AI Just Broke The Stock Market

Summary

The video argues that AI is challenging the economics of software businesses by making applications cheaper and faster to build. This could threaten subscription revenue and per-user (“per-seat”) pricing. However, the presenter says investors may have overreacted: established software companies still benefit from reliability, proprietary data, customer relationships, and the difficulty of replacing systems embedded in large organizations.

Market impact and financial context

• The US software index reportedly fell more than 26% from late January to its April low. Estimates of the value erased from software companies ranged from $1 trillion to $2 trillion.
• The sell-off followed new AI tools and features, including Anthropic’s Claude Cowork and plugins. Thomson Reuters had its largest one-day drop, while LegalZoom fell nearly 20% in one session.
• Despite the downturn, the video says US software stocks were hitting fresh highs for the year by early October, while expected sector earnings growth had risen to over 20%. The presenter cautions that a near-term recovery does not rule out longer-term disruption.
• As an example of the sector’s past returns, the presenter says $10,000 invested in the biggest software ETF in April 2016 would have grown to just over $41,000 ten years later.
• The presenter compares AI’s potential effect on software to Napster’s disruption of recorded music. US recorded-music revenue fell from $14.6 billion in 1999 to just under $7 billion by 2014, with the change unfolding gradually.
• Some strategists cited in the video see the second half of 2027 as a possible important test, when more AI computing capacity may come online and AI coding tools could pose a stronger challenge.
• For broader market perspective, the presenter notes that the Nasdaq lost almost 80% over the two and a half years after the 2000 dot-com bubble burst, despite the internet itself proving transformative.

Why software companies are exposed

The video characterizes subscription software as a highly profitable model because it combines recurring revenue, low marginal costs, and customer switching costs. AI coding tools could weaken that model by letting businesses build simpler applications themselves.

A Retool survey cited in the video found that:

• 35% of surveyed teams had already replaced at least one paid software product with an in-house tool.
• 78% planned to build more software themselves.

The presenter notes that Retool sells tools for building software, so its survey may be biased. The video also describes risks to per-seat pricing: if AI agents perform work previously done by employees, businesses may need fewer human software licenses even if the underlying software remains in use.

Risks and potential sources of resilience

• Copying an application is not the same as operating a reliable business system. The video contrasts AI-built copies of Adobe apps with the originals. A reviewer found that the Photoshop copy was much smaller (0.08 GB versus 6.4 GB) but could not open iPhone photos. The Premiere Pro copy reportedly could not perform basic clip trimming.
• The presenter argues that switching mission-critical systems in a large company is risky, and that contracts, established workflows, and integration make disruption gradual.
• The video identifies four potential company advantages, or “moats”:
  1. Proprietary data competitors cannot readily access.
  2. A large existing customer base and distribution reach.
  3. Physical infrastructure, such as chips, data centers, and power.
  4. Deep integration into how customers operate, making replacement difficult.
• AI startups that merely package third-party models may also be vulnerable if model providers add the same features themselves.

Employment and business-model signals

• The video cites Stanford research using ADP payroll records: by July 2025, employment among software developers aged 22–25 had fallen by nearly 20%. Older, more experienced workers in the same roles reportedly were not seeing the same decline. The researcher cautioned that the data did not prove AI caused every job change.
• The presenter links potential job displacement with software revenue risk: fewer employees could mean fewer paid software seats. Computer-based, repeatable work—such as coding, customer service, administration, data entry, and basic design or editing—is highlighted as more exposed.

Investor and career framework

The presenter gives a three-part framework rather than specific stock picks:

• Investments: For individual holdings, assess what prevents AI or competitors from reproducing the company’s product. Consider whether it has proprietary data, loyal customers, or hard-to-replicate infrastructure. The presenter warns that correctly predicting a technology’s impact does not necessarily mean correctly predicting the investment outcome.
• Career: Evaluate whether AI replaces or amplifies the value of your work. Learn to use AI tools and develop skills that are harder to automate, including judgment, relationships, experience, and earning trust.
• Business or personal brand: AI can copy content, designs, or products, but an established audience, reputation, and customer trust may remain scarce advantages.

Assets, companies, and instruments mentioned

• Companies and stocks: Adobe, Figma, Wix, Salesforce, GoDaddy, Intuit, ServiceNow, HubSpot, Thomson Reuters, LegalZoom, Microsoft, Anthropic, Spotify, and Tesla.
• Funds and indices: The largest software ETF (not named), the S&P 500, the Nasdaq, and a US software index.
• Other references: AI agents and AI models; chips, data centers, and power infrastructure; software subscriptions and per-seat licenses. No specific stock tickers are provided.

Disclosures and sponsorship

The presenter says the video is not financial advice and that he is not a financial adviser. The video includes a Hostinger sponsorship and promotes a discount; the presenter says he contacted the company about sponsoring that segment.

Presenters and sources

Presenter: Mark Tilbury. Kai appears in the Hostinger demonstration.

Sources and references mentioned include Retool’s survey; Brandon Thomas and his Adobe-style apps; a TechRadar reviewer; Reuters; BlackRock’s annual report for its flagship software fund; Microsoft’s CEO (not named in the subtitles); a Stanford research team and ADP payroll records; and historical figures on Napster and recorded music.