will they keep taking every committed unit at the same price? I have a hard time picturing memory makers holding customers to the high contract price all the way through. The LTA itself may never be broken and may stay in place. But if both sides agree to push this year's volumes into next year or two years out, this year's earnings estimates come down. If price gets renegotiated too, margin estimates have to come down as well. Again, I agree LTAs are sturdier than they used to be. What I question is whether, after demand turns, memory makers can still earn 80% margins on contracts signed during the upturn and keep collecting hundreds of billions of dollars in profit because of those LTAs.
54
88
15
646
148,030
Wait, I can’t just sort stocks by P/E, buy the first one, and retire? Groundbreaking stuff. Looking forward to part two: revenue isn’t profit.
the ai trade won’t bottom until market participants finally learn *not* to value companies based on fwd pe. a fwd pe of 40 is not necessarily expensive if earnings grow durably over decades. a fwd pe of 4 is not cheap if earnings last 18mo then go down 90%. a company’s value is the net present value of *all* future cash flow, not the net present value of *next year’s* cash flow.
1
1
122
I actually feel bad for Elon, even though I deeply disagree with him on some things. Imagine trying to build a car company in California in the mid-2000s, when American manufacturing is being completely offshored and almost all of Wall Street thinks you’re a joke and is rooting for your failure, while the entire oil and gas industry is using all their power and leverage to make you fail. But against all odds, you actually pull it off. You build cars that make electric vehicles desirable to millions of people, create an enormous number of quality American jobs, and force the global car industry to play catch up and move away from oil and gas. You would think California would be proud of you and the team. Instead you find yourself having to fight a completely corrupt and incompetent left government impeding all your efforts while you sleep at the factory night after night to avoid collapse. At some point, you would go to the right too. And because of your politics, it becomes a justification for people to celebrate your companies failing in their personal ideological fight with you. Even threats against you become something people explain away and use mental gymnastics to justify. Meanwhile, people like Scott Galloway and Bernie Sanders who have made absolutely nobody better off other than themselves get to lecture you about your value to society and even call for your murder without grounds from their vacation home while you're on 20 hour benders in your factories trying to get things built for all of society. If this were me I would lose my mind. Criticize his opinions. I do sometimes, especially on Ukraine. But the pleasure people take in wanting to destroy everything he’s built is hideous and frankly suicidal for the masses, especially coming from people who insist they care about workers, the climate, and civilization as a whole. As you try to make good things for society you are confronted with a world that worships people who care only about the perception of being good people, while they actually fight against your tangible efforts of making improvements for humanity. I don't think almost anyone would have made it out of that and if they did, they would probably not come away feeling so warm towards the left either. So I am rooting for Elon to succeed, and for humanity.
I believe Elon is a good person
441
1,329
116
12,449
422,731
Bingoto retweeted
현금이 귀해진 시기 현금을 가진 기업의 힘이 가장 강하다는 것은 당연한 세상의 이치
4
1
43
3,447
Bingoto retweeted
Interesting
I am excited to be an investor in trade(xyz) as they partner with ICE to bring 24/7 perpetual futures on Hyperliquid to the United States. As a market participant, I am looking forward to to using their innovative product to enhance my investment process.
3,716
3,600
2,266
28,323
6,554,700
I am excited to be an investor in trade(xyz) as they partner with ICE to bring 24/7 perpetual futures on Hyperliquid to the United States. As a market participant, I am looking forward to to using their innovative product to enhance my investment process.
2,943
3,372
2,374
29,042
9,176,145
We did our job today. Don't fuck this up Korea.
🤖 Made with AI
25
59
4
1,114
50,748
Bingoto retweeted
“who should we lay off first” CEOs:
BREAKING: Microsoft, $MSFT, is offering buyouts to 60,000 workers as part of the company’s efforts to cut costs amid AI adoption, per Bloomberg. Personnel on select teams in the US, Canada, India, and the UK at a vice president level and below will be eligible. This reflects ~27% of Microsoft's total workforce.
