Buybacks Did Not Die. The Free Buyer Did.

Apple bought back $36.989 billion of stock in the six months ended March 28, 2026, according to Apple’s SEC companyfacts data accessed June 22, 2026. Meta bought back zero in Q1 2026, after $12.754 billion in Q1 2025, according to Meta’s SEC companyfacts data accessed June 22, 2026. Amazon spent $44.203 billion on productive assets against $26.032 billion of operating cash flow in Q1 2026, according to Amazon’s SEC companyfacts data accessed June 22, 2026.

That is the setup. Not the conclusion.

The lazy version says the buyback era is over. That is too clean. Apple disproves it. The better version is harsher: the free buyer is no longer free. The largest technology companies can still buy stock, but now they have to fight their own AI infrastructure budgets for the cash.

For years, the market did not have to think very hard about this. Big tech printed cash. It bought back stock. Share counts came down. Earnings per share got help. Multiples got a quiet bid.

Now the same companies are writing enormous checks for data centers, chips, power, servers, and cloud capacity. The cash did not disappear. It got reassigned.

The cash-flow table that matters

The three companies do not tell the same story. That is the point.

Company	Current cash-flow fact	Buyback fact	What it proves	What it does not prove
Apple (AAPL)	$82.627 billion of operating cash flow and $4.344 billion of capex for the six months ended March 28, 2026, per Apple SEC data	$36.989 billion of repurchases in the same six-month period, after $90.711 billion in fiscal 2025 and $94.949 billion in fiscal 2024, per Apple SEC data	Apple is still a buyer of its own stock	It does not prove Apple can keep the old pace forever
Meta (META)	$32.226 billion of operating cash flow and $18.997 billion of capex in Q1 2026, per Meta SEC data	$0 of repurchases in Q1 2026, down from $12.754 billion in Q1 2025, per Meta SEC data	Meta is the cleanest buyback pause in the group	It does not prove Meta will issue stock
Amazon (AMZN)	$26.032 billion of operating cash flow and $44.203 billion of productive-asset spending in Q1 2026, per Amazon SEC data	$0 of repurchases in fiscal 2024 and fiscal 2023, after $6.000 billion in fiscal 2022, per Amazon SEC data	Amazon is already an infrastructure-spend story, not a buyback story	It does not prove the spending is bad

The table is not a bear case by itself. Apple is still buying. Meta may restart. Amazon has never been the clean buyback machine Apple was.

But the table does show a regime change. Buybacks are no longer the default use of excess cash. They are one bidder in a room full of larger bidders.

AI capex is the loudest bidder.

Apple is the counterargument, and that matters

If someone says buybacks are dead, start with Apple.

Apple (AAPL) still bought $36.989 billion of stock in the six months ended March 28, 2026, and it still had $45.572 billion of cash and equivalents at quarter end, according to Apple’s SEC data accessed June 22, 2026. That is not a dead buyer.

But the direction is not clean either. Apple bought back $90.711 billion in fiscal 2025, down from $94.949 billion in fiscal 2024, according to the same SEC data. That is not a collapse. It is a slower machine.

The distinction matters. The market does not need Apple to stop buying tomorrow for the math to change. It only needs the largest buyer to become less aggressive at the margin.

Margins are where multiples live.

Meta is the cleaner warning

Meta (META) is a better test case because the repurchase line went to zero.

In Q1 2025, Meta repurchased $12.754 billion of common stock. In Q1 2026, that number was zero, while capex rose from $12.941 billion to $18.997 billion, according to Meta’s SEC data accessed June 22, 2026. That is a 46.8 percent increase in capex and a complete stop in buybacks for the quarter.

This is where the market has to be careful. Zero buybacks in one quarter is not the same thing as permanent dilution. It is not a secondary offering. It is not a confirmed board decision to sell stock.

It is still important.

When a company that used to retire shares decides not to retire any, the per-share tailwind is gone for that period. Earnings can still grow. The stock can still rise. But the old mechanical help is absent.

That is not analysis. That is arithmetic.

