Sgòr Gaoith, Cairngorms: Our rucksacks are heavy with crampons and ice axe, and the wind is a menace, but it’s all worth it for the views The weather app shows wall-to-wall spiky yellow suns. We haven’t had a day like this in over a month, nor is one foretold in the weeks to come, so I abandon the desk and, with a friend and a dog, head for the hills. We choose Sgòr Gaoith , the high point of the ...
Sgòr Gaoith, Cairngorms: Our rucksacks are heavy with crampons and ice axe, and the wind is a menace, but it’s all worth it for the views The weather app shows wall-to-wall spiky yellow suns. We haven’t had a day like this in over a month, nor is one foretold in the weeks to come, so I abandon the desk and, with a friend and a dog, head for the hills. We choose Sgòr Gaoith , the high point of the ridge above the River Feshie which forms the western edge of the Cairngorms. Our path starts in Scots Pine forest, where we walk in dappled light and can hear the soft chuckle of Allt Ruadh, the Red Burn. The trees thin as we gain height, but everywhere, new saplings are appearing across this once bare landscape and you can almost feel the rising of life. Our rucksacks are heavy with crampons and ice axe – just in case – but the air is fresh and spangling as the sun pours down the brown slopes and into our winter-dull eyes. Continue reading...
Worawith Ounpeng/iStock via Getty Images One IPO priced this past week, joined by one SPAC, and one major deal joined the pipeline. The biggest news of the week came from the backlog, though: Elon Musk’s SpaceX ( SPACE ) has reportedly submitted a confidential filing ahead of its mammoth IPO. Reports from the prior week indicated that the offering could raise $75 billion. HMH Holding ( HMH ) price...
Worawith Ounpeng/iStock via Getty Images One IPO priced this past week, joined by one SPAC, and one major deal joined the pipeline. The biggest news of the week came from the backlog, though: Elon Musk’s SpaceX ( SPACE ) has reportedly submitted a confidential filing ahead of its mammoth IPO. Reports from the prior week indicated that the offering could raise $75 billion. HMH Holding ( HMH ) priced slightly below the midpoint to raise $210 million at an $880 million market cap. HMH provides equipment and services for the oil and gas industry, with a focus on offshore operations. The company has a presence in 16 countries and sales in over 60 countries and emphasizes a more asset-light business model compared to its peers. The energy industry is cyclical, though, and HMH partly depends on new builds and equipment wear and tear. Despite launching alongside surging energy prices, HMH broke issue on its first day; it finished the week down 6%. 2 IPOs During the Week of March 30th, 2026 Issuer Business Deal Size Market Cap at IPO Price vs. Midpoint First Day Return Return at 04/03 KPET Ultra Paceline ( KPET.U ) $200M $252M 0% +0% +0% Blank check company led by former TPG Pace Group executives. HMH Holding ( HMH ) $210M $880M -2% -4% -6% Provider of drilling equipment and services. Click to enlarge One major issuer joined the pipeline this past week. The Elmet Group ( ELMT ), which makes specialized products for the aerospace, defense, and other industries, filed to raise $115 million. 2 Filings During the Week of March 30th, 2026 Issuer Business Deal Size Sector Lead Underwriter Churchill Capital XII ( CXIIU ) $300M SPAC Citi Blank check company founded by dealmaker and former Citi executive Michael Klein. The Elmet Group ( ELMT ) $115M Industrials Cantor Fitz. Makes critical materials and high-power microwave products for aerospace and defense. Click to enlarge Week Ahead One direct listing is currently scheduled in the week ahead, although some smaller issuers may join...
AlexSecret/iStock via Getty Images By Jennifer Nash The S&P 500 had its best day since May on Tuesday, which led to the index's largest weekly gain in four months and its first in six weeks. The index rose 3.4% from the previous Friday and is now 5.67% off its all-time high from January 27, 2026. Here is a snapshot of the index from the past week: The table below summarizes the number of record hi...