150
1,129
84
14,528
900,061
Noo but china is going to oversupply memoryyy
中 빅테크 알리바바·텐센트·바이트댄스, 삼전닉스에 D램 장기계약 ‘SOS’ $SKHY, Samsung 통상 3~5년 LTA 넘어 2035년까지 초장기 계약 요청 美 빅테크 선점 속 中도 가세…서버용 메모리 확보전 미국 빅테크에 이어 중국의 알리바바와 텐센트, 바이트댄스 등 대형 클라우드 서비스 제공업체(CSP)들이 삼성전자와 SK하이닉스에 서버용 메모리 장기공급계약(LTA)을 요청하고 나섰다. 통상 3~5년인 장기공급 기간을 넘어 일부 고객은 2035년까지 10년간 메모리를 공급하는 초장기 계약을 요구한 것으로 파악됐다. 27일 업계에 따르면 알리바바·텐센트·바이트댄스 등 중화권 CSP는 국내 메모리 업체들과 서버용 D램 장기공급계약을 타진 중이다. 미국 빅테크들이 이미 LTA를 통해 메모리 물량을 확보한 가운데, 중국 CSP들도 사업을 확대하며 메모리 장기계약에 나선 것으로 풀이된다. 요청 품목은 서버용 DDR(더블데이터레이트)과 저전력 DDR인 LPDDR이다. DDR은 서버의 중앙처리장치(CPU)가 데이터를 빠르게 읽고 쓸 수 있도록 돕는 주기억장치다. LPDDR은 전력 소모를 줄인 메모리로 주로 스마트폰 등에 쓰인다. 인공지능(AI) 서버의 전력 효율을 높일 수 있는 데다 메모리 공급난까지 겹치면서 서버 업체들이 LPDDR까지 확보에 나선 것으로 전해진다. 가격 부담을 감수하더라도 메모리 물량을 먼저 확보하는 것이 AI 시장 주도권을 잡는 데 중요하다는 판단에서다. 중국에도 창신메모리(CXMT)와 양쯔메모리(YMTC) 등 메모리 업체가 있지만, 서버용 고성능 메모리의 제품 경쟁력과 양산 규모는 국내 업체에 미치지 못한다는 평가다. 중화권 CSP들이 필요한 물량을 자국 내서 조달하기 어려운 점도 한국 업체와의 장기 계약을 추진하는 배경 중 하나로 꼽힌다. 미국 빅테크가 먼저 장기계약으로 물량을 선점한 점도 중국 CSP들의 움직임을 서두르게 한 요인으로 보인다. 삼성전자는 글로벌 5대 데이터센터 고객과 계약을 마쳤고, 대형 고객 5곳과 추가 협상을 진행 중이라고 밝힌 바 있다. 업계에서는 이들 고객을 아마존웹서비스(AWS)·마이크로소프트·구글클라우드·메타·오라클로 보고 있다. SK하이닉스도 핵심 고객을 포함한 10여 곳과 LTA 협상을 마무리했다고 밝혔다. 중국 CSP들의 요청은 이 같은 기존 계약과 별도 물량인 것으로 전해진다. $AMZN, $MSFT, $GOOGL, $META, $ORCL AI 열풍에 메모리 확보 경쟁이 치열해지면서 메모리 업체와 고객사 간 장기계약도 확대되는 추세다. 계약 기간과 물량이 늘고, 공급 안정성을 높이기 위한 선수금 지급도 확대되고 있다. 삼성전자와 SK하이닉스의 내년 메모리 생산 예정 물량 가운데 70~80%가 글로벌 빅테크와 맺은 5년 장기공급계약 물량일 것이라는 게 업계 추정이다.
4
124
internet demand was bottlenecked by human access and time inference demand will only be bottlenecked by physical infra and electrons
Replying to @aaronburnett
Agentic demand is growing faster than I’ve ever seen anything grow
17
4
53
8,024
Yup, memory is now more than 50% of the book, way too much hate without grasping the fundamentals let's see
These kinds of tweets always surprise me if I’m being honest. I think part of it is that people missed the memory trade, but happy to break down the thesis One of the common themes at Token was agents. Yet when I pushed investors on how to express that thesis, many just said “AI x crypto,” which is a cop out answer If you want to be long agents beyond “agentic trading,” one of the clearest ways is through inference infrastructure like compute (GPUs) and context (memory) Training is mostly about how much GPU compute you have, but once you start serving those models memory becomes a much bigger bottleneck. Inference breaks down into prefill which processes the prompt and is generally limited by GPU compute, and decode which generates the response and is limited by memory bandwidth That’s why caching is so important. It lets agents reuse context instead of paying for expensive GPU prefill again. DeepSeek Flash charges $0.006 per million cached input tokens versus $0.30 uncached at peak rates. That’s 50x cheaper input pricing for reused context The bandwidth of memory also affects how many tokens per second you can generate. While capacity affects how much context and how many concurrent users you can serve It's gotten so competitve now that there are different memory tiers making different tradeoffs between bandwidth, capacity and cost Groq and Cerebras are optimize around extremely fast tokens per second on SRAM chip. Nvidia's Groq has 150 TB/s of memory bandwidth but only 500 MB of SRAM per chip. Since the capacity is so limited you need a rack that connects 256 of these chips to reach 128 GB of SRAM to be able to run the more modern models. Cerebras takes a different approach with a wafer design with 21 PB/s of aggregate memory bandwidth and 44 GB of SRAM, but the main point is that both of the these chips need to be massively networked to fit larger and larger models or more context as the model continue to scale in parameter count Then you drop to high bandwidth memory from companies like SK hynix and Micron which trades bandwidth for much more capacity. Nvidia's next gen chip the Rubin has 288 GB of capacity and 22 TB/s of memory bandwidth, versus the Groq configuration’s 500 MB and 150 TB/s That is roughly 7x less memory bandwidth but nearly 600x more capacity. Going further down the higharchy ram and nand based SSDs provide even cheaper capacity further down the hierarchy One interesting note is that DeepSeek has made a ton of hardware innovations in caching context on SSDs because of the export bans by the US on HBM, further pushing down expensive GPU prefill work. Nvidia has also acknowledge SSDs are a increasing important part of the stack and