Tim Cook will move to executive chairman on September 1, with John Ternus — Apple's SVP of hardware engineering — becoming only the third Apple CEO this millennium. The board approved it unanimously. Cook stays on to run policy and government relations. Arthur Levinson steps up to lead independent director.

The numbers under Cook are staggering. Apple went from roughly $350 billion in market cap to $4 trillion. Every iPhone since the 4S. Apple Watch, AirPods, Vision Pro, and the full Intel-to-Apple-silicon transition. A services line that grew into a $100 billion-a-year business. Sam Altman's reaction on X landed in one sentence: "Tim Cook is a legend."

What makes this interesting for private markets is the timing. Most analysts had 2027 penciled in for the handoff. Pulling it a full year forward — and installing a 51-year-old hardware engineer who's 15 years younger than Cook — says something about what the board thinks the next era needs. Creative Strategies' Carolina Milanesi caught the sharpest implication: "Now Ternus will be the one to introduce the first foldable iPhone. The most consequential hardware moment in years belongs to a hardware engineer from day one."

Ternus spent 25 years at Apple. His second job out of college. His public philosophy is telling: "We never think about shipping a technology. We always think about how can we leverage technology to ship amazing products." Read that as a signal for where AI lives on Apple's 2 billion active devices: on-chip, not in the cloud. Apple also expanded Johny Srouji's role to lead Hardware Engineering, keeping the Apple Silicon architect in the seat that drives the chip roadmap.

The private-markets read: every AI company with a dependency on Apple's ecosystem — Siri partnerships, App Store distribution, developer tools — is now negotiating with a hardware-first CEO whose instincts favor silicon over APIs. That shifts the calculus for startups that assumed cloud-model access to Apple's users was the long game. Shares were down 2.5% today. TechCrunch has more.

Amazon said Monday it will add up to $25 billion to its Anthropic investment, on top of the $8 billion already committed. Five billion funds now. The other $20 billion is gated on "commercial milestones." As part of the same deal, Anthropic has committed to spend at least $100 billion on AWS over ten years.

On paper: hyperscaler funds frontier lab, lab buys compute from hyperscaler. In practice: the dollars move in a tight loop. Mukund on X summed up the reaction: "This is insanely circular. Why are the SEC and FTC not all over this?" Corey Quinn asked: "How much of Amazon's investments in these AI labs is in the form of AWS credits?"

Set aside the accounting debate. The capacity math is what matters for the pre-IPO names. Anthropic has now locked up roughly 5 gigawatts from Amazon plus another 3.5 gigawatts from Google and Broadcom. Four months ago OpenAI executives were reportedly calling Anthropic's compute stance a "strategic misstep." That critique no longer fits. And as we wrote last week, Bloomberg reported that Anthropic is fielding investor offers at roughly $800 billion — more than 2x the $350 billion February round — and is resisting. If any round eventually prints near that level, AI infrastructure comps on the secondary market get repriced overnight. TechCrunch on the Amazon investment.TechCrunch on the $800B offers.

A tranched investment is capital committed upfront but released in stages, with each stage gated on a milestone — revenue, product, regulatory, or otherwise. Amazon's new $25 billion Anthropic commitment is structured exactly this way: $5 billion funded now, $20 billion contingent on "commercial milestones." The structure lets the lead manage risk while giving the company a larger headline number to cite. It also means the announced round size and the actually-deployed capital are two different numbers. On every large AI round going forward, it's worth reading the tranche language before taking the valuation at face value.