For most of the week, the market treated Nvidia's buyback announcement like the whole story.

$235 billion authorized. The largest in company history.

The read everywhere was simple: a company this confident in its own stock must know something the market doesn't.

Then I pulled the insider filings.

Nvidia executives have sold just under a billion dollars worth of stock in the past three months alone.

That's not a rounding error, and it's not index-fund turnover. It's the people closest to the business cashing out while the company tells shareholders the stock is cheap.

Today I want to walk through what the buyback is actually signaling, what the selling is actually signaling, and why both can be true at once without canceling out.

Missed Friday's Playbook on Broadcom?

The bigger story is that AI infrastructure is running into political and social limits, not just technical ones. Public opposition has already affected roughly $42 billion of European data-center investment through delays and cancellations, with similar resistance emerging in South Korea over power use, water, land, and proximity to homes. In plain English, the AI buildout may still be enormous, but getting projects approved is becoming a real investment risk.

Sandisk the slight edge because its NAND business offers more upside if AI storage demand stays strong, with quarterly revenue up 372% and gross margins near 85%. Western Digital is the steadier choice, with revenue up 44% and less exposure to the boom-and-bust swings of flash memory. The simple takeaway is that Sandisk offers more growth, while Western Digital looks like the more conservative AI storage bet.

McDonald’s has fallen to its lowest levels since 2022, pushing its RSI to around 25 and its forward P/E to roughly 19, well below recent norms. Analysts still see nearly 30% upside, but investors are worried about weak U.S. traffic and backlash over AI-powered pricing plans. In plain English, this is starting to look like a contrarian setup, but the stock still needs to prove the selloff has actually bottomed.

What do Molly Ringwald, Dan Levy, Amy Porterfield & a world record have in common? They'll all be at GURU Conference 2026. 100% Free. 100% Virtual.

If you're obsessed with marketing like we're obsessed with marketing, GURU is the must-attend conference of the year. We'll be covering all things email marketing: B2B, B2C, newsletters, deliverability, email design, AI & more.

You can expect to walk away with new email strategies, the very latest digital trends, and how to step up your email performance. But don't worry, we also like to have fun. This year's theme is rom-com, so there will be DJs, meet-cutes, and a cutest pet contest. (Start prepping your dog's headshot now.)

Don't miss out. Join us Nov 12th & 13th for the largest virtual & free email marketing conference, powered by Constant Contact.

Yes! I Want to Register!

Nvidia just hit a fresh all-time high of $237.27, pushing its market cap past $5.7 trillion.

Three days earlier, it added $150 billion to its buyback authorization, bringing the total to $235 billion.

A company announcing the largest buyback in its history, trading at 29.57x forward earnings against a five-year median near 56x, is telling you it thinks its own stock is cheap.

That's the story most of Wall Street is running with this week.

Here's what complicates it. Nvidia insiders have sold $967.6 million worth of stock in the past three months.

Institutional positioning is split almost down the middle, 20 funds trimming against only 12 adding.

If the people running this company, and the money that tracks it closest, genuinely believed that $400 number, you'd expect more buying, not more selling.

So which signal is lying to you? The buyback, or the people walking away from their own stock?

Nvidia is about to launch the DGX Spark, a $4,999 desktop system that can run AI models with up to 100 billion parameters, shipping October 23.

That's Nvidia pushing its chips further down the food chain, from hyperscaler data centers into the hands of individual developers and smaller labs.

Analysts expect Nvidia's upcoming quarter to show revenue growth near 91% year on year, past $109 billion.

The GF Score of 95 out of 100 backs that up, perfect marks on both profitability and growth, rare for a company this size.

Jensen Huang has used GTC to unveil Nvidia's biggest moves before, and GTC lands again this month.

Here's the complication the buyback headline buries.

Nvidia insiders sold $967.6 million in stock over the last three months.

Jensen Huang's selling isn't new, he trades under a pre-set plan and has been steadily reducing his stake for over a year.

But this scale, paired with 20 institutional holders trimming against just 12 adding, is a different signal than one executive diversifying.

Not a single rogue seller. A pattern across the people and funds closest to the company.

