Big Tech Lost $900 Billion as AI Moats Face a Stress Test

By Daniel Okafor

Big Tech lost nearly $900 billion after third-quarter AI spending rose sharply, the Financial Times calculated. The two-day technology rout tested whether companies funding AI can defend their cash flows.

Alphabet started the rout. Shares fell 12.6% Thursday, erasing about $500 billion after quarterly results showed weak core-search revenue and faster cash burn. FactSet expected negative $2.23 billion in free cash flow; Alphabet reported negative $9.1 billion. Axios said part of the search shortfall came from less-profitable “AI mode” searches.

ServiceNow, the S&P 500’s worst performer, fell 27% on slowing revenue and weak guidance, Barron’s reported. Amazon and Intel came under pressure Friday; Intel was down about 25% premarket, and Reddit fell more than 22% as it moved toward a logged-in experience amid AI’s pressure on search.

How did AI turn the moat into a financial test?

Business language has used castles for decades. Jamie Dimon invoked a fortress balance sheet in a 2015 JPMorgan shareholder letter, then warned in 2021 that its moats were not particularly deep. Warren Buffett told Fortune with Carol Loomis in 1999 that investors should favor durable competitive advantages and wide moats.

“With AI disruption potentially around the corner, there is a lot of discussion about moats at the moment,” Apollo chief economist Torsten Slok told clients after Alphabet fell. The issue was whether AI weakens current advantages or creates stronger ones.

Slok built on New York University professor Aswath Damodaran, the dean of valuation. Damodaran lists brand, switching costs, network effects, cost advantages and legal protection as moat sources. Slok’s AI test asks:

• Does the company own data rivals cannot get?
• Is its product embedded in how customers work every day?
• Can it raise prices without customers walking away?

Hyperscalers raised $326 billion in stock and bonds this year, more than triple 2025, S&P Global Market Intelligence data showed. AI pressured cash-rich search businesses while owners redirected cash into AI infrastructure.

OpenAI’s GPT-6.1 Terra costs up to half as much per token as its predecessor. Ramp’s index of what companies pay Anthropic and OpenAI showed token prices down roughly 35% since end-June; Axios called it a price war.

In an Oct. 24 note, Slok argued that AI had produced no financial payoff for corporate America outside tech. He said revenue per worker and profit margins were rising for the Magnificent Seven but flat for the rest of the S&P 500, as token spending became a cost center rather than a growth source.

Why did capitalism adopt castles and vampires?

Horror took 15.5% of the domestic box office in 2025, according to The Numbers. U.S. horror-book sales rose 7% in the first nine months of 2025, according to Circana BookScan, while adult fiction grew 1%. On Friday, The New York Times asked why every book was gothic, noting labels on family sagas and space operas.

Slate called Nathan Fielder’s You Can See Everything twisted; The Hollywood Reporter framed it as gothic romance in the tradition of Alfred Hitchcock. For roughly 250 years, castles, moats, vampires and laboratory-born creatures have mirrored economic fears.

Horace Walpole, son of UK Prime Minister Robert Walpole, inherited enough wealth to draw a salary without working. He published The Castle of Otranto in 1764, now generally regarded as the first gothic novel.

Walpole passed the book off as an obscure Italian translation supposedly found in a northern English Catholic family’s library and printed in Naples in 1529. A 2014 Çankaya University paper by Ertuğrul Koç and Neslihan Atcan Altan called it a dramatization of aristocrats displaced by a greedy new bourgeoisie.

Literary historian Robert Miles linked the modern novel to print technology and the circulating library, which Walpole’s false provenance mimicked.

In her 1831 introduction to Frankenstein, Mary Shelley wrote that the story began in June 1816 after she heard Lord Byron and Percy Shelley discuss galvanic experiments. A 2018 Live Science piece compared the arguments over electrical life with those over AI.

Karl Marx made the vampire official in Capital in 1867, calling capital “dead labour, that, vampire-like, only lives by sucking living labour.” The image ran across 30 years of his writing. Bram Stoker’s Dracula followed in 1897, as farm depression threatened aristocratic estates and cheap farm goods from booming American farms flooded Britain.

Franco Moretti read Stoker’s count as a rational entrepreneur in a 1982 New Left Review essay: he invested gold to expand his dominion and conquer the City of London.

Manchester Centre for Gothic Studies researcher Xavier Aldana Reyes told the Times that gothic interest surges amid social anxiety and political turmoil. Literary agent Mary C. Moore described the genre as exposing what is rotten beneath something shiny and wealthy.

Where do software credits resemble zombie firms?

Matt Taibbi’s 2009 Rolling Stone story “The Great American Bubble Machine” branded Goldman Sachs a vampire squid wrapped around the face of humanity. Then-CEO Lloyd Blankfein conceded the article had tapped into something.

Bank for International Settlements researchers Ryan Banerjee and Boris Hofmann define zombie firms as businesses at least 10 years old whose earnings miss interest payments for three straight years.

Across 14 advanced economies, zombie firms’ share of listed nonfinancial companies rose from 4% in the late 1980s to 15% in 2017, the researchers found. Weak banks that roll over loans instead of writing them off keep them alive; lower rates tend to raise their share.

PitchBook counted $131 billion of leveraged loans below 80 cents on the dollar, a common distress threshold and the most since May 2020. Software accounts for 32% of those loans but 12% of the index; private equity backs 84% of stressed software companies.

Zando launched horror imprint Evil Twin, planning six to eight titles a year. Warner Bros.’ R-rated body-horror film Clayface opened No. 1 the weekend of Oct. 23 with about $24 million, according to The Numbers.

Inkfluence AI said 3,929 horror books had been started on its fiction-writing platform during the 180 days through Sept. 12. Haunted-house and ghost premises led at about one-third.

The next test is measurable. Investors will look for AI-driven gains in revenue per worker and margins beyond the Magnificent Seven, Slok said. Until then, cheaper tokens, heavier borrowing and a $326 billion capital bill leave AI castles dependent on cash and credit they have not yet produced.