- Two Headlines, One Decision
- What a Buyback Actually Announces
- The Buyback Hurdle Rate Test
- Why Nvidia Probably Clears the Bar
- Why Boeing Did Not
- The Three Questions That Separate Them
- The Successor Test
- People Also Ask
- About the Stagnation Assassin
Two Headlines, One Decision
On September 28, 2026, two stories landed that almost nobody connected. Nvidia authorized the largest share repurchase increase in corporate history. Boeing learned its newest jet would not be certified because of a software bug. Same day, same decision, opposite verdicts, and the connecting logic went unexamined.
Nvidia’s board authorized an additional $150 billion under its existing share repurchase program, bringing the total remaining authorization to $235 billion. The company said it expects to execute the remaining program through fiscal year 2028. Jensen Huang framed it as confidence in the long term opportunity ahead, and noted that the company’s cash generation gives it capacity both to invest in the technology and to return capital to shareholders.
The same day, the head of the Federal Aviation Administration said certification of the Boeing 737 Max 10 would be delayed while the agency assesses a newly disclosed software defect. The bug sits in the 737 Max flight management computer operating system and can cut off automated flight guidance if pilots have to execute a go around. Boeing shares shed nearly 7 percent. CEO Kelly Ortberg had told investors earlier that same month that certification was expected very soon. The approval process had already stretched on for years.
Most readers will file these as unrelated. One company is winning. One company is losing. Move on.
They are the same story told from opposite ends of the same decade, and the thing that connects them is the most misunderstood decision in corporate finance.
I break down the four operator moves for testing a valuation model in my podcast episode A DCF Model Is Dressed-Up Guessing.
What a Buyback Actually Announces
A buyback is not a reward, a gift, or a financial engineering trick. It is a public statement about hurdle rate. When a company repurchases its own stock, it is claiming that no internal project, acquisition, segment, or capability available to it earns a higher return than buying itself. That claim deserves proof.
Let me say the unpopular part first, because I am not interested in the lazy version of this argument. Buybacks are not evil. The people who want them banned have not thought it through.
Your job, the only job, is to make as much money as you possibly can, every minute of every day. A dollar sitting on your balance sheet has exactly one obligation: go where it earns the most. If the highest returning use of that dollar is buying your own stock, then buy your own stock. Buy it aggressively. Buy it without apology and without a press release full of hedging.
That is a legitimate answer. It is sometimes the correct answer. And a management team with the discipline to say “we found nothing better, so we are returning it” is behaving more honestly than a management team that empire builds into adjacencies it has no right to win, purely so nobody accuses it of lacking vision.
But notice what that legitimate answer actually requires. It requires that the company examined every internal project, every acquisition target, every new segment, every piece of capacity, every engineering investment on the table, and found that none of them clears the return available from repurchasing its own shares.
Nothing available to us earns more than this.
That is the threshold. Not “we have extra cash.” Not “the stock looks cheap.” Not “shareholders expect a return of capital this quarter.” The threshold is a comparison against the full opportunity set, and the announcement is the company asserting it ran that comparison and lost.
I have never once heard an analyst ask the only question that matters on a buyback call: what was the next best use of this money, and what would it have returned instead? Until somebody asks that out loud, a $235 billion authorization is a claim that nobody has tested.
Here is what should make every operator uncomfortable. The investment community does not hold buybacks to that standard either. Analysts model the accretion to earnings per share, note the reduced share count, adjust the target price, and move to the next question. The comparison against the alternative almost never happens in public. Which means the single largest capital allocation decision most mature companies make is the one facing the least scrutiny.
The Buyback Hurdle Rate Test
The test has five steps and a binary verdict. Inventory every available use of capital. Score each on expected return. Compare the best alternative against the return on your own shares. Correct for time horizon, because decade length payoffs always score poorly on a quarterly lens. Break close calls toward whatever builds durable position.
This is not complicated arithmetic. It is uncomfortable arithmetic, which is why it gets skipped.
Notice that four of the five steps have nothing to do with the buyback itself. The repurchase is the benchmark, not the subject. What you are really doing is forcing an honest inventory of where your company has a genuine right to win, and then refusing to let the easy option win by default.
Why Nvidia Probably Clears the Bar
Run the test on Nvidia and the case holds up. The company is buying an asset it understands better than any other, at a multiple below the broad market, in a business where the demand side is not the constraint. The reinvestment runway may be genuinely saturated. That is a defensible hurdle rate claim.
Start with the price. Nvidia stock has climbed 24 percent over the past twelve months, lifting market capitalization to $5.42 trillion. One former Wall Street analyst noted that the company trades near 18.7 times forward earnings, meaning management is buying its own shares at a multiple cheaper than most of the S&P 500.
Now think about what that asset actually is. Nvidia is purchasing the one business on earth it understands better than any outside buyer possibly could, at a valuation below the market average, in a sector where combined hyperscaler capital expenditure is projected by S&P Global Ratings to exceed $1.3 trillion by 2027.