Insider selling doesn't automatically mean a stock is about to fall, people sell for personal reasons unrelated to the business.

Because if insiders genuinely believed the $400 fair value number, the easiest, lowest-risk way to express that would be to simply hold what they already own.

They are not doing that. The gap between what the buyback says and what the selling says is the real story here.

Nvidia's chart tells its own version of this story, and it's been a wild one.

The stock bottomed around $189.60 back in July, after a selloff that had bulls questioning whether the AI trade had run out of road.

From there it clawed back in a textbook double bottom, neckline near $213, and once it broke above that level the buying accelerated fast.

This week it hit a fresh all-time high of $237.27, finally clearing the old 2026 high near $236 that had capped it twice before.

The momentum reading backs it up, the Average Directional Index climbing toward 16, early innings for a trend, not a blow-off top.

Bulls are already talking $250 next, and $300 if that breaks.

The risk sits below, not above. A stock that just broke a multi-month base on conviction needs that $213 neckline to hold if this run stalls.

Lose that level and the double bottom stops meaning anything.

Jensen Huang has turned GTC into the venue for Nvidia's biggest announcements, and it lands again this month.

Bulls want another roadmap reveal or a surprise customer deal that extends the story past DGX Spark.

Bears will be watching for demand commentary that sounds more cautious than recent quarters.

A quiet GTC, with nothing new to point to, would be its own kind of signal.

TSMC and ASML report October 14 and 15, effectively doing Nvidia's homework before Nvidia reports its own.

Strong numbers there support the idea that chip demand is still outrunning supply.

Weak guidance from either would be the first real crack in the AI infrastructure story this quarter.

Watch the commentary on advanced packaging and memory capacity specifically, that's where Nvidia's own bottleneck has lived all year.

A $235 billion authorization is a ceiling, not a promise, companies routinely announce numbers they never fully use.

Meaningful repurchases against this authorization over the next couple of quarters would be a real vote of confidence with cash behind it.

A number that sits mostly unused while insiders keep selling starts looking more like a headline than a strategy.

I covered this exact pattern back in March, when Nvidia beat estimates and the market sold it anyway. The setup rhymes: strong fundamentals, a market not fully buying the obvious bullish read.

Nvidia's own earnings land in November, estimates calling for revenue growth near 91% to north of $109 billion.

The number that ultimately settles this argument isn't the buyback or the insider filings, it's what shows up on that report.

A beat with accelerating guidance makes the selling look like noise.

A beat that gets sold anyway, the way it did back in March, makes the insider activity look a lot more prescient.

I keep coming back to one simple test with insider selling: would I rather own what the executives are buying, or avoid what they're selling?

Right now Nvidia's own people are choosing the second option, at a scale that's hard to wave away as routine diversification.

And honestly? A $235 billion buyback sounds impressive until you remember it's authorized, not spent, and Nvidia has a long history of letting these numbers sit mostly on paper.

I'm not calling Nvidia a short here. The growth is real, the GTC pipeline is real, DGX Spark pushes the company into markets it's barely touched.

But I'm also not going to pretend nearly a billion dollars in insider selling is just noise. The people with the best information about this business just told you, with their own money, what they think $237 is worth relative to where it's headed next.

That's the trade worth watching, not the headline number on the buyback press release.

Wispr Flow turns your voice into clean, professional text inside any app. Emails, Slack, client updates: speak once, send without editing. 4x faster than typing. Works on Mac, Windows, iPhone, and Android.

Try Wispr Flow free

Buybacks and insider selling aren't opposites, they're just two different people making two different bets on two different time horizons.

A company can believe its stock is cheap on a five-year view while the person who built it decides thirty years of concentrated risk is enough for one lifetime.

The mistake is treating either signal as the whole truth.

The honest move is to weigh both, notice when they disagree this sharply, and size your own position accordingly.

Nvidia is still one of the best businesses on the planet. That was true before this buyback and it'll be true after it.

Being a great business and being fully priced aren't the same question, and this week Nvidia is forcing investors to answer both at once.

Know someone who'd rather think than chase? Forward them this edition.