Tech • AI • Robotics
North Korea’s Secret US Workforce, Canceling the SaaSpocalypse, MSL’s Latest Exits | Diet TBPN
7/10TL;DR
The feared “SaaS apocalypse” has not arrived on schedule, as many software companies have rebounded sharply by proving their value lies in distribution, networks, infrastructure and pricing power rather than code alone.
KEY POINTS
A market panic left room for a rebound
The broad selloff in software erased roughly $2 trillion in market value as investors bet that generative AI would quickly commoditize many products. The core concern remains that large code bases are easier to replicate than they were a decade ago, but the timetable has proved far slower and the outcomes more uneven than the market initially assumed.
Why some software businesses held up
Companies once treated as interchangeable “piles of code” often turned out to have stronger defenses. Those included embedded distribution, large sales organizations, customer trust, network effects and products that take only a tiny share of a client’s total spending, making them hard to justify replacing even if alternatives can be built more cheaply.
Shopify shows the limits of replacement theory
Shopify became a key example of software that looks vulnerable in theory but remains sticky in practice. For a business doing nearly $100 million in annual sales, Shopify Plus can cost only about $1,000 a month, a negligible expense relative to revenue. Rebuilding a comparable commerce stack in-house would still require sustained engineering effort despite advances in AI coding tools.
Platform winners still look durable
Several large software and internet platforms were cited as likely long-term winners despite AI disruption. Google, Meta, Spotify, Roblox and Salesforce benefit from scale, user bases, infrastructure and distribution. Even if AI changes how content is created, games are built or music is produced, those activities may still flow through the same dominant platforms.
Cybersecurity has emerged as a clear AI-era winner
Security spending has strengthened as AI raises both the sophistication and the volume of threats. Palo Alto Networks rose about 121% over the past year to a market value near $320 billion, while CrowdStrike gained about 107% to roughly $230 billion. Palo Alto chief executive Nikesh Arora also drew attention for investing $10 million of his own money in the company during the dip, a stake now worth about $26 million.
Twilio’s comeback highlights infrastructure moats
Twilio, up about 150% over the past year, has also challenged the idea that AI can simply replace all software layers. Its business depends on deep relationships with mobile carriers, message delivery systems and anti-spam controls. Even companies that could technically build messaging functions themselves still rely on Twilio’s network and reliability, underscoring the distinction between software tools and infrastructure.
Some companies have in fact been structurally broken
Not every software business has recovered. Chegg remains a stark loser, with its stock down 99% over five years as basic large language model services replaced much of its homework-help utility. Revenue fell to $376 million in 2025, down 39% from $617 million in 2024, leaving the company with a market value of only about $80 million to $90 million.
Canva and other design tools face similar pressure
The same replacement risk is now being discussed around design software. New image models can generate presentations, graphics and infographics in a single prompt, raising questions about whether products such as Canva can preserve pricing and engagement if core creation tasks become native features inside general AI systems.
Remote hiring has become a national security vulnerability
A separate Wall Street Journal investigation detailed how North Korean operatives used stolen identities, AI tools and U.S.-based facilitators to win remote tech jobs at American companies. The FBI says thousands of North Korean IT workers are applying across the country. One documented cell applied to more than 1,000 companies in just three months, while some workers reportedly earned as much as $300,000 a year and generated nearly $800 million for the regime in 2024.
CONCLUSION
The software shakeout is proving less like extinction and more like separation. AI is eroding weak products quickly, but companies with networks, infrastructure, low-cost positioning and entrenched distribution remain difficult to dislodge.
Explain this