Amazon shows the endgame

Amazon (AMZN) is what the model looks like when infrastructure eats the cash first.

In Q1 2026, Amazon generated $26.032 billion of operating cash flow and spent $44.203 billion on productive assets, according to Amazon’s SEC data accessed June 22, 2026. Productive-asset spending was about 170 percent of operating cash flow.

That is not a buyback setup. That is an infrastructure setup.

Amazon also reported $101.816 billion of cash and equivalents, $119.074 billion of noncurrent long-term debt, and $2.832 billion of current long-term debt at March 31, 2026, according to the same SEC data. The company has resources. The point is not distress. The point is priority.

The priority is capacity.

What the market is saying

Public market prices do not look scared. Apple closed at $300.0092, Meta at $561.04, and Amazon at $233.69 on June 22, 2026, according to public market price history checked June 22, 2026. Options data checked on June 18, 2026 showed put-call volume ratios below 1.0 for Apple, Meta, and Amazon, and expected earnings moves of 5.8439 percent for Apple, 10.0026 percent for Meta, and 7.8295 percent for Amazon.

The options market is not screaming.

Polymarket was also not screaming. It showed 53 percent odds that OpenAI IPOs by December 31, 2026, 14 percent odds by September 30, 2026, and 20 percent odds for an AI bubble burst in 2026, checked on June 22, 2026 via Polymarket’s public market search.

So the market is saying two things at once. It believes the AI capital cycle is real. It does not yet treat the buyback tradeoff as a major valuation problem.

Maybe that is right. Maybe AI capex earns high returns and the missing buybacks do not matter.

But that is the bet.

The episode claim should stay conditional

The June 16, 2026 What Are Your Thoughts discussion asked a sharp question: what if Apple, after years of buying back stock, decided to sell a large block instead? The hosts said many large companies had been “sucking up every share” and now faced “a different world,” in segment 60 of the episode.

That is a useful framing. It is not a filing. It is not an Apple announcement. It is not proof that Meta will sell stock.

Treat it as a warning label, not a fact pattern.

The fact pattern is narrower and stronger. Some of the biggest technology companies now have a cash-use problem they did not have in the old software era. They can fund AI infrastructure. They can buy back stock. They can preserve the balance sheet.

Doing all three at the old scale is harder.

What this does not tell you

It does not tell you that Apple has stopped buying stock. Apple has not. The company repurchased $36.989 billion in the six months ended March 28, 2026, according to Apple’s SEC data.

It does not tell you that Meta will issue shares. Meta reported zero Q1 2026 buybacks, not a secondary offering, according to Meta’s SEC data.

It does not tell you that AI capex is wasted. Amazon’s $44.203 billion of Q1 2026 productive-asset spending could be rational if it creates durable capacity and revenue, according to Amazon’s SEC data. The filing proves the cash went out. It does not prove the return.

It does not tell you full-year 2026 capital allocation. The cleanest data here is Q1 2026 for Meta and Amazon, and six months ended March 28, 2026 for Apple. The next few quarters matter.

FAQ

Doesn’t Apple prove the buyback era is still alive?

Yes. Apple is the best counterargument. It bought $36.989 billion of stock in the six months ended March 28, 2026, according to Apple’s SEC data. That is why the right claim is not “buybacks are dead.” The right claim is that the automatic buyer is less automatic.

Why does a buyback pause matter if earnings still grow?

Because buybacks help earnings per share. If net income rises but share count stops shrinking, per-share growth loses a tailwind. That does not kill the stock. It changes the multiple investors should pay.

Is a buyback pause the same as share issuance?

No. A pause stops reducing the share count. Issuance increases it. Meta’s Q1 2026 data shows a pause, not issuance, according to Meta’s SEC data. A real equity sale would be a much larger supply event.

What would change the bearish version of this thesis?

Two things. First, AI capex starts producing clear, high-return revenue. Second, buybacks resume without stretching balance sheets. If that happens, the market is right to look through the spending.

What is the verdict?

Revenue growth alone is not the question anymore. Returns on the spend are.