Russia’s military satellite Kosmos-2612 has maneuvered to within 26 meters of ICEYE-X36, a Finnish-American radar satellite whose imagery is supplied to Ukraine. Space analyst Georgy Trishkin said the near-zero relative speed could indicate an attempt to force ICEYE-X36 to make orbital corrections, potentially disrupting operations and using up its fuel. RAND researcher Michael Bohnert called the approach a show of force.

very helpful

On the great @I_Am_NickBloom AI survey... and on the question of concentration... sort of mirrors the economy and ordinary software. cc @Noahpinion @ramez @hsu_steve @PeterBerezinBCA

Incredible scoop by my @WSJ colleagues: Elon Musk ordered the slowdown of a military-grade version of Starlink relied on by Ukraine, a decision that sparked panic in Kyiv and confusion inside the Trump administration. wsj.com/world/ukraine-starli…

yeah but i tend to write stuff down.

exhaustion

I would add this isn't particularly concentrated by standards of the US economy.

We covered off concentration. It's reasonable common level of concentration if you look at the facts. Unpicking this for AI, it's about 110k US firms or about 5000 firms in the UK. Don't forget the chart looks more when you look at employment (see below). Datadog, for public filings, is more concentrated than the above.

who are they?

Great new paper from @i_am_nickbloom A few thousand firms in the US and UK represent half of all AI spending-- roughly top 2%. Trouble is n=845 & 1273; so one firm here or there could make a huge difference. Bottom 50% of firms are <1% of AI spendin by eyeball. Average US firm spends $2000 per employee (in 2025) -- obviously skewed by high spenders. Marginal productivity benefits. Worth comparing to the BCG survey (n=1273), which has more recent fieldwork and may be more self-selected. nitter.siddz.com/azeem/status/211258097…

It isn't quite as easy as restocking a supermarket shelf. I'm not sure how big the unplugged GPUs are (i.e. is it a localised problem) vs just the build-up of a construction backlog investmentbrief.ai/p/ai-inve…

A BCG survey showed how companies worldwide are seeing the impact of AI. There are confounders, of course. Maybe advanced AI users just have better management teams, and therefore better commercial performance. It is a self-reported survey, etc., etc.

That's not really what's happening, as the Amazon, Google and Microsoft results will show.

bankers in notting hill

What this one from August 2026 ?

The easy thing to do to sense-check is to exclude Chinese clouds, and then use whatever proportion you think is chips and servers. Then you have a number for what NVDA, AVGO and AMD have to ship; if they wildly disagree, I would trust the earnings announcments over a chart.

what % of OAI and ANT revenue do you think you cover within Ramp?

This chart shows the key risk to the AI trade, and it's missing from the latest revenue announcements. The AI boom is fundamentally driven by a top 1% of companies that drive 80% of spend. Within that top 1%, 80% are venture-backed. That's all fine if the AI trade works out, and if, as some prominent investors say, everyone wins. But if it doesn't work out, then we have a cascading effect of large customers simultaneously pulling back their AI spend.

we do calculate meta’s ad uplift as well.

Both sets of analysts get it wrong. But solar has different dynamics: a technology on a steep learning curve and highly modularised, very fragmented market.