Markets are heading into a particularly gothic Halloween as the artificial-intelligence trade exposes what may be a weakening aristocratic structure at the heart of early-21st-century capitalism. Wall Street has described itself in castle-adjacent terms for years.

Jamie Dimon has long discussed JPMorgan’s “fortress balance sheet” with shareholders. His 2015 letter included a section explaining what the phrase meant. By 2021, Dimon had added another castle-related term, warning that “the moats that protect this company are not particularly deep.”

The idea of a corporate moat came from Warren Buffett. In a 1999 Fortune article with Carol Loomis, Buffett argued that the key to investing was a company’s competitive advantage and the durability of that advantage. The businesses worth owning, he said, had wide moats around them.

The gothic novel’s defining image of danger is a castle falling apart. For roughly 250 years, the genre and the language of capitalism have traded in related images: the castle and the moat, the vampire and the creature built in a laboratory. This Halloween season, those images have moved to the front of the conversation as a two-day rout in technology stocks raised questions about whether the AI giants’ moats can hold, while gothic horror has surged across bookshelves and movie screens.

Horror accounted for 15.5% of the domestic box office in 2025, according to The Numbers. U.S. horror-book sales rose 7% during the first nine months of 2025, according to Circana BookScan, while adult fiction overall grew 1%. The New York Times asked Friday why every book seemed to be gothic, with publishers applying the label to works ranging from family sagas to space operas. Several film critics also described Nathan Fielder’s business-scandal documentary You Can See Everything as a twisted, even gothic romance in the tradition of Alfred Hitchcock.

The moats get tested

The selling began with Alphabet. The stock fell 12.6% Thursday, erasing roughly half a trillion dollars in market value after the Google parent reported its quarterly results. Revenue from its core search business came in below expectations, partly because of the popularity of less-profitable “AI mode” searches and the company’s faster-than-expected cash use. Alphabet’s free cash flow—the cash left after operating expenses and capital investments—was negative $9.1 billion, compared with the negative $2.23 billion analysts had expected, according to FactSet. Axios reported on the stock’s decline.

Software giant ServiceNow was the S&P 500’s worst performer, falling 27% the same day on slower revenue growth and disappointing guidance. By Friday morning, it was Amazon’s and Intel’s turn. Intel was down roughly 25% before the opening bell. Big Tech lost nearly $900 billion in market value after another sharp increase in AI spending during the third quarter, according to a Financial Times calculation reported by Fortune. Nasdaq 100 companies lost more than $1 trillion in market value Friday alone.

Reddit, whose shares fell more than 22% in premarket trading Friday, is pulling up the drawbridge by moving toward a logged-in experience as AI eats into search. The irony is unsettling investors: AI is putting pressure on the businesses that generate cash, while that cash is still being directed into AI.

Hyperscalers—the largest cloud-computing operators—have raised $326 billion through stock and bond issuance this year to fund the build-out, more than triple the amount raised in 2025, according to S&P Global Market Intelligence data.

The underlying question was posed by Apollo chief economist Torsten Slok on the morning Alphabet fell. “With AI disruption potentially around the corner, there is a lot of discussion about moats at the moment,” Slok wrote in his Thursday Daily Spark note.

Slok built on the work of finance professor Aswath Damodaran, known as the “dean of valuation.” Damodaran has identified several sources of competitive advantage, including brand strength, switching costs, network effects, cost advantages and legal protection. Slok reduced the issue to a three-part AI moat checklist: Does a company own data its rivals cannot obtain? Is its product part of how customers perform their jobs every day? And can it charge more without customers leaving?

By the third test, the companies selling AI do not appear to measure up. OpenAI’s new GPT-6.1 Terra model costs up to half as much per token as the previous version. An index from expense-management firm Ramp, based on what companies actually pay Anthropic and OpenAI, shows token prices down roughly 35% since the end of June, in what Axios described as a price war. Axios reported on the decline in AI token prices.

The companies buying AI have not yet received much financial benefit, either. In an Oct. 24 note, Slok argued that AI “has delivered zero financial payoff to corporate America outside of tech.” Revenue per worker and profit margins are rising for the Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—but remain flat for the rest of the S&P 500, he wrote. Mainstream companies are beginning to treat token spending as a cost center that weighs on margins rather than as a source of growth.

The gauges for that checklist are public and regularly updated: Ramp’s index tracks what companies actually pay for AI tokens, while S&P Global Market Intelligence tallies the financing behind the build-out.

