Software stocks have bled out. More than $2 trillion vanished from the sector in the first quarter of 2026 alone, as fears mounted that artificial intelligence could render entire business models obsolete. The iShares Expanded Tech-Software Sector ETF plunged over 30% in six months, outpacing the Nasdaq’s 9% drop. Yet revenue keeps climbing at many firms. ServiceNow’s subscriptions grew more than 20%. Datadog’s bookings surged 37%. Adobe trades at a third of its historic price-to-earnings ratio. Investors smell value. Wall Street started bottom-fishing this week; the software ETF jumped 6.4% in two sessions, Oracle soared 18%, Microsoft and Palantir gained 6% each. Bloomberg captured the shift. But is this a dead-cat bounce, or the start of a rebound?
Take Chegg. Its stock cratered 99% since 2021. Free tools like ChatGPT gutted its core product. Revenue tumbled 40% year over year. “The stock is down by 99%… free AI tools like ChatGPT have literally replaced its core product,” said Matt Frankel on a recent Motley Fool Money podcast. Chegg’s plight stands as exhibit A for AI’s disruptive punch. A handful of skilled prompters can now whip up custom software in days, mimicking what SaaS giants spent years perfecting. High margins. Recurring revenue. Those hallmarks suddenly look vulnerable. Jon Quast nailed it: “A couple of skilled AI prompters can create software products that do what some of these SaaS companies do… in just a matter of days.”
Pay-per-seat pricing amplifies the pain. Asana’s net retention dipped below 100%. Atlassian’s enterprise seat counts fell even as revenue rose 23%. Agentic AI promises efficiency; companies might slash from 300 seats to 100 and keep output steady. “If agentic AI allows a company to function with 100-seat licenses instead of 300… it’s bad for business,” Frankel warned. HubSpot and Constellation Software face similar scrutiny. Both hit revenue highs, yet trade at rock-bottom multiples—three and four times sales, their lowest ever. Listeners like Scott Pounders, holding SaaS names, want to know: Who’s next?
Not everyone sees apocalypse. Nvidia’s Jensen Huang pushed back hard. “The market got it wrong on SaaS… AI agents won’t replace enterprise software, but the agents themselves will use the tools,” he said, name-checking ServiceNow, Cadence, and Synopsys. ServiceNow’s CEO Bill McDermott agrees. He’s rebuilding around AI, ditching pure seat reliance for hybrid models that embed intelligence directly. “There was a rerating of SaaS companies, across the industry,” McDermott told the Wall Street Journal. His firm now powers AI workflows, turning threat into tailwind.
Cybersecurity emerges as a clear winner. AI breeds novel threats—malware that evades detection, agentic attacks. The global market should triple in seven years. CrowdStrike calls AI “the largest opportunity in cybersecurity.” Zscaler eyes a $19 billion untapped niche securing agentic operations. Both stocks sit 30% to 60% off highs. Duolingo holds firm too. Sure, AI could clone language apps. But brand loyalty and network effects demand an order-of-magnitude edge to switch users. Management leans into free tiers, projecting solid growth.
Developers feel the squeeze firsthand. AI handles boilerplate, tests, docs—up to 80% of tasks at mature firms. GitHub Copilot, Claude, Devin turn one engineer into a five-person team. “No one writes code anymore,” laments a YouTube analyst. Juniors vanish first; entry-level roles evaporate as AI commoditizes syntax. Yet jobs grow. Software postings on Indeed rose 11% annually, outpacing the market. Citadel Securities data backs it. Bureau of Labor Statistics forecasts 15% expansion by 2034. Why? AI spurs more software everywhere—marketing automation, onboarding, data research. Aaron Levie of Box puts it bluntly on X: “There are far more categories where AI agents… induce demand for that skill than spaces where agents eliminate the work.”
Bugs swarm too. AI models like Mythos unearth flaws in legacy code at unprecedented rates. Patching lags; volume overwhelms small devs. “Welcome to the bug armageddon,” declares the Wall Street Journal. Cybersecurity shares dipped on the news. Deloitte’s 2026 outlook warns of intensifying rivalry: AI-natives erode incumbents, agentic tools claim 60% of the market by 2030. Agentic AI reshapes workflows, drafting full SDLC cycles for human oversight. “In 2026, agentic AI won’t just help engineers code—it’ll run first drafts,” predicts CIO.com.
Productivity soars unevenly. Pragmatic Engineer’s survey of 900 engineers shows “shippers” loving AI, while others hit usage caps and costs. Google reports 3% code quality gains, but more bugs and debt. Reddit threads echo frustