Here is the question worth sitting with this morning: if Nvidia’s business were fully valued at $5.4 trillion, why would its board just authorize what Nvidia described as the largest increase to a share repurchase program in U.S. corporate history?
On September 28, 2026, Nvidia’s board approved a $150 billion increase to its existing repurchase program, which the company said was the largest authorization increase on record. That brings the total remaining authorized repurchase amount to $235 billion, with the company expecting to execute the program through fiscal year 2028. The $150 billion increment alone is nearly double the $80 billion the board added in May.
Why This Matters Now
The announcement landed on a punishing macro day. The 10-year Treasury yield briefly topped 5.27% before ending the day around 5.23%, a level not seen since 2007. Against that backdrop, management chose to lean in rather than pull back. That sequencing is the signal.
This is not a first move from a standing start. Nvidia repurchased 203 million shares for $39.8 billion during the first half of fiscal year 2027. That pace is up roughly 64% from $24.2 billion a year earlier. The company has been doing this at scale and accelerating. The new authorization extends the runway, but the execution record is what gives it credibility.
The Investment Thesis
The standard worry about large buybacks is that they arrive when growth is fading and management has run out of better uses for cash. Sometimes companies increase their capital returns when they are running out of growth opportunities, but the stepped-up capital returns here suggest management expects sufficient growth to enable it to return a larger share of cash to shareholders.
The underlying business supports that reading. Q2 sales surged 106% to $96.2 billion, with Data Center sales jumping 117% to $89.0 billion. Adjusted earnings reached $2.22 per share, up 120%, while gross margin stood at 75%. Q3 revenue guidance stands at $108.0 billion, plus or minus 2%. At that revenue rate, this is a company that can fund both an AI arms race and a record capital return program simultaneously.
CFO Colette Kress said that against a plan to return 50% or more of free cash flow, the company had returned 60% year to date, adding that going forward Nvidia intends to increase and return excess free cash flow net of strategic uses. That language, “excess,” is the operative word. Management is saying the AI reinvestment budget is handled, and what is left over is getting returned.
The Risks
The $150 billion is an authorization: permission to repurchase shares, with the pace left to management. An authorization is not a commitment. Nvidia repurchased $19.7 billion of stock in Q2 alone, so the mechanical capacity to execute is real, but conditions change. In Q2, days sales outstanding rose to 60 days as Nvidia extended payment terms for certain large, investment-grade customers. If working capital continues to absorb cash, the pace of actual buybacks could slow even as the authorization looks large on paper.
Rate risk compounds this. With the 10-year briefly above 5.27%, every dollar returned via buyback is being weighed against a risk-free alternative that did not exist eighteen months ago. Nvidia’s stock still needs to outrun a higher hurdle rate to justify repurchases over debt reduction or pure cash preservation.
What Investors Should Watch Next
The Q3 earnings report, expected in November, will show the actual pace of repurchase execution. Nvidia’s original buyback program, announced on August 9, 2004, authorized just $300 million. The latest increase alone is 500 times that amount. The trajectory is extraordinary, but the stock’s response will ultimately depend on whether management converts authorization into consistent, visible buying, particularly on days like yesterday when yields spike and the AI complex pulls back. Those are precisely the conditions under which an aggressive repurchase program can do the most work for remaining shareholders.
Management’s read appears to be that the market is underpricing the duration and scale of the AI infrastructure cycle. The $235 billion authorization is, at minimum, the most expensive bet they have ever made on themselves.
