AI's closest insiders sold into a turbulent quarter: what their timing suggests
In September 2024, while tech stocks were swinging on doubts about how soon artificial intelligence would turn a profit, Nvidia CEO Jensen Huang finished selling about 6 million of his company's shares, worth roughly $713 million. Across the three months ending that September, Nvidia's executives and directors sold around $1.8 billion in stock. That was one of the heaviest stretches of insider selling the company had seen in years. Over the same period, insiders bought exactly zero shares on the open market.
Headlines like that read as a warning: the people closest to the AI boom were cashing out during a shaky quarter, so surely they knew something. A month later, in October 2024, Nvidia hit new all-time highs on demand for its Blackwell chips. That episode is worth remembering as this year's third-quarter insider filings get tallied. The timing of a sale tells you when a decision was made, and that decision is usually months older than the market mood around it. If you hold US tech through a Canadian ETF or mutual fund, as plenty of Edmonton investors do, knowing this changes how you react.
Most insider sales were decided before the turbulence
Huang's sales ran under a Rule 10b5-1 plan. Executives set these up in advance, choosing how many shares to sell and on what schedule, then let the plan run on its own. The plan protects them from accusations of trading on information the public doesn't have. Every trade is still reported to the US Securities and Exchange Commission on a Form 4, which is why the totals become news.
So selling during a rough quarter mostly reflects a choice made in calmer weeks. Tech leaders are also paid mainly in stock. When shares vest, a tax bill arrives, and selling some of those shares is often the only practical way to pay it. Founders and executives at Amazon and Meta sell regularly for the same reasons: taxes, spreading their money around, buying property. Someone trimming a sliver of a fortune held almost entirely in one company is managing personal risk, and the trade says little about next quarter.
Before reading meaning into a sale, check whether it ran under a pre-set plan. If it did, the date the plan was adopted matters more than the date of the trade.
The quieter signal is the buying that didn't happen
Selling has a dozen dull explanations. Buying has one. An insider who spends their own cash on open-market shares thinks the price looks cheap. During that 2024 quarter, nobody at Nvidia did. That didn't predict a fall, and the stock went on to rise. It does tell you that the people with the best view didn't see a bargain at those prices.
Heavy, visible selling can also weigh on a stock for a while, because other investors get wary of buying what insiders are unloading. That pressure comes from investor nerves and tends to fade, but it can leave a share price feeling stuck for a stretch.
Watch for the buying that's missing at least as closely as the selling that shows up.
Selling on a schedule is the habit worth copying
The most useful part of the Nvidia story may be the behaviour itself. Executives with most of their wealth tied up in one asset decide in advance to trim it on a schedule. Once that's set, a scary week can't talk them out of it or push them into something rash.
Plenty of Alberta investors are more concentrated than they realise. It might be a single tech fund that has run up, a rental property that now holds most of their net worth, or a business that doubles as the whole retirement plan. Turning part of that into something else is the careful choice, whether it becomes a down payment, a smaller mortgage balance or a broader portfolio. It's far easier to do with rules you set while you were calm.
Write your rule down before the next turbulent quarter. Decide how large a share of your wealth any one asset is allowed to become, and what you'll sell, and when, if it crosses that line.
In September 2024, while tech stocks were swinging on doubts about how soon artificial intelligence would turn a profit, Nvidia CEO Jensen Huang finished selling about 6 million of his company's shares, worth roughly $713 million. Across the three months ending that September, Nvidia's executives and directors sold around $1.8 billion in stock. That was one of the heaviest stretches of insider selling the company had seen in years. Over the same period, insiders bought exactly zero shares on the open market.
Headlines like that read as a warning: the people closest to the AI boom were cashing out during a shaky quarter, so surely they knew something. A month later, in October 2024, Nvidia hit new all-time highs on demand for its Blackwell chips. That episode is worth remembering as this year's third-quarter insider filings get tallied. The timing of a sale tells you when a decision was made, and that decision is usually months older than the market mood around it. If you hold US tech through a Canadian ETF or mutual fund, as plenty of Edmonton investors do, knowing this changes how you react.
Most insider sales were decided before the turbulence
Huang's sales ran under a Rule 10b5-1 plan. Executives set these up in advance, choosing how many shares to sell and on what schedule, then let the plan run on its own. The plan protects them from accusations of trading on information the public doesn't have. Every trade is still reported to the US Securities and Exchange Commission on a Form 4, which is why the totals become news.
So selling during a rough quarter mostly reflects a choice made in calmer weeks. Tech leaders are also paid mainly in stock. When shares vest, a tax bill arrives, and selling some of those shares is often the only practical way to pay it. Founders and executives at Amazon and Meta sell regularly for the same reasons: taxes, spreading their money around, buying property. Someone trimming a sliver of a fortune held almost entirely in one company is managing personal risk, and the trade says little about next quarter.
Before reading meaning into a sale, check whether it ran under a pre-set plan. If it did, the date the plan was adopted matters more than the date of the trade.
The quieter signal is the buying that didn't happen
Selling has a dozen dull explanations. Buying has one. An insider who spends their own cash on open-market shares thinks the price looks cheap. During that 2024 quarter, nobody at Nvidia did. That didn't predict a fall, and the stock went on to rise. It does tell you that the people with the best view didn't see a bargain at those prices.
Heavy, visible selling can also weigh on a stock for a while, because other investors get wary of buying what insiders are unloading. That pressure comes from investor nerves and tends to fade, but it can leave a share price feeling stuck for a stretch.
Watch for the buying that's missing at least as closely as the selling that shows up.
Selling on a schedule is the habit worth copying
The most useful part of the Nvidia story may be the behaviour itself. Executives with most of their wealth tied up in one asset decide in advance to trim it on a schedule. Once that's set, a scary week can't talk them out of it or push them into something rash.
Plenty of Alberta investors are more concentrated than they realise. It might be a single tech fund that has run up, a rental property that now holds most of their net worth, or a business that doubles as the whole retirement plan. Turning part of that into something else is the careful choice, whether it becomes a down payment, a smaller mortgage balance or a broader portfolio. It's far easier to do with rules you set while you were calm.
Write your rule down before the next turbulent quarter. Decide how large a share of your wealth any one asset is allowed to become, and what you'll sell, and when, if it crosses that line.
Read Next
Five Money Checks Canadian Snowbirds Should Finish Before Heading South
Canadian Bank Security Assumes a Human Is Logging In. AI Agents Break That
Before Storm Winds Arrive, Turn 5 Yard Hazards Into Protective Buffers
A 25 Basis Point Hike Doesn't Erase Variable's Long-Run Edge Over Fixed