manch

@yuyy614893671 不用懂技術，只要懂華為就夠了。

周二美股开盘，半导体股票会不会血流成河？ 不懂技术，华为的这个公告到底含金量如何？ 懂的来说说？

@jbulltard1 Trump didn’t want to bankrupt his casinos either. Sometimes reality has a say in things.

LOL you bears think this guys gonna sink markets to fight iran?

@seanpianka @gave_vincent China doesn’t have enough internal demand. It’s trying hard to export supply and grab demand from other countries. It’s fine if it’s from a smaller economy. Not fine for China’s size. Hence the rise of global protectionism. China’s fate will be exactly like Japan’s in the 90s.

I recently read @gave_vincent’s piece “The Chinese Equity Bull Market Quandary,” and it gave me a useful frame for something I’ve been experiencing firsthand while traveling in China over the last few weeks, something I don’t think Americans fully grasp until they spend time here: The “industrial deflation” story is not just a macro chart or an equity-market debate. It is something you physically experience. I’ve been in China since early May, truly the best time to visit, and walking around here with USD savings / US income assumptions inside this economy feels like discovering an entirely different consumer frontier. A massive, spotless mall. Morning dim sum with 10+ items for ~¥100 / $14. QR-native ordering. No tipping. Fast service. Clean public space. Portable power banks everywhere. Tailors who can turn around alterations in 20 minutes. Restaurants and shops competing brutally on speed, price, quality, and reviews. And it’s not just inside the mall. A 20-minute premium DiDi, effectively the local equivalent of a nicer Uber, can be ~$5. The roads are full of green-plate EVs, bicycles, and electric scooters from delivery drivers, with tons of people walking and comparatively few blue-plate ICE cars, whose license plates alone can cost around $10k in cities like Shanghai. The whole thing feels like a high-throughput urban mobility system: cheap rides, dense foot traffic, electrified transport, and services that can reach you almost instantly. In the US, a comparable meal in a major city can easily be $100+ after tax and tip, often with worse service, worse execution, and more friction. In China, the same category of experience can feel cheaper, faster, cleaner, and more operationally competent. This is the part that gets lost in abstract discussions about “China overcapacity” or “deflation.” From the consumer side, that overcapacity often manifests as abundance: too many restaurants, too many malls, too many delivery riders, too many EVs, too many manufacturers, too much competition, too much pressure to improve. Of course, this is not how the average Chinese worker experiences prices. The arbitrage is being a foreigner with dollar purchasing power inside a lower-cost service economy. But that is exactly the point: the gap between US nominal incomes and Chinese urban consumer prices is enormous. It also changes how you think about Chinese industrial companies. The same system that produces cheap, fast, high-quality urban services is also producing firms that are no longer merely “cheap China” competitors, but increasingly world-class operators in EVs, batteries, drones, hardware, solar, logistics, robotics, and consumer electronics. The US still has the higher-income, higher-trust capital-market machine. But China has built a consumer/industrial machine where the lived experience of abundance is hard to explain to people who have only seen it through US media, macro commentary, or geopolitical framing. My prior view of China was too abstract. Being here makes the deflationary industrial economy feel less like weakness and more like a brutal, hyper-competitive operating system that exports falling prices and rising quality to anyone able to access it. The uncomfortable part is going back to the US and paying 5–10x more for slower, worse, more fragmented versions of things that here feel ordinary.

@MoodyWriter13 Great write up. Do we know the real reasons why Bosch ended their partnership with Ceres?

