👀 1929 Is Trending Again

Stax Payments, a Hubspot customer, now books 80% more meetings with the exact same team they had before.

No new hires. No new territories. Just HubSpot's Prospecting Agent finding the right accounts, surfacing the right contacts, and drafting outreach that actually gets responses.

Your best rep is great because they know who to go after and when. Prospecting Agent gives every rep that same starting point.

Get a demo

Uncle Sam is buying his own debt back. At double the size.

The Treasury just announced it will at least double the maximum size of its liquidity-support buyback operations, aimed at longer-dated nominal coupon securities.

Plain English: the government is going into the open market and hoovering up its own long-term IOUs.

Official reason? Better market functioning and more flexibility in managing the supply of long-term debt. Call it plumbing maintenance.

Look at what it costs the US to borrow right now.

Yields rise when buyers get fussy. That whole curve is investors saying they want paying properly to lend to the world's largest economy.

The bond market is the landlord who never misses a rent review. It just put the rate up across every unit in the building.

Here's the part doing the rounds online.

In the months before the October 1929 crash, the Treasury was also buying and retiring debt. One operation on September 16, 1929 involved $100 million in securities.

Roughly six weeks later: Black Monday and Black Tuesday, October 28 and 29.

Spooky. Also basically meaningless.

Two dots do not make a trend line. Treasury buybacks are a routine liquidity tool that have run in plenty of years which ended with absolutely nothing dramatic. 1929 had a few other things going on besides a bond desk being busy.

If you want a reason to be cautious, the yields are right there. You don't need a 97-year-old coincidence.

Buybacks help the long end trade more smoothly. They don't change who ultimately has to absorb all this debt.

And that matters to you for one simple reason: a 5.26% long bond is competing directly with your stock portfolio.

When the boring option pays over 5%, equities have to work harder to justify the risk. Growth names priced for a decade of perfection feel that pressure first.

Elevated long yields also feed into mortgage rates and corporate borrowing costs. That bleeds into earnings eventually.

So watch the 30-year. It's the tell.

Opening entities in every country can be slow, expensive, and hard to scale.

That's why more companies are using EOR to hire globally faster.

See how Oyster helps teams hire, pay, and support talent in 180+ countries while staying compliant along the way.

Get all the answers

Trump wants America launching rockets 1,000 times a year.

That's roughly three a day. Every day. Including Christmas.

He signed a memorandum Thursday titled around a "Golden Age of Space Transportation," with more than 1,000 US launches and reentries annually by 2030 as the target.

Most of the memo is aimed at paperwork, which tells you where the traffic jam actually is.

It directs agencies to expedite permitting and environmental reviews, and to develop range-scheduling criteria and publish the schedules so launch resources get allocated properly.

Right now a launch range is an airport with one runway and no departures board. Everyone turns up and argues.

There's also an order to set fair and transparent cost-recovery policies for shared space services and infrastructure, so companies know what using the government's stuff will cost them before they commit.

Boring? Yes. Also the difference between launching monthly and launching daily.

The same memo tells agencies to facilitate commercial transportation to and from the Moon, support commercial robotic access to Mars, and explore commercial opportunities for human missions to and from Mars.

Casually slipped in there. Like adding "buy milk" to a list that already says "colonise a planet."

Backing it up: a space transportation industrial-base strategy, streamlined procurement, and incentives for private companies to co-develop launch infrastructure with the government instead of waiting for it.

A memo doesn't build anything. What it hands the industry is predictability, and predictability is what capital needs before it shows up.

If a company knows its permit timeline, its range slot and its cost of using shared infrastructure, it can plan. Planning is how you get from a handful of launches to three a day.

The knock-on list is long: launch providers, satellite operators, component suppliers, insurers, anyone charging rent on a piece of the pipeline.

Two things worth keeping in mind. 2030 is four years away, which is one election and several budget fights from here. And a memorandum can be undone by the same instrument that created it.

Direction of travel is real. The timeline is a wish until someone funds it.

AI has to earn $10 trillion a year for any of this to make sense.

Not cumulatively. Not by 2040. Every single year.

That's the number BCA Research chief economist Peter Berezin landed on when he ran the maths on what today's data centre spending actually has to pay back.

Here's how you get there 👇

Amazon, Microsoft, Google, Meta and Oracle are on track for $1 trillion of annual capex by 2027, driven almost entirely by AI infrastructure.

A trillion a year. On warehouses of chips that age like milk.

Which raises the question nobody wants asked on the earnings call: what exactly do you have to sell to make that back?

Start generous. Assume a conservative 30% pre-tax return on invested capital, and a 50% EBITDA margin, which is the forecast, not the history.

Run those numbers and the hyperscalers on their own need more than $4.3 trillion in annual revenue.

That works out at $526 from every person alive. Every year. Including the several billion who have never touched a GPU.

And that was the friendly version.

Let margins drift back to their historical norm of 30% and the bill climbs to nearly $7.2 trillion.

Now invite everyone else to the party. Alibaba, Baidu, the "neoclouds," plus adjacent moonshots like SpaceX.

Total: $10 trillion.

Berezin's own yardsticks make it uncomfortably clear:

Put another way, AI has to grow into a line item the size of feeding humanity. That's the hurdle rate.

Right now the spending is funded by some of the most profitable companies in history, so it looks painless.

The catch is timing. Capex hits the balance sheet today and the depreciation hits the income statement for years afterwards.

If the revenue arrives on schedule, nobody ever mentions this maths again.

If it arrives late, the write-downs show up first, and they land on the same five companies that carry an enormous share of the index most people passively own.

Worth knowing the size of the bet you're already in.