👀 Who's right?

SCOREBOARD

Tuesday close. S&P down 0.4%, down for a second day running, with Amazon and Oracle among the heavier drags. On the week: Nvidia −6.01%, S&P −1.14%, Nasdaq −1.66%.

ON WATCH

The two men with the most to say about AI safety said it on the same stage yesterday, and they flatly disagreed. At Salesforce's Dreamforce conference, Nvidia's Jensen Huang and Anthropic's Dario Amodei took opposite views of how worried anyone should be.

Huang, whose chips power almost the whole boom, brushed the worry aside. He called the choice between moving fast and staying safe a "false choice," and said there's no need for new AI laws because the market already forces firms to get it right.

That's a straight shot at Amodei, who spent last week arguing the industry must deliberately slow down. On stage he held the line, calling a slower pace the way to lead instead of racing everyone off a cliff.

It matters because Huang is the field's most powerful supplier, and "no new laws needed" is what Washington wants to hear. When the man selling the shovels says the gold rush is safe, the diggers relax.

To be fair, Huang isn't waving off safety, he's saying firms can police it themselves and pause when they must. The catch is that two DeepMind safety researchers quit last week saying the labs can't be trusted to do that, so it's the sellers versus the leavers.

SPONSORED

Every customer signal is captured in one shared context layer, always current and compounding. Agents and workflows build pipeline, chase every buying signal, and move deals forward, an always-on revenue engine running alongside your team.

With Attio, you’ll get:

Teams like Parallel, Turbopuffer, and Wordsmith build on Attio. Are you one of them?

Get started now

THE MACHINE

While the bosses argue about safety, the money isn't slowing at all. The Financial Times reported yesterday that OpenAI is in early talks to raise fresh cash at a valuation above $1.2 trillion.

That's up from $852 billion in March, so investors have pencilled in roughly another $350 billion of worth in under six months. It would make a company that isn't yet nine years old worth more than any bank on the planet.

The business underneath is growing fast and burning cash just as fast. Revenue is running at more than $40 billion a year, but OpenAI is on track to lose around $27 billion in 2026, and Sam Altman has called going public this year "ill-advised."

The bull case is real, because that run-rate has roughly doubled in a year and demand for its models keeps outrunning supply. The counter is just as plain, because $1.2 trillion is around thirty times sales for a firm still bleeding money and leaning on a handful of giant backers.

SOMEONE SMART SAID

Gomez runs Cohere, one of the smaller frontier labs, so he clearly has a horse in this race. His point still lands: when the biggest companies get to write the safety rules, the rules tend to fit the biggest companies.

He isn't asking for less oversight, he wants more, but shaped by a wide group instead of by Anthropic and OpenAI alone. He also isn't playing the danger down, calling today's models the most potent cyber weapon ever built.

The gap is the story. One camp wants to slow the whole field down, and a rival building the same tech says that's a tidy way to freeze the pecking order in place.

THE CHART

This is the Nasdaq Composite, the index the AI boom has carried to record after record. It's slipped about 1.7% this week as the chip stocks wobbled and long-term interest rates pushed toward levels last seen in 2007.

Higher rates hit the AI trade twice: far-off future profits are worth less today, and the debt-funded data-centre build gets dearer to finance. The case against the gloom is that the index still sits near an all-time high and the economy has kept growing through rising rates, so this reads as a wobble, not a rout.

HEADS UP

ON THE HORIZON

MEERKAT ON YOUTUBE

The Unthinkable Is About To Happen To The Fed. On the morning the Fed decides, the Finance channel digs into why the central bank may be losing the wheel, and what a rate move now could do to mortgage costs. Watch on Meerkat Explains Finance.

THE BURROW

Quick vote. Nvidia's Jensen Huang says AI needs no new laws and the market will keep it honest. Do you buy it?

Trivia. The U.S. Federal Reserve was created in 1913, largely in response to one earlier banking panic. Which year was that panic? Answer's at the bottom of the email. No peeking.

ALSO WORTH YOUR INBOX

Meerkat's other daily read: Crypto Nutshell, the 5-minute crypto brief for 95,000+ readers.

HOW DID WE DO?

One tap, and Meerkat listens. How did today's issue land? Hit a button, and add a line if you want to tell Meerkat why.

Rates land at 2pm. Meerkat's reading the dot plot so you don't have to.

Know someone whose savings are quietly riding on a $1.2 trillion bet? Forward them this.

Meerkat is not telling you to buy anything, sell anything, or short anything. Meerkat is not a financial adviser. Meerkat is a small animal with a whiteboard and a Fed dot plot it can't read. Please do not sue Meerkat.

TRIVIA ANSWER

1907. The Panic of 1907 got so bad that the financier J.P. Morgan personally corralled a rescue, and the fright pushed Congress to create the Federal Reserve six years later.