AlexSecret/iStock via Getty Images By Jennifer Nash The S&P 500 had its best day since May on Tuesday, which led to the index's largest weekly gain in four months and its first in six weeks. The index rose 3.4% from the previous Friday and is now 5.67% off its all-time high from January 27, 2026. Here is a snapshot of the index from the past week: The table below summarizes the number of record highs reached each year dating back to 2013. Here is a snapshot of the index from the past six months with a 50-day moving average: S&P 500: A Perspective on Drawdowns On October 9, 2007, the S&P 500 reached an all-time high, closing the day at 1565.15. Then on March 9, 2009, the index dropped ~57% off of its high from exactly 17 months before, closing the day at 676.53. This time period became known as the Global Financial Crisis. It took over 5 years before the index reached a new all-time high on March 28, 2013, where it closed out at 1569.19. The chart below is a snapshot of record highs and selloffs since the 2007 peak reached on October 9, 2007. What happens if we take out the Global Financial Crisis? Here's a snapshot of the same chart above where the start date has been changed to the trough reached on March 9, 2009. Note the recent selloffs in 2022. Here are a few tables with the number of days of a 1% or greater change in either direction and the number of days of corrections (down 10% or more from the record high). And here is a linear chart of the index since October 9, 2007: Here is a linearly scaled version of the same chart with the 50- and 200-day moving averages. The index has been below the 50-day moving average since February 27th, 2026, and below the 200-day moving average since March 19th, 2026. Additionally, the 50-day moving average has been above the 200-day moving average since July 1st, 2025. S&P 500: A Perspective on Volatility For a sense of the correlation between the closing price and intraday volatility, the chart below overlays the S&P 500 sinc...
Microsoft (NasdaqGS:MSFT) is committing a US$10b investment in AI and cloud infrastructure in Japan. The plan includes new AI data center capacity and partnerships with Japanese firms such as SoftBank and Sakura Internet. Microsoft aims to support training for 1 million Japan based engineers by 2029 and deepen cybersecurity cooperation with local authorities. For investors tracking Microsoft, this...
Microsoft (NasdaqGS:MSFT) is committing a US$10b investment in AI and cloud infrastructure in Japan. The plan includes new AI data center capacity and partnerships with Japanese firms such as SoftBank and Sakura Internet. Microsoft aims to support training for 1 million Japan based engineers by 2029 and deepen cybersecurity cooperation with local authorities. For investors tracking Microsoft, this move extends the company’s cloud and AI footprint in one of the world’s largest technology...
Many of Tokyo’s popular and iconic Somei Yoshino cherry blossom trees were planted during Japan’s post-war advancement in the 1960s, and are now getting old and frail. Some have fallen and many others require support, triggering safety concern as the Japanese celebrate the season of their favourite flower. Two cherry blossom trees collapsed on Thursday, one at Kinuta Park in downtown Tokyo and the...
Many of Tokyo’s popular and iconic Somei Yoshino cherry blossom trees were planted during Japan’s post-war advancement in the 1960s, and are now getting old and frail. Some have fallen and many others require support, triggering safety concern as the Japanese celebrate the season of their favourite flower. Two cherry blossom trees collapsed on Thursday, one at Kinuta Park in downtown Tokyo and the other at the Chidorigafuchi greenway. The one in Kinuta Park damaged a fence while the other tree...
PeopleImages/iStock via Getty Images Weak demand for labor and job destruction at federal & state governments should push up unemployment. But the supply of labor has plunged. The federal government shed another 18,000 jobs from its payrolls in March, including employees who’d departed earlier but whose severance packages ended in February. Since January 2025, the federal government has shed 352,0...
