Pivot News AI Briefing — Jun 25, 2026

Show Notes

Hosts: James & Maya In this episode: • Hi, I'm Maya, and today we're looking at OpenAI building its own chip, a memory-chip boom, AI flooding kids' feeds, a women's health startup, and a po... • Let's start with the chip. OpenAI just announced Jalapeño, its first custom inference accelerator, built with Broadcom and supported by Celestica. • Inference is the part that runs every time you ask ChatGPT a question. Doing that on your own silicon means you control the cost instead of paying som... • And the timing matters. This comes nine months after they revealed the Broadcom partnership. The real story isn't the headline chip — it's OpenAI want... • Worth noting the caveats here. A custom chip is a long bet. Designing, fabbing, and deploying takes years, and Nvidia isn't standing still. But for a ... Subscribe to the newsletter at pivotnews.ai for the full written briefing.

What is Pivot 5: Today's Top AI Headlines?

Pivot5 | 5 Headlines & Unprompted

James: Hi, I'm James.

Maya: Hi, I'm Maya, and today we're looking at OpenAI building its own chip, a memory-chip boom, AI flooding kids' feeds, a women's health startup, and a political race tech money couldn't buy.

James: Let's start with the chip. OpenAI just announced Jalapeño, its first custom inference accelerator, built with Broadcom and supported by Celestica.

Maya: Inference is the part that runs every time you ask ChatGPT a question. Doing that on your own silicon means you control the cost instead of paying someone else's margin.

James: And the timing matters. This comes nine months after they revealed the Broadcom partnership. The real story isn't the headline chip — it's OpenAI wanting to own the whole stack, from the model down to the metal.

Maya: Worth noting the caveats here. A custom chip is a long bet. Designing, fabbing, and deploying takes years, and Nvidia isn't standing still. But for a company spending heavily on compute, even a modest efficiency gain pays for itself.

James: Which brings us to who's selling the hardware. Micron just posted a quarter that says the spending isn't slowing.

Maya: The numbers are striking. Fiscal Q3 revenue of $41.46 billion, adjusted EPS of $25.11, both above consensus. And they guided Q4 to $50 billion — more than 40% above Wall Street estimates.

James: That's not a small beat. The stock jumped 15%.

Maya: And there's a $22 billion backlog behind it. The data tells a clear story: hyperscalers are still ordering memory chips hard, despite recent market swings. The AI buildout looks healthy through at least August.

James: So if you're a business leader wondering whether the capex wave is cresting, Micron is your read. Demand is real and it's booked.

Maya: For now. One quarter doesn't make a cycle, but it does push back on the slowdown talk.

James: Now to a story closer to home. A new analysis found 59% of videos in a new TikTok user's For You feed are AI-generated, and that content dominates kid hashtags like nursery rhymes and cartoons.

Maya: Let's look at what actually happened. These are toddlers getting served machine-made cartoons. The animations are off — balloons with double knots, trucks that almost look like familiar characters. Close enough to pass, wrong enough to notice.

James: What nobody's talking about yet is the economics. Human creators making kids' content now compete with free, high-volume AI output. That's hard to win on price.

Maya: And it puts real pressure on TikTok's moderation and on parent trust. The platform has to decide how much of this it tolerates before families walk away.

James: From slop to something more hopeful. Xella Health came out of stealth with $4.7 million in seed funding to build an AI precision health platform for women's diagnostics.

Maya: The number that frames this: women's health gets only 6% of private healthcare investment. That gap slows research and leaves a lot of women without clear answers.

James: The founder, Kelly Lacob, watched her mother's cancer return during her own MBA. She saw the gaps even with good access. Backers include Precursor, Ulu, and Swizzle Ventures.

Maya: It's early, and a $4.7 million seed is small. But investor interest here suggests women's health diagnostics is starting to look like a real commercial opening, not just an underserved cause.

James: Our last main story is political. Micah Lasher won New York's 12th district primary with 39.1% against a field of eight, and he openly called out the AI companies that spent millions on the race.

Maya: Tech giants treated this Manhattan seat as an AI-regulation test. A candidate who rejects that money winning anyway is a result worth watching.

James: The general election will tell us whether voters reward that stance. If direct tech spending faces real resistance, that changes how these local races get fought.

Maya: Though it's a deep-blue district, so the primary was the real contest. We'll see if it holds as a pattern.

James: And that's our briefing. Five stories, one thread — the money and the politics of AI are both getting more concrete.

Maya: Thanks for listening. We'll be back tomorrow with another Pivot briefing.