The demand side is not Nvidia’s constraint. Absorption is. There is a hard physical limit to how fast you can stand up fabrication capacity, hire and integrate engineers, and close acquisitions without wrecking the thing that made you good in the first place. When the constraint on reinvestment is absorption rather than opportunity, excess cash genuinely has nowhere better to go.
Huang’s own framing gets the sentence structure right, and the structure matters more than the sound bite. Cash generation gives the company capacity to invest in the technology first, and to return capital second. Investment, then remainder. That is the correct order, and it is the order most buyback announcements quietly invert.
I will flag one thing that keeps this from being a clean pass. In June 2026, Nvidia issued an aggregate of $25 billion of senior unsecured notes across seven tranches for general corporate purposes. Borrowing while repurchasing can absolutely be correct if the spread works in your favor. It is also precisely the moment where the hurdle rate question stops being rhetorical and becomes arithmetic. If you are borrowing at one rate and buying an asset yielding another, show me the gap on the slide.
Nvidia can probably show it. I would still like to watch somebody make them.
Why Boeing Did Not
Between 2013 and 2019, Boeing repurchased $43.5 billion of its own stock. In 2017 alone the figure was $9.2 billion, against capital expenditures of $1.7 billion. That ratio was a hurdle rate decision made seven consecutive times, and the certification delay now on the front page is the return on the road not taken.
The year by year record, drawn from Boeing’s own filings and a Leeham News analysis of the full program, looks like this:
| Year | Share repurchases |
| 2013 | $2.8 billion |
| 2014 | $6.0 billion |
| 2015 | $6.8 billion |
| 2016 | $7.0 billion |
| 2017 | $9.2 billion |
| 2018 | $9.0 billion |
| 2019 | $2.7 billion |
| Total | $43.5 billion |
Both ends of that table come straight from Boeing’s annual reports. The 2013 filing records 25.4 million shares repurchased for $2.8 billion, with no repurchases at all in 2012 or 2011. The 2019 filing records 6.9 million and 26.1 million shares totaling $2.7 billion and $9.0 billion for 2019 and 2018 respectively.
Now set that next to what the company was spending on itself. Boeing’s capital expenditures totaled $1.8 billion in 2019, relatively consistent with $1.7 billion in both 2018 and 2017.
In 2017 Boeing spent $9.2 billion buying its own stock and $1.7 billion on capital expenditures. Five and a half dollars into the share count for every single dollar into the physical business. That was not a financing decision. That was a hurdle rate decision, and the bill arrived right on schedule.
Understand what that ratio asserts. Every year, for seven straight years, Boeing’s leadership was effectively claiming that no available investment in engineering capability, certification throughput, supplier quality, software rigor, or manufacturing capacity could beat the return on repurchasing shares. Nobody framed it that way out loud. That is exactly the problem. The claim was made by default rather than by analysis, and it went unchallenged because the share price cooperated.
We now know what those alternative investments were worth, because we are watching the invoice arrive. The current defect sits in the 737 Max flight management computer operating system, discovered by Boeing in August, and it can cut off automated flight guidance when pilots have to perform a go around. It affects the Max 7 as well as an undisclosed number of jets already in service. The FAA administrator said the agency has not concluded whether it is a safety of flight issue, but will delay the Max 10 until it is satisfied. The certification process had already stretched on for years before this.
A software defect in a flight management system is not bad luck. It is a return on investment, and a deeply negative one, realized in an engineering capability that was funded at roughly $1.7 billion a year while $9.2 billion went out the door to the share count. Anyone who has run a genuine operational turnaround recognizes the pattern immediately: the capability you declined to fund does not disappear, it just sends its bill later, with interest, at the worst possible moment.
The ending was not subtle either. Boeing’s recorded debt went from $13.8 billion at the end of 2018 to $27.3 billion at the end of 2019, with $17.3 billion still sitting unused under the repurchase authorization and the program suspended.
Same instrument as Nvidia. Same mechanics. Opposite verdict. The difference was never the buyback. The difference is whether anybody ran the test.
I Have Been On The Wrong Side Of This Test
I am not writing this from the cheap seats. I ran the same broken arithmetic with my own money, at a scale where I had no excuse and nobody else to blame.
I bought a wood flooring retail store at the COVID-inflated peak. Then tariffs landed on the category. Then I discovered, slowly and expensively, that I did not understand retail. Every one of those is a real problem. None of them is the actual problem.
The actual problem is that I never ran step one. I had capital and I had an opportunity set, and I never honestly inventoried it. The business where I held a genuine right to win was the industrial tank liner manufacturer I already understood down to the shop floor. Putting more money into the thing I knew how to operate was sitting right there on the list, unscored, because I never wrote the list. I bought what was available instead of what returned the most, and I told myself the purchase price made it a good decision.
That is the same failure as a lazy buyback. Different instrument, identical mistake: the alternative never got a number, so the easy option won by default. Boeing did it with $9.2 billion. I did it with a flooring store. The math does not care about the size of the check.
