X（推特）上的财经、科技与宏观账号
A few assumptions behind the charts, which compile figures found in the public domain and published research. Global forecasts use a 52% US share, North American forecasts 97.5%, and ex-China forecasts 68.4%. Where a facility-power conversion is needed, the assumed PUE is 1.20;
Something that blows my mind is the variance of the estimates we're still seeing for data center capacity. We’re already in mid-October and the selected estimates for 2026 US data-center growth still span roughly 10–21 GW on an IT basis. A range more than 2x for the year we’re https://t.co/VL1dm0sQw
Why I believe Muse (and Muse-like agents) are game changers for SMBs $META full piece: https://t.co/rQBpOpGBKj https://t.co/VbDyCAJj9J
Intel would make more sense I'd have thought, given Qualcomm's recent acquisitions. But Jim continually says he has no plans, so 🧂 https://t.co/4qNcjHwala
Super interesting: "Anthropic's Claude Fable 5.1 scored 41% in Claude Code and 61% in OpenAI's Codex. It made 28 tool calls at the median in one and 66 in the other." "No model did best in its maker’s own harness" https://t.co/BONmxl9hiE 引用 @chamath：Model Benchmarks Are Worthless How can a model
Love people opening up the convo to debate/learn in public and Matt is great at it. Sharing for broader awareness. I'll leave the chip-based GW forecasts to others, but on power, I've presented several frameworks for thinking about US IT capacity and the very real limits that https://t.co/fTp3ZzdW5
Very interesting to see which analysts are going to Semicon West and which ones are going to OCP.
Great episode. Meta's $100B+ AI bet is built around PERSONAL superintelligence: an agent that knows your context, remembers it and gets things done for you. Zuckerberg and Wang on Cheeky Pint: >They see room for a consumer-focused frontier lab while much of the industry's 引用 @collision：Cheeky Pi
DTE与Google在Van Buren Township签订1 GW、期限为20年的电力合同。作者称，密歇根州10月1日的相关命令披露了超大规模云服务商的最低电力承诺，并提到为公用事业提供下行保护的基础设施投资安排。
Sure... Random train of thoughts on markets now, optics like $LITE / memory + $IREN / Anthropic. On Memory: - Demand visibility on $MU / Samsung / SK Hynix extend into 2030-2031. Samsung has LTAs covering 70% of their capacity into 2031 (likely with SK Hynix type price 引用 @Wolfex_Yeat：@aleabi
高通CEO: “什么是AI手机？当OPENCLAW面世时，AI手机的未来就清晰了” 💯💯💯 引用 @TJ_Research：以前我们买手机看哪家的照相功能好，未来我们买手机看哪家的端侧AI强
COLUMN: “Buy to the sound of cannons, sell to the sound of trumpets.” The ancient stock market maxim resonates in today’s Middle East. While the US and Iran fight, Big Oil is heading into neighboring Iraq. @Opinion https://t.co/6rcxadx1Em
NEW ODD LOTS: @tracyalloway and I talk to @So8res to get the actual details on how AI would actually kill us all https://t.co/GBuT46fmSA
As a structural bond bear, I actually don’t find this recent bear surge in bonds surprising. The yield curve is still flat, nothing is broken. It’s almost expected. But now, the more people talk about it, the more I might become neutral on bonds. https://t.co/uOrymG3qRA
The AI boom is increasingly shifting from parabolic to linear growth. Most investors are totally unprepared for that shift. This week, free on the blog app. https://t.co/EUji9tuxSA https://t.co/qQk0lvaJcR
2000年互联网泡沫破裂，科技股持续了很长时间的低迷。核心不是泡沫，而是2000年前后的互联网都没找到商业模式，从00年一直到06、07年 引用 @market_sleuth：“Too big to fail. Active trading isn’t needed anymore. In fact it’s more of a waste of time.” Those were the words I heard as I was speaking with a finance person yesterday who is way smarter than I & who mana
We are coming off a curious week of economic and financial dualism ... and are now looking at a packed one featuring US inflation data, lots of other global releases, and the IMF-World Bank annual meetings. For details, please use one of the following links to read my Weekly Look https://t.co/iuRLp9
I hear all the time that the semis & hyperscalers see plenty of AI-related demand ahead, with no connection to who will pay to meet those lofty expectations. In short, it would require all non-tech companies to bet their entire free cashflow on AI by 2030. Good luck. https://t.co/tllPSskzj8
As this chart (h/t @Mike_Weilandt) shows, periods of CNY depreciation/ weakness have consistently been associated with Chinese export outperformance. (the IMF China team ignored this from 22-24, and as a result missed the jump up in China's surplus) Time to get back to BoP https://t.co/ow8GUgbld2
A chart that says more than a thousand words The acceleration in China's export volume growth clearly followed the RMB's depreciation. Sometimes international economics is simple (even if the PBOC and a few others want to fuzz it up) https://t.co/dY7aZH8yUB
3/ She prefers "smaller, more frequent" moves to avoid bigger hikes later. Her neutral-rate estimate is among the Fed's highest, and she says if she's wrong, it's too low. Also, higher yields are "absolutely not" a substitute for Fed hikes https://t.co/Amj2pqRriQ
2/ Her take: Yes, rate hikes won't open the Strait of Hormuz or immediately slow the data-center buildout. But they might make other firms "think twice" about expanding in this environment, leaving scarce parts or supplies to data centers and easing pressures at the margins.