Fuel cells for AI data centers are, a genuine structural trend, not a short-lived hype cycle, no doubt. The power demand of data centers is growing faster than the grid can keep up, new gas turbines have lead times of five to seven years, and securing a grid connection often takes years as well. Fuel cells solve exactly this problem. They can deliver clean, permit-friendly power directly on-site within weeks to months. What matters, though, is being honest about why they win. They do not win on pure electricity cost. Against highly efficient combined-cycle gas turbines (CCGT), fuel cells are clearly more expensive at roughly 1.5 to 2.5 times the cost per MWh, while they are roughly on par with or slightly cheaper than open-cycle gas turbines (OCGT) due to ~15–20% lower fuel consumption versus OCGT’s ~35–42% efficiency, and they are significantly cheaper than diesel generators in continuous operation, as diesel is a high-cost fuel intended only for backup use rather than baseload power.They win on availability and permitting. This is a scarcity premium, not a permanent cost advantage. As long as grid bottlenecks persist, the business is excellent. That is the key point investors need to keep in mind. And one more thing matters greatly. Even among fuel cells themselves, efficiency determines the ultimate power cost, because the company that burns more gas to produce the same amount of electricity will inevitably generate more expensive power. This is where the field starts to separate into winners and losers. $BE is the undisputed market leader and, operationally, the strongest of the three companies. The advantages are obvious. mature and highly efficient SOFC technology, a massive backlog supported by real contracts (Oracle, AEP, Brookfield, and most recently Nebius), proven gigawatt-scale manufacturing, native 800-volt architecture for Nvidia racks, and delivery timelines that are faster than customers can construct their buildings. Bloom is effectively sold out on the demand side for years. The bottleneck is production capacity, not orders. That is an enviable position to be in. At roughly 60% efficiency, Bloom also sets the benchmark the others have to compete against. The drawbacks are not operational but valuation and supply-chain-related. At around an $87 billion market cap, the stock is priced for perfection. Virtually all of the expected success over the next four to five years is already reflected in the valuation imo, leaving little to no margin of safety. There is also a concentration risk: Bloom depends on scandium, which is roughly two-thirds concentrated in China. Fantastic company, wrong price. Not a buy for me. I should add that I already identified Bloom and its advantages back in 2024 and was invested. Unfortunately, I sold far too early at around $13.... Ceres Power has, in my opinion, the most attractive business model. Ceres manufactures nothing itself. Instead, it licenses its technology to industrial partners and collects high-margin royalty streams. The model is asset-light, the balance sheet is debt-free, margins are structurally attractive, and there is no scandium exposure because the technology relies on inexpensive stainless steel. The key insight is that Ceres can potentially earn royalties from almost everyone except Bloom: Doosan in Korea, Delta in Taiwan, Weichai in China. Weichai claims to have already built the world’s largest SOFC production facility in China,but installed capacity is not the same as deployed systems, and Weichai has already failed to commercialize successfully with Ceres on multiple occasions in the past. Until there are real delivery and deployment figures, even the “largest facility” remains more of a promise than proof.The technology itself is theoretically even more efficient than Bloom’s. That matters economically because higher efficiency means lower gas consumption and therefore potentially the cheapest electricity of the three, assuming the performance translates from theory into commercial reality. If the product gains traction, Ceres can participate across multiple continents simultaneously without ever having to build a single factory. But that still needs to be proven. The disadvantages are just as real. The model remains largely unproven. Very little has actually been shipped so far, and the only meaningful order came from a Doosan-affiliated party and was tiny. Bosch’s withdrawal in early 2025, after investing more than €1 billion, demonstrated how vulnerable a pure licensing model can be. Ceres is entirely dependent on the strategic decisions of its partners and does not control its own commercialization. On top of that, the stock is far from cheap. To me, Ceres is the opportunity with the highest upside potential, but it only becomes truly compelling once a real third-party large-scale order from Doosan or Delta finally validates the model. $FCEL is the least convincing of the three in my opinion. To be fair, there are some positives: an initial data-center reference customer appears to be emerging, the company has more than two decades of operating fleet experience with real-world field data, and its carbon-capture capability is something neither Bloom nor Ceres currently offer. If execution succeeds, the upside from such a small base could be substantial. But the disadvantages clearly outweigh the positives for me, starting with the electricity cost itself. FuelCell’s MCFC technology is by far the least efficient of the group: roughly 47% efficiency versus Bloom’s approximately 60%, implying approximately 25 to 30% higher gas consumption for the same amount of electricity generated. Translated into power costs, FuelCell therefore likely produces electricity roughly 15 to 30% more expensively than Bloom, precisely in the kind of 24/7 data-center environment where continuous gas consumption maximizes the impact of that efficiency disadvantage. And that is merely the lower bound of the problem. The company has been burning cash for years. Its one genuine differentiator, carbon capture, is also aimed at a different market than the one that matters here. Carbon capture might become relevant for the Hyperscalers, when they are reminded of their climate neutrality targets. Within the data-center market itself. FuelCell’s competitiveness likely depends on a temporary premium on speed and deployment time rather than a sustainable cost advantage in electricity generationArguably, it has the opposite. For me, this is a clear pass. One more dimension worth understanding is the materials base, because it quietly shapes both cost structure and supply-chain risk. All three technologies share one major advantage over PEM fuel cells: none of them require platinum or other precious-metal catalysts, thanks to their high operating temperatures. The differences lie elsewhere. Bloom's SOFC depends on scandium for its electrolyte, which is its one genuine raw-material vulnerability given the roughly two-thirds Chinese concentration of supply. Ceres has the most robust material profile by far:. Its metal-supported design places a thin ceramic layer on ordinary stainless steel, meaning no scandium, no exotic bottleneck, and a cell that is mostly cheap, widely available steel, this is a real and often overlooked part of the asset-light, low-cost thesis. FuelCell's MCFC, by contrast, is nickel-based, with both electrodes made of porous nickel and an electrolyte of molten alkali carbonates. Nickel is cheap and abundant, but the corrosive 650°C carbonate melt is precisely what drives the limited stack lifetime and demands expensive corrosion-resistant steels. In short, Ceres wins on material simplicity, Bloom buys top performance with a scandium concentration risk, and FuelCell relies on cheap nickel but pays for it with corrosion and shorter stack life. My conclusion: Bloom is the best company with the best product, but too expensive. Ceres has the best business model and the highest potential, but right now also very expensive, and it still has to prove itself commercially. FuelCell is the weakest of the three imo. This reflects only my personal opinion. This is the result of one day of research, but I have effectively been following this sector for about 3 years. Under normal circumstances, I would research more deeply and over a longer period before commenting on something like this.If anyone has more information, I would be happy to hear it.

@mukund I also don’t understand the hate Meta is having. Zuck pays extremely well and these people have been getting insanely good TC. Getting paid 300k for not doing much is not anyone’s birth right.

Controversial take. $META If in 2 years or so Meta starts to hire again and posts jobs. They will still get a ton (1000+) people applying for each job. This sentiment will turn. The people saying “I hate Meta” will say “Ok. Pay me well and I will hold my nose and work there”.

@HeLLaTaLL510 Walnut Creek erasure

Concord = Hayward. Richmond = Vallejo. Fairfield = Antioch. Suisun City = Pittsburg. Oakland aesthetic + Marin Culture = Berkeley. El Cerrito, San Pablo and El Sobrante don’t exist. Those are all parts of Richmond. Emeryville also doesn’t exist. It’s just where Oakland and Berkeley go to Target and IKEA. Daly City is a neighborhood in San Francisco. San Jose is the worst parts of modern San Francisco culture combined with the worst parts of Los Angeles Suburban sprawl. Pinole, Hercules and Rodeo are all different neighborhoods in one city. Crockett, Port Costa and downtown Martinez are also the same city. That is all.

@christinelu Isn’t Arcadia where the rich and powerful men from China keep their mistresses?

Arcadia is the city next door to me in Pasadena. Back in the day it was known as the Chinese Beverly Hills. Lots of Chinese government officials and tycoons along with their mistresses in Maseratis.

@BourbonCap What’s up with the CEO’s sunglasses? They make him look like a sleazy used car salesman, which he is sort of TBF.

During the last earnings call, $NOW CEO said they were going to reduce stock based compensation. I thought he was joking, but it looks like he's gonna do it They gonna get cash instead

@citrini So far all the investments are driven by real revenue. Anthropic and to a lesser extent OpenAI are still growing revenue exponentially by the month. As long as we don’t see the equivalent of “eyeballs instead of dollars” circa dot-com late 90s we still have ways to go.

People keep confusing a bubble with “stocks go up and get overvalued”. A bubble is when when a prevailing trend and a prevailing misconception about that trend interact reflexively, each reinforcing the other until the gap between perception and reality becomes unsustainable. A bubble is not when everyone realizes that right now every iota of AI demand eventually, at some point upstream, must move through memory OEMs. Nor is it when estimates continue rising because things are better than expected. And it’s not just when stocks trade expensive to historical valuations. The reason behind the moves in the AI infrastructure layer so far have been simply that we don’t have enough. They’ve been driven by the fundamental reality more than the perception of the future. It’s why the bulk of the most bullish parts of this cycle have been lumpy and centered around earnings season when companies uniformly come out and confirm there’s still not enough. In the bubble, the reality is driven by the market - not the other way around. Everyone keeps saying “people are gonna freak out if it’s not a bubble!”. I think that’s silly, we have a transformative new technology that needs crazy capital to fuel it coming to fruition, that has and always will result in a bubble as long as we have financial markets. But if you want to call the top in a bubble, you need a much stronger view on what the misconception is and what negative catalyst forces broad perception to align with realizing it than you do on valuation.

@aleabitoreddit $TSLA all-in investors entered the chat

I have high conviction that majority/full port $IREN investors are the dumbest people you’ll meet in this world.

@vshih2 “Dropping out of a window” with Chinese characteristics.

What happened here? The guy looks pretty healthy

@daniel_p_gross What was the impact on wages of those more specialized telephone operators who survived automation? Did their wage rise or fall in real terms post automation?

Perhaps useful to point out that history rhymes: this is effectively what happened with telephone operators, 100 years ago. Young, entry-level operators, whose work was mainly connecting local calls--the simplest version of the job--and not much else, were wiped out by automatic call switching. More senior operators, who had a wider range of tasks and did more complicated work like information service, emergency service, and long distance calls, were not as affected. Even when local telephone operators were effectively eliminated across AT&T's network, the others remained. Much had to do with the complexity of the work and how entangled it was with the rest of the organization. Extended discussion in work with @jamesfeigenbaum in @QJEHarvard and Management Science: academic.oup.com/qje/article-ab…, pubsonline.informs.org/doi/10.1287/mn… Lots of credit to @joshgans for helping us sharpen some of these ideas in this work as editor of the ManSci paper

@covie_93 The most important thing to a White Christian isn’t the Christian part. It’s about being White.

How the hell did donald trump manage to convince Christians that he is one of them?????

@gtconway3d “Worst attorney general in history”

"Pam Bondi actually had decent reputation. She was a real prosecutor and a real DA. I don't think she was a rocket scientist, but she was respected in her own way. And now she's basically going to be remembered forever as being the worst attorney general in history."

@avidseries Glad we can use the word retard again. These MAGA clowns are all retards.

@adamscochran We need sleepy joe back! Just so we can actually get some good sleep at night.

Not a great sign. Hopefully this is just for authorizing rescue operations… But Trump being quiet and actually working, is seldom followed by good news.

@Robert_E_Kelly I think we may need to factor in Trump’s cognitive decline. The guy doesn’t have all his marbles anymore and it shows.

Prediction: Hegseth is gonna be the fall guy for the war 72% of Americans oppose a ground war, which is the only way Trump can ‘win’ this thing now. Everyone is upset about gas prices & market decline. Trump’s approval rating is 35% He’ll bomb for a few more weeks, hoping for a break. He won’t get it But his political instincts are good enough not to invade. The GOP would get hammered in November if a ground war went badly, and that would mean Congressional investigations next year When this whole mess ends in the next month or two, then come the firings

@DashaBurns @alexanderburns Nutlick is not licking hard enough. 🤣

NEW: Trump is considering more changes to his Cabinet, according to four officials. He has expressed frustration and disappointment with Commerce Secretary Howard Lutnick and Labor Secretary Lori Chavez-DeRemer. "He's very angry and he's going to be moving people," an administration official told @politico. No final decisions have been made on Chavez-DeRemer and Lutnick, and Trump has contemplated firing people and then backed off before. My latest reporting: politico.com/news/2026/04/0…

@conorsen Netscape IPO’ed in 1995 and officially kicked off the dotcom boom. So…… we are still early?

Bloomberg reporting that Anthropic is weighing an IPO as soon as October:

@amconmag Rogan is low key one of the dorks. 💀

Joe Rogan says there are “a lot of dorks” in MAGA: “That phrase sucks. America is great. Make America greater? I’m down. But MAGA, and then it becomes a movement of a bunch of dorks? A lot of them are these really weird, f-cking uninteresting, unintelligent people.”