Meta (NASDAQ:META) delivered 28% revenue growth last quarter, but its costs and expenses rose nearly twice as fast, surging 55%.
This gap, according to Steve Eisman, is getting hard to ignore.
“I just found it astonishing,” the investor of “The Big Short” fame said on the latest episode of The Real Eisman Playbook. “We want the reverse.”
Eisman’s question is whether Meta’s AI spending can generate enough profit to justify the rapidly rising cost base.
The Cost Curve Is Moving the Wrong Way
Part of the expense surge reflected unusual items. Meta booked $2.4 billion in legal charges and $1.18 billion in severance costs during the quarter.
According to Meta’s second-quarter 10-Q, research and development expenses jumped 67% to $21.66 billion, with the company citing higher employee compensation, data centers, technical infrastructure, third-party cloud services and AI token costs.
‘The Depreciation Is Starting to Really Explode’
Eisman said the depreciation tied to Meta’s infrastructure buildout is becoming increasingly important. “The depreciation is starting to really explode,” he said. “It’s like a weight on your shoulders.”
Server and network depreciation reached $4.62 billion in the quarter, up roughly 48% from $3.12 billion a year earlier, according to Meta’s second-quarter 10-Q.
And the spending is still accelerating. Meta reported $31.08 billion in quarterly capital expenditures and expects to spend $130 billion to $145 billion for the full year to support its AI efforts and core business. Free cash flow fell to $784 million from $8.55 billion a year earlier.
Meta also issued $25 billion of senior unsecured debt in May, bringing its long-term debt to $83.66 billion at the end of June.
Zuckerberg Says AI Is Already Paying Off
CEO Mark Zuckerberg has a direct answer: “AI is accelerating our core business today.” Advertising revenue rose 27% last quarter, ad impressions grew 14% and the average price per ad climbed 12%.
The debate may be less about whether AI works than whether those gains justify the capital burden.
Strategas Sees a Cleaner Trade in the Chips
Strategas founder Jason Trennert, appearing on the same episode, said AI demand still appears to exceed supply and hyperscalers are likely to keep spending, but he is unsure they will generate an “appropriate return” for shareholders. He warned of a possible “race to the bottom,” and is more bullish on semiconductors.
Prediction market traders also see little sign of an imminent AI bust. A Polymarket market with roughly $2.9 million in volume currently gives just a 12% chance that the AI bubble bursts by the end of 2026.
For Trennert, that makes semiconductor suppliers such as Nvidia (NASDAQ:NVDA) the cleaner side of the trade: hyperscalers can keep pouring billions into AI infrastructure, and chipmakers can keep getting paid, even if the returns ultimately disappoint Meta and its shareholders.
Image: Shutterstock
Login to comment