Mag7应该叫“Mag700”，这个观点还挺有意思。核心是当前美股“缺乏广度”（主要由Mag 7等少数大票推动）不必过度担心，原因有两点： 引用 @EricBalchunas：The Stock Market Is Currently Out of Breadth But It's No Reason to Worry.. new from me on the latest worry being hurled at investors from the columnist class and what they aren't telling you: 1) over past 100 ye
More on how the size of the question ahead has gotten so large here on the blog app: https://t.co/bifKj6KlvA https://t.co/Zd3Ku1pPin
红杉两年前提出的6000亿美元问题，如今已变成3万亿美元问题。相关支出规模扩大之际，实际收入仅增加略多于1000亿美元，前瞻性需求与现实收入之间的差距已扩大。
Barry Eichengreen sees 2 different scenarios 1️⃣ Favorable: gradual erosion of US Dollar along with creation of a diverse global multipolar monetary ecosystem 2️⃣ Worrisome: sudden collapse of USD before other systems are mature and robust: global liquidity shortage @nabe_econ https://t.co/OlUuV5
I really enjoyed this chart and history book by @MebFaber For people who like macro and market history, it’s a really cool physical resource. https://t.co/V6n5Iqv5Ld
Remember two weeks ago when risk premium sceptics were looking at ACM and saying risk premiums were falling not rising. Where are the sceptics today. Using one model is stupid but thinking risk premiums were falling back then was even stupiderer https://t.co/KnWYUsuzHk
5/5 Diversification across company names and diversification across investment themes aren't necessarily the same. Concentration risk means half the index is exposed to the AI theme. Half is not. The choice is not necessarily to go all-in on AI or all-in against it. It is
4/5 Concentration is only part of the story. Over the past year, the rolling 6-month correlation (green) has turned slightly negative. Another sign that the headline index conceals two different market experiences. https://t.co/HDiCAqPo3a
This is a funny tweet. Because do people actually associate @pmarca and Thiel with having been prescient about AI? The former I associate with b2b SaaS. The latter with kind of Silicon Valley woo ideas. With Elon it’s more complex, but Brian has a substantive point to make. 引用 @brianchau57：Bay Area
On the flipside, it’s not obvious to me that the doomers have better arguments. But they’re definitely better at arguing, having sharpened the blade on message board the last 15 years.
It would help if all their critics didn’t come off publicly sounding like this guy https://t.co/L6Xntz6Ge1 引用 @MattZeitlin：i still don't really understand how AI can cause human extinction BUT i do become more convinced the doomers/ASI-is-coming people are right when i read the arguments against th
And its standard fiscal advice would more or less result in a much bigger current account surplus -- i.e. it wants China to rely more on exports (think the advice used to be a 7 pp consolidation over 10ys, though that good downplayed for s-term cyclical reasons in the last A4)
The IMF's standard approach has a hard time with a country with massive savings and a public sector that has real assets as well as large debts -- by all of the IMF's standard models, China's large augmented fiscal deficit should have generated a current account deficit ...
作者认为，将中国公共部门债务整体合并分析是一个分析错误，因为这会掩盖中央政府明显的财政空间：利率低、债务占GDP比重低于35%，且拥有大量金融资产。引用内容称，中国公共债务占GDP比重为122%，高于疫情前的74%；扩展口径财政赤字为GDP的10%，偿债支出占中央政府预算的19.2%。
Difficult for the fund to do but I hoped the fund would use this excellent paper to rethink its fiscal advice and move away from the augmented debt/ augmented deficit and treat the LGFVs as financial institutions that might give rise to a bailout cost not as ordinary government
But key points in the essay were asserted not demonstrated -- this one for example "As cheap imports undercut traditional producers and provide low-cost inputs, labour and capital have shifted towards more productive industries" what industries are expanding on a scale able to
I do want to know more about Europe's increased production of electronic components (The firms, sectors and scale)
FT's Tej Parikh really stretched the industrial production data here to fit into into a narrative around market competition, creative destruction and moving upmarket -- the rise in pharma production? Ireland winning a game of tax competition the rise in optical equipment? https://t.co/me7tctXCBy