spin up a new platform called CMX which uses flash storage for reusable inference context. That adds inference caching to traditional storage demand and is one reason I’m bullish on NAND suppliers like SanDisk I think if you’re bullish agents, you should be bullish memory, because you want to be long context. In the theoretical limit, you’d want effectively unlimited memory and context, like the movie Her, where your agent knows your history, your preferences and everything you’ve worked on With paid AI adoption at only a few percentage of the world's population, and even fewer people using agents, my bet is that many of us will graduate from one agent to having multiple agents working for us, which will only multiply compute and memory demand On the supply side of things, Amazon CEO is saying AWS’s capacity blocks are sold out for 2026 and largely sold out for 2027 “There’s only so much memory, and I don’t think that that shortage ends in 2027” teapaw.org/firesidealpha/status/2… Memory suppliers are also securing multi year contracts. Micron has agreements extending into 2031, backed by take or pay commitments and substantial cash deposits. Breaking those contracts has financial consequences, and I think customers also risk future allocation. The point I want to drive home is that securing more memory capacity directly gives Labs the ability to support more concurrent users or larger contexts windows aka more revenue This is all while people are missing that earnings have gone up so much that the foward PE ratios are actually coming down. The results keep backing up the thesis while people keep calling the top. I believe we are just climbing the wall of worry I’m bullish on crypto as the future of finance, and I’m bullish on AI. I don’t think we need to mash them together and they don’t need to compete. Just be long both
1
4
178
The circularity of the AI capex problem If you are worried about supply of fresh equity, why are those companies who benefit from the spend of this new capital being sold? Why is everyone worried about dilution? Don't these projects have good margins with utility like demand than meta's business as a whole? There is a gap of a few years but tech investors have rarely been worried about things like that before. The only proper sense bearish take to explain a day like today is that there will be insufficient demand for compute in the future and that all this spending is on a white elephant. I haven't really heard many make this case. But I think it's the only thing that underlies these sort of moves. People doubt AI fundamentally and when they get asked to make a sacrifice for it (handing out capital), they flinch. It's not going to work, there's not enough demand. This makes selling across the AI supply chain rational. Companies like coreweave are the ultimate expression of this. Individually unimportant in the whole scheme of things being opportunistic extremely highly leveraged (not just financially either, but to input pricing and execution risk) players that will just have their supply absorbed by someone else if they fail. There pricing offers a barometer in the belief in this future demand. So, it's all just sentiment. And this makes sense. It's an enormous bet on something that is doubted by and is frightening to many and is many years into the future. Much harder than selling ads now. AI has to generate enough economic productivity to pay for itself and the margins that justify these capital raises in the first place. I think it will. I don't think most agree and they are just getting pulled along as beta. Even if it is successful in generating enough productivity to pay for itself (plus some juicy inference margin) you have many that are frightened of it in addition to doubting it. The corporate factory line worker is at risk just like they were to machines in the industrial revolution. This will prompt a response from government. How do you price that? By discounting the ROI on these investments? This would get you to the point of being worried about dilution. The trade is just too uncertain and needs so much capital that the only realistic way it fails in the medium term is that we become forever compute constrained and AI has a small percentage of the effect it could've had, taking down the labs with it.
11
8
4
103
25,521
"it was always going to be this way. we tried to warn the world."
Replying to @hinapansari
yes. it was always going to be this way. we tried to warn the world.
1
2
109