PeopleImages/iStock via Getty Images Weak demand for labor and job destruction at federal & state governments should push up unemployment. But the supply of labor has plunged. The federal government shed another 18,000 jobs from its payrolls in March, including employees who’d departed earlier but whose severance packages ended in February. Since January 2025, the federal government has shed 352,000 civilian employees, or nearly 12% of its staff. Federal government employment is now down to 2.66 million, the lowest since 1966, when LBJ was President. State governments shed another 4,000 jobs in March, and since January 2025 have shed 49,000 jobs or about 1% of their payrolls, according to data from the Bureau of Labor Statistics today. Local governments added 14,000 jobs in March. Local government employment is largely composed of educators, first responders, and healthcare workers. Since January 2025, local governments have added 161,000 workers to their payrolls. The private sector added 186,000 jobs in March. But the data has been yo-yoing, made worse by big revisions in both directions. Today, January was revised up further, February was revised down further. And earlier this year, the BLS adjusted everything to its annual massive benchmark revisions. So the six-month average, which irons out the silly month-to-month ups and downs – a labor market doesn’t turn on a dime every single month – shows the trend better: The private sector added 53,000 jobs on average per month over the past six months. Total payrolls, including government, rose by 178,000 in March. But the six-month average inched up by only 15,000 after having been in the negative in four of the past five months, dragged down by job cuts at federal and state governments. These kinds of dynamics – negative to low positive growth in nonfarm payrolls – would normally cause the unemployment rate to spike because normally, the US labor force would grow substantially through immigration, legal and illegal,...
This story has been made freely available as a public service to our readers. Please consider supporting SCMP’s journalism by subscribing. Hong Kong is bracing for a series of thunderstorms on the weekend, with the weather forecaster issuing an amber rainstorm warning on Saturday morning. The Hong Kong Observatory announced the amber warning, the lowest of the three-tiered rainstorm warning system...
This story has been made freely available as a public service to our readers. Please consider supporting SCMP’s journalism by subscribing. Hong Kong is bracing for a series of thunderstorms on the weekend, with the weather forecaster issuing an amber rainstorm warning on Saturday morning. The Hong Kong Observatory announced the amber warning, the lowest of the three-tiered rainstorm warning system, at 11am. It also said rain was particularly heavy in Tai Po district and might cause serious...
Would you pay for (or even bother to use) a calculator if you knew it was only going to be right 98% of the time? You might if you knew enough about what you were calculating to tell if an answer was wrong. But otherwise? Probably not. Keep that thought in mind—we’ll come back to it. This week’s newsletter is about artificial intelligence, or more particularly about the large language models (LLM)...
Would you pay for (or even bother to use) a calculator if you knew it was only going to be right 98% of the time? You might if you knew enough about what you were calculating to tell if an answer was wrong. But otherwise? Probably not. Keep that thought in mind—we’ll come back to it. This week’s newsletter is about artificial intelligence, or more particularly about the large language models (LLM) we currently refer to as AI. It’s about what they need from us and what we think we need from them. There are two parts to the first bit: computing power and data. Compute isn’t 100% straightforward: The electricity needed for giant data centers is in short supply everywhere (no one has invested enough in their grids for decades). There’s also a growing helium problem: a third of helium is produced in Qatar. Without it (as a coolant) chip production is impossible. No chips, no AI. Still, this is not an existential problem. With good policy choices and supply chains sorting themselves out, energy and material shortages can be (eventually) surmounted everywhere. The inputs may be increasingly expensive, but compute is always buyable. The second part, the data on which all LLMs are trained, is not. Its supply is limited. Up to ChatGPT4, the internet provided enough data for each new iteration to be better. But that version was completed a few years ago, trained on the lot. There is little more for new models to train on. The data on the internet might have expanded over the last few years, but not in a particularly helpful way. Much of it has been produced by other LLMs: train your new model on that and you might end up degrading it. Why? Because LLMs are horribly prone to errors (confabulations or hallucinations), which means they can’t give us what we most need from them: accuracy. An LLM is not a continuous learning machine. Its knowledge stops with its training. It also isn’t deterministic (like, say, a calculator), says AI expert Janusz Marecki (who I interviewed for a p...
Roommates overall are skewing older, as young people stay with their parents for longer. The share of older adults looking to rent with a roommate has tripled from a decade ago.
Roommates overall are skewing older, as young people stay with their parents for longer. The share of older adults looking to rent with a roommate has tripled from a decade ago.