Tanay (@TS66555) profile picture

Thematic & long-term Investor. Aim is to share my research, not recommend specific stocks. The longer version is at: https://teapawt.net/t.co/BsO212HoI2

Joined July 2020
TRANSOCEAN - attracting a bid from the grifters - $rig getting bought by politicans ! The long consolidation will end after mid-terms. I expect a very strong ramp in $OIH into EOY. $TDW $NE $SDRL $val $borr
BREAKING: We just caught another interesting trade. Representative Lisa McClain just filed a purchase of stock in a company called Transocean, $RIG It's a Swiss offshore drilling company. This is the first time we've seen a member of Congress trade the stock in over 5 years.
8
45
4,852
Alibaba CEO says new Zhenwu V900 is China’s most powerful AI chip, with 3x predecessor performance
8
12
1
103
27,884
$BABA UNVEILS NEW AI CHIP, PLANS 5T-10T PARAMETER MODELS AND 20GW+ OF DATA CENTER CAPACITY Alibaba unveiled its new Zhenwu V900 AI chip, designed for both model training and inference, with performance roughly 3x higher than its predecessor released just four months ago. The V900 is expected to enter mass production and commercial release in Q1 2027. Alibaba says clusters built around the chip can scale to as many as 500,000 accelerators for training and running frontier AI models. Alibaba is also targeting more than 20GW of global data center capacity by 2032 and plans to launch its new Yitian 720 and 730 CPUs for agentic AI workloads in 2027. Qwen4 is currently in training and expected soon, while future Qwen generations are expected to scale to roughly 5T-10T parameters, versus 2.4T for the current Qwen 3.8 Max. Alibaba has committed more than $53B over three years toward AI infrastructure and capabilities.
13
42
15
347
30,330
$BABA is the Chinese $NVDA
Alibaba unveils new AI chip it says is China's most powerful and targets 20 gigawatts of data center capacity by 2032. bloom.bg/46H35gf 📷️: Qilai Shen/Bloomberg
4
4
89
6,846
*ALIBABA UNVEILS NEW AI CHIP IT CALLS CHINA'S MOST POWERFUL
China already dominates open models. Soon it will dominate the hardware that powers them.
27
76
3
503
143,539
Just too much FUD and short-term focus on the Chinese tech platforms after recent earnings. So putting out some long thoughts on them, especially given I have held these for a while now and plan to hold out long-term. The earnings were rather subdued no doubt, but one has to take a step back and also contextualize that against one of THE worst macro backdrops in China in decades combined with a highly, highly dynamic AI environment that requires these leading platforms to spend (just like their US counterparts are). If one takes at least a 3 year view, these businesses are undervalued. I expect all these three stocks to be trading much higher vs current levels in the next 12-18 months. Would welcome constructive thoughts, for or against, my opinions / reasoning. ----- Alibaba -- The old business funds the new business. The attraction is that you have a large, cash-generative commerce franchise funding what is increasingly one of China's most credible full-stack AI platforms. You do not need Taobao/Tmall to return to its old growth rates (and it likely won't given macro + competition). You mainly need commerce share and monetization to stabilize and remain strongly profitable. That cash engine can then fund Cloud, AI infrastructure, Qwen, proprietary T-Head chips and higher-layer MaaS / applications. That is already becoming tangible: Cloud has reaccelerated, AI-related revenue has been growing triple digits for multiple consecutive quarters, and Alibaba has built genuine silicon-to-application vertical integration. The current earnings look artificially messy because Alibaba is deliberately spending heavily on AI and quick commerce. The key 3-5 year question is therefore not whether near-term margins fall but whether Alibaba emerges with a still-dominant commerce franchise plus a much larger, profitable Cloud/AI business. Baidu -- The balance sheet pays for most of the stock. Baidu is the most straightforward asset-value dislocation of the three. At ~US$32bn market cap, strip out the substantial net cash / liquid assets and listed investments, and you are left with only ~US$9-10bn for the operating businesses. Then I simplify the operating businesses further: Legacy Search: Structurally impaired (as ad dollars are shifting away from Search to Short-form video). I value it at zero. If Baidu eventually monetizes AI Search successfully, great -- that is upside. Apollo: Credible technology + operating scale, but still too early financially. Put ~US$1-2bn against it using the Chinese listed peers like Pony.ai as a rough anchor only. That leaves what actually matters as of today: AI Cloud + Kunlun. These are both exposed to structural 5–10 year growth markets, and they are already real commercial businesses. Cloud grew 50% YoY in 2Q26 and GPU Cloud 283%; Kunlun has meaningful shipments and external customers. My own reasonable valuation for Cloud + Kunlun is ~US$18-20bn today, vs. an implied market allocation of only ~US$9–10bn after the other assets. And that still gives effectively no value to AI Applications / AI-native Marketing. Tencent -- The core business is cheaper than the headline multiple looks. Tencent is basically a high-quality operating compounder with a large investment portfolio sitting on top. The reported investment book is ~RMB875bn; even haircutting it quite heavily, I can underwrite ~RMB650bn. Strip that from EV and the "core" Tencent operating businesses are only around 9–10x 2026E non-IFRS operating profit, rather than the ~12-13x headline number. And what you are getting at 9-10x is not a melting-ice-cube business: Games, WeChat, Ads and FinTech are still growing and generating very high returns, with AI increasingly improving monetization across those existing businesses rather than requiring Tencent to invent a completely new earnings engine. Importantly, the investment portfolio -- PDD, Spotify, Epic etc. -- is then additional NAV, and actual disposals such as Kuaishou in July 2026 show that at least parts of it are monetizable rather than purely paper value.
