DAVID MARIUZ/iStock via Getty Images It’s been a little over a year since the last time I covered Luxfer Holdings ( LXFR ), so I thought I’d give an update and see how the company has performed over the last while. The management has good plans for returning to growth, but unfortunately, I would like to see some results before updating my stance. By the Numbers Looking at the company’s top-line pe...
DAVID MARIUZ/iStock via Getty Images It’s been a little over a year since the last time I covered Luxfer Holdings ( LXFR ), so I thought I’d give an update and see how the company has performed over the last while. The management has good plans for returning to growth, but unfortunately, I would like to see some results before updating my stance. By the Numbers Looking at the company’s top-line performance for the latest quarter, we can see that there is a marked decline. Revenue came in at around $84m, down 13.5% y/y, which also missed estimates by $600k. Let’s look at the revenue breakdown in more detail to see the main reason the top-line performance was so underwhelming, to say the least. The company is kind enough to provide us with not only the breakdown of its revenue segments, which now only have two since divesting the Graphic Arts segment earlier, but also with end market performance. Below, we can see that the performance across Gas Cylinders and Elektron was quite disappointing, with Elektron’s performance pulling the company down considerably. Author If we dig even deeper into the end-markets, we can see that the reason Elektron performed so badly was due to the Specialty Industrial end-market taking a massive hit. Author Also, the performance was amplified because last year, the company was still reporting its Arts segment, and that was, as you can see, $6.5m in revenues. Excluding Arts, the top-line performance was -7.2%. Elektron sales declines were due to overall lower volumes across end markets, and nothing else in particular. Gas Cylinders managed to maintain their performance with stable volumes and even stronger specialty industrial demand. Let’s move on to the company’s profitability and efficiency. Gross margins improved by quite a bit, around 400bps. The company doesn’t attribute the increase to anything specifically, and only said it is the result of “positive pricing and cost discipline”. So what we can take from this is an improved price m...
China’s sharp move away from deflation last month is a welcome development for Beijing, although it comes on the back of higher energy costs caused by the Iran war. Producer prices rose 2.8% from a year earlier, the fastest since July 2022 and higher than all estimates in a Bloomberg survey of economists, whose median was 1.8%. Consumer inflation unexpectedly climbed to 1.2%. “China has officially...
China’s sharp move away from deflation last month is a welcome development for Beijing, although it comes on the back of higher energy costs caused by the Iran war. Producer prices rose 2.8% from a year earlier, the fastest since July 2022 and higher than all estimates in a Bloomberg survey of economists, whose median was 1.8%. Consumer inflation unexpectedly climbed to 1.2%. “China has officially exited deflation,” said Carlos Casanova, senior Asia economist at Union Bancaire Privee in Hong Kong. China had been trapped in a deflationary spiral since late 2022, as a manufacturing glut and sluggish domestic demand led to intense price wars. Price increases in mining and processing of non-ferrous metals, as well as crude oil, contributed to last month’s jump in factory inflation. Economic damage from the conflict in the Middle East will likely feature prominently in talks this week between Donald Trump and Xi Jinping. Trump is expected to press Xi over Beijing’s approach to Iran, senior US officials said. While China has been adding its voice to global pressure to stop the conflict, revenue that it provides to Iran as well as potential weapons exports would be among the topics discussed, one of the officials said. The US leader and Iran have rejected each other’s latest peace proposals to end the war as the two sides struggle to maintain a fragile ceasefire. In final preparations for the first US presidential trip to China in nearly a decade, Treasury Secretary Scott Bessent will meet his Chinese counterpart He Lifeng for last-minute talks in Seoul on Wednesday, both sides confirmed. Keep track of the twists and turns — and the global fallout — of the war with Iran here . What You Need to Know Today Rising prices in China are fueling gains in the yuan, which has climbed to a more than three-year high. That strength is likely to continue over the coming year, according to Goldman Sachs, which said the currency is more than 20% undervalued against the US dollar. Goldman...
一直以来，中美 AI 的发展差异，始终是行业讨论的核心话题。 美国负责创新，中国负责产业化，几乎已经成了很多人的共识。 但过去，大多数讨论都停留在外部视角：谁的模型更强、融资更多、产品更领先。真正从AI实验室内部出发，去观察中美研究文化、组织方式和技术生态差异的人，其实并不多。 不久前，美国 AI 研究员 Nathan Lambert 访问了多家中国 AI 实验室，包括月之暗面、智谱等等，并在之后...