The Three Questions That Separate Them
Three questions convert the hurdle rate test into something you can run on Monday. Which segments can absorb the capital at a better return? What time horizon do the alternatives carry? And on a genuinely close call, does the option leave behind a position a competitor cannot immediately copy?
First, which segments can actually absorb the capital? Every business has a small number of places where it holds a genuine right to win, and a much larger number of places where it does not. Map them honestly using a Right-to-Win Matrix and stop pretending the map is longer than it is. If a green segment exists that can absorb your cash at a better return than your own stock, and you bought stock instead, you misallocated. Full stop. If no such segment exists, the buyback is correct and you should defend it in public without flinching.
Second, what time horizon do the alternatives carry? This is where Boeing’s decision looked most defensible in the moment and worst in hindsight. Engineering capability, certification infrastructure, supplier development: these are ten year investments. Score them against a quarterly lens and they lose to a buyback every single time. They are designed to. Capital allocated to decade length payoffs will always look inferior to capital returned this quarter, right up until the decade actually arrives and presents itself at your door. Decade Allocation is not a philosophical preference. It is the correction factor that keeps your scoring model from systematically lying to you.
Third, does it build anything? When two uses of capital produce genuinely comparable returns, break the tie toward the one that leaves you holding a position a competitor cannot immediately copy. This is the Moat Mandate applied to the treasury. A buyback builds no moat. It cannot, by construction. It is a perfectly respectable financial transaction that produces exactly zero durable competitive advantage, and that is fine when nothing else clears the bar. It stops being fine the moment it becomes the default because it is easy, the market applauds, and nobody in the room has to defend a ten year number.
The Successor Test
Kelly Ortberg did not make any of Boeing’s capital allocation decisions between 2013 and 2019. He inherited them. Every week he now spends explaining a certification delay is a week spent living inside somebody else’s hurdle rate math, which is the clearest argument for the Inheritance Standard I have seen in years.
He inherited a certification backlog. He inherited a supplier base, a software organization, and a balance sheet that were all shaped by seven consecutive years of one particular answer to the hurdle rate question. He did not get a vote on any of it.
That is the test I would apply to every dollar, and it is the one almost nobody applies to a repurchase. Not the share price. Not earnings per share. The question is simpler and far more brutal: would I want my successor to inherit this? What is left standing after the check clears and the announcement scrolls off the page?
Nvidia’s answer to that question may well be yes. Huang’s stated rationale is confidence in the long term opportunity ahead, the balance sheet supports it, and the absorption constraint is real rather than invented.
Boeing’s answer, from 2013 through 2019, was also yes at the time. Everybody applauded. The stock cooperated beautifully for most of a decade.
The answer changed. Not because the buyback was immoral, and not because returning capital is a sin. It changed because the decision was never held to the only standard that counts, and by the time anyone got around to checking the math, the money was gone and the airplane still was not certified.
Every buyback is a confession. Make absolutely certain you know what yours is confessing.
People Also Ask
Executives searching this topic tend to ask four questions: whether buybacks destroy long term value, what return threshold a repurchase should clear, how to tell a disciplined buyback from a lazy one, and whether Nvidia’s authorization is comparable to Boeing’s. Short answers follow.
Are stock buybacks bad for long term company value?
No. A buyback is neutral on its own. It becomes destructive only when the company had a higher returning investment available and chose the repurchase anyway. The instrument is not the problem. The unexamined hurdle rate behind it is the problem.
What return should a share buyback have to clear?
It should clear the return on the single best alternative use of that capital, corrected for time horizon. If the best available internal or external investment earns more than repurchasing your own shares, the buyback is a misallocation regardless of how the share price responds.
How can investors tell a disciplined buyback from a lazy one?
Look at the ratio of repurchases to capital expenditures and research spending over several consecutive years. Then ask what the company declined to fund during that window. Boeing spent $43.5 billion on repurchases from 2013 to 2019 while capital expenditures ran near $1.7 billion annually.
Is Nvidia’s buyback comparable to Boeing’s?
Mechanically yes, strategically no. Nvidia is buying near 18.7 times forward earnings while facing a physical absorption constraint on reinvestment. Boeing was buying while underfunding the engineering and certification capability that later failed. Same instrument, different answer to the hurdle rate question.
About the Stagnation Assassin
Todd Hagopian is a Fortune 500 transformation executive who has generated $3B in shareholder value across Berkshire Hathaway, Illinois Tool Works, Whirlpool, and JBT Marel, where he serves as Global VP of Product Strategy and Marketing. Known as The Stagnation Assassin, he is the author of two published books: The Unfair Advantage: Weaponizing the Hypomanic Toolbox and Stagnation Assassin: The Anti-Consultant Manifesto. His blog is published in 15+ languages and read by operators worldwide. Bring him to your stage via his speaking page or connect with him on LinkedIn.
Your treasury is making a claim every quarter whether you test it or not. Run a 15-minute Capital Allocation Audit on your own hurdle rate: name every investment you declined to fund this year, score what it would have returned, and find out what your last allocation decision was actually confessing. Book the audit here.
See also: Home Depot Earnings: Two-Strategy Risk