5
287
Alibaba Group Holding founder Jack Ma has bought more than HK$600 million worth of the company’s Hong Kong-listed shares on consecutive days, according to people familiar with the matter, a show of confidence in the firm’s long-term prospects for artificial intelligence.-SCMP #ALIBABA $BABA
5
16
1
190
28,453
Alibaba's QwenWork ranks first in Jefferies’ evaluation of 8 global AI agents, scoring 95/100! 🎉 Recognized for industry-leading agentic and engineering capabilities, QwenWork stood out as the most consistent performer.
25
22
2
167
971,398
Chinese bros with inside info running Tencent to the ground 🤦‍♂️Not complaining, it is an amazing business and the lower the multiple, the better it is.
5
264
100% agree. Chinese LiDARs, both Hesai and Robosense, will continue to scale and dominate globally in the years to come. This is a growth industry, both are scaling rapidly, enough to go around. OUST and the like will fundamentally suffer for the most part.
This whole post is exchange is a banger. Case study of how: (a) The dirt cheap robosense is used orders of magnitudes more than the richly valued oust, (b) The statement that Robosense is low-end lidar v HSAI is a mischaracterization. I believe HSAI is more dominant for traditional Lidar use like autos and robotaxis. Robosense actually has an edge in created an integrated component that for these more advanced type humaniod (or dogoid?) forms - where they are integrating lidar, camera, motion sensors and software to solve the whole vision problem. This is great because there are thousands of robotics programs - and very few of them want to support a whole vision team. If you want exposure to lidar within humanoid robots, Robosense is much more pure and continues to get more pure (v both OUST and HSAI) by the day as they are heavily leaning into this use case. (c) Robosense has more access to the west. HSAI is pretty integrated with Huawei and itself has some restrictions / potential restrictions in the west. Robosense does not have this coupling with Huawei nor the same restrictions concerns.
4
311
Semiconductor stocks are quietly running the S&P 500 show right now. Over the last 6 months the index is up roughly 8.3%. A handful of chip and memory names have delivered the majority of that move. Look at this Bloomberg chart of the biggest contributors (in percentage points): • Micron $MU alone added ~1.4 points, the single largest contributor of any stock in the index, after skyrocketing 188%. Memory demand (especially HBM for AI) is on fire. • AMD $AMD came in 2nd with ~0.9 points on a 158% rally. • Apple $AAPL and Intel $INTC each chipped in ~0.8 points. Intel’s 205% surge is especially wild after being written off last year. • Broadcom $AVGO Nvidia $NVDA and Sandisk $SNDK each added ~0.4 points with returns of 9%, 5%, and a staggering 395% respectively. The pattern is clear. Even though the S&P looks broadly healthy, it is really being powered by a concentrated group of semiconductor and memory companies. Remove these names and the broader market return looks a lot more modest. Data centers need massive amounts of memory and chips, and the winners (so far) are printing. Investors should keep a close eye on this group. #Semiconductors #AI #Stocks #Investing
5
1
14
4,180
$TSM Q2 EARNINGS DOUBLE BEAT 🔥 EPS: $4.31 vs $3.77 est REV: $40.20B vs $39.76B est Gross Margin: 67.7% Operating Margin: 60.3% Net Profit Margin: 55.6%
65
179
47
1,491
287,984
CXMT on Trade.xyz. Good indication of where the market is likely to send it once it starts trading. It was listed at $5 (~4.5x the IPO price). Won't be surprised if this trades up to $10 in the immediate aftermath of the IPO driven by euphoric Chinese retailers.
2
125
Tanay retweeted
I would be looking to invest in $CCXI (or any western robotics company) with an edge in production - regardless of where they stood in terms of raw technical performance. One problem with $CCXI is they have practically no hard numbers on production. The hardest number is in a re-occurring footnote. "Note: (1) Reflects customer orders for Digit v5, as of May 2026, representing potential multi-year value expected to be realized over time, subject to the realization of certain contractual milestones, and relates to 1,000 Digit v5 robots with three-year term RaaS contract, which includes warrants issued to purchaser vesting proportionately to robots deployed; figures are not a measure of current period revenue." 1,000 bots over three years That's trash compared to numbers already coming out of China. It's simply not close to being competitive.
5
1
20
3,824
Ofcourse, everyone on X suddenly has an opinion on Gavin and Satya's latest posts / views 🤦‍♂️
3
73