The first castle was a fake

Xavier Aldana Reyes, a researcher at the Manchester Centre for Gothic Studies, told The New York Times that interest in gothic themes tends to rise during periods of social anxiety and political turmoil. Literary agent Mary C. Moore described the genre’s purpose more directly, telling the newspaper: “Gothics are all about exposing what is rotten at the core of something shiny/wealthy.”

The genre was created by a wealthy insider who understood the old order and received a salary without working. Horace Walpole was the son of UK Prime Minister Robert Walpole. His The Castle of Otranto, published in 1764, is generally considered the first gothic novel.

Walpole presented the book as a translation of an obscure Italian work discovered in the library of an old Catholic family in northern England and printed in Naples in 1529. In reality, it dramatized an aristocratic system being displaced by what a 2014 paper by Çankaya University researchers Ertuğrul Koç and Neslihan Atcan Altan called the “greedy” new bourgeoisie. The paper is available through Çankaya University.

Literary historian Robert Miles has written that the modern novel was understood to have been revolutionized by print technology, which made circulating libraries possible and helps explain Walpole’s pre-modern false origin story.

The gothic tradition soon produced other stories of science, power and social disruption. In her 1831 introduction to Frankenstein, Mary Shelley wrote that the idea came to her in June 1816 after she overheard Lord Byron and Percy Shelley discussing galvanic experiments. A 2018 article republished by Live Science drew a direct comparison, arguing that electrical life provoked a range of debates similar to those surrounding AI today.

Karl Marx made the connection to capital explicit. In Capital, published in 1867, he wrote that capital was “dead labour, that, vampire-like, only lives by sucking living labour.” The image appeared repeatedly across three decades of his writing.

Bram Stoker’s Dracula established the great modern vampire myth in 1897. It appeared as an agricultural depression threatened the fortunes and estates of Britain’s noblemen while cheap agricultural products flowed in from expanding American farms. Literary critic Franco Moretti interpreted the count as a businessman. In a 1982 essay in New Left Review, Moretti described Stoker’s Dracula as a rational entrepreneur who invested his gold to expand his dominion and conquer the City of London.

The undead on Wall Street

In more recent times, Matt Taibbi’s 2009 Rolling Stone article “The Great American Bubble Machine” called Goldman Sachs “a great vampire squid wrapped around the face of humanity.” The phrase remained prominent enough that then-Goldman CEO Lloyd Blankfein acknowledged in August of that year that the article had “tapped into something.”

Economists developed their own monster: zombie firms. Bank for International Settlements researchers Ryan Banerjee and Boris Hofmann define zombies as companies at least 10 years old whose earnings have failed to cover their interest payments for three consecutive years. Across 14 advanced economies, the share of listed nonfinancial companies classified as zombies rose from 4% in the late 1980s to 15% in 2017, according to the researchers’ BIS paper.

The researchers found that weak banks can keep zombie companies alive by rolling over loans instead of writing them off. They also found that lower interest rates tend to increase the share of zombie companies.

The newest candidates are in software. PitchBook counted $131 billion in leveraged loans priced below 80 cents on the dollar, a common threshold for distress and the highest level since May 2020. Software companies accounted for 32% of those distressed loans while making up only 12% of the index. Private-equity firms backed 84% of stressed software companies.

These are castles unable to pay for their own upkeep, kept standing by someone else’s money—a dollar-princess arrangement, recalling the American heiresses whose fortunes propped up the estates of Britain’s landed aristocracy, with a credit line in place of a dowry.

The anxiety is also driving business during the haunted boom of 2026. Book publisher Zando launched a horror imprint called Evil Twin and plans to publish six to eight titles a year. In Hollywood, Warner Bros.’ R-rated body-horror film Clayface opened at No. 1 during the weekend of Oct. 23 with approximately $24 million, according to The Numbers.

The machines are writing horror, too. Inkfluence AI, a fiction-writing platform, says 3,929 horror books were started on the service during the 180 days ending Sept. 12. Haunted-house and ghost premises led the categories, accounting for about one-third of the projects. The platform’s AI horror-story generator is among the tools now participating in that trend.

Generative AI was used for research assistance and/or transcription of this article. The reporter independently reported and verified the factual claims, and a human editor reviewed the article before publication. The article was updated with the Nasdaq movement on Friday.

This story was originally featured on Fortune.com.