一直以来，中美 AI 的发展差异，始终是行业讨论的核心话题。 美国负责创新，中国负责产业化，几乎已经成了很多人的共识。 但过去，大多数讨论都停留在外部视角：谁的模型更强、融资更多、产品更领先。真正从AI实验室内部出发，去观察中美研究文化、组织方式和技术生态差异的人，其实并不多。 不久前，美国 AI 研究员 Nathan Lambert 访问了多家中国 AI 实验室，包括月之暗面、智谱等等，并在之后撰写了一篇长文，系统分享了他对中国 AI 产业的观察。 Nathan Lambert 曾在 Meta AI、DeepMind、Hugging Face等机构工作，也是 AI 行业里颇具影响力的研究者和独立观察者之一。 在他的视角里，中国 AI 的竞争力，并不只是“追赶速度快”这么简单。 更深层的差异，来自工程文化、人才结构、组织方式，以及中国科技公司对于技术栈控制权的重视 。 某种程度上，这篇文章也提供了一个难得的“内部视角”——帮助我们理解，中国为什么能够在这一轮 AI 竞争中，始终紧紧咬住美国，甚至在部分方向上开始形成自己的优势。 以下是 Nathan Lambert 文章内容： 为什么中国实验室更适合“今天的大模型” 中国的大模型公司，某种程度上可以说是“快速追赶型技术体系”的最佳样本。 如果只看结果，中美之间其实已经非常接近。无论是支持 Agent 工作流的新一代模型，还是研究人员、数据规模、算力资源等基础条件，中国头部实验室与美国前沿实验室之间的差距，已经远没有外界想象得那么大。 真正的差异，不在资源本身，而在于这些资源是如何被组织、管理和激励的。 Nathan Lambert 在与大量中国 AI 实验室交流后，一个最强烈的感受是： 中国的组织文化，可能天然更适合当前这一阶段的大模型竞争 。 因为今天的大模型研发，本质上已经不再是单点突破，而是一场复杂的系统工程。 从数据清洗、模型架构、训练策略，到强化学习、后训练、Agent 能力，每一个环节都可能带来提升。但真正困难的地方在于，如何把这些细碎优化整合成一个整体。 很多时候，一个模型是否足够领先，并不取决于某一个“天才想法”，而取决于整个团队能否持续围绕最终结果做协同优化。 而这恰恰是中国实验室相对擅长的部分。 在 Lambert 看来，美国 AI 研究文化更强调个人表达。研究人员需要不断强调自己的贡献、技术路线...
Benjamas Deekam/iStock via Getty Images Since my last article , Cass Information Systems ( CASS ) does not look weak anymore to me. At the time my rating was hold, I was more cautious because Q2 2025 profit jump was strongly dependent on one-offs, especially TEM sale profit and Rubicon recovery. Now the situation is quite better. Q3 , Q4 , and Q1 2026 showed that EPS momentum did not disappear whe...
Benjamas Deekam/iStock via Getty Images Since my last article , Cass Information Systems ( CASS ) does not look weak anymore to me. At the time my rating was hold, I was more cautious because Q2 2025 profit jump was strongly dependent on one-offs, especially TEM sale profit and Rubicon recovery. Now the situation is quite better. Q3 , Q4 , and Q1 2026 showed that EPS momentum did not disappear when one-offs ended. I have to admit that the results are better. Still, I like to see clear growth stories. Transaction volumes remain weak, processing fees are pressured, and profit growth is mostly pulled by net interest income, expense control, and buybacks. So, the stock is not that expensive anymore, but I still do not see a strong argument to re-rate it to buy. What Changed? Since my last article, the main change is that CASS profitability remained. In Q3 2025 the company made $9.1 million in net income, diluted EPS was $0.68, NIM increased to 3.87%, and net interest income increased by 19.3%; overall that was an improvement since my Q2 coverage. Later, full-year 2025 results were even better because net income reached $35.1 million, diluted EPS was $2.61, net interest income grew 19.8%, and the company bought back 617 thousand shares. Q1 2026 was not that bad either, because adjusted EPS from continuing operations was $0.66 and grew 26.9%, while NIM increased to 3.95%. My first thesis and old argument were that Q2 2025 was inflated by one-offs, and I stand by it. It is just not in play anymore. CASS really maintained earnings momentum, and after Q2 we saw earnings get better and better. Processing fees went down, invoice volumes remained weak, and NIM is still pushing the company. Due to this reason, my current thesis is now less bearish, but I would not call it bullish either. Momentum As I said, I have to admit that currently CASS looks better than it did in my previous article. Q4 adjusted EPS from continuing operations increased 38.8% per year, and Q1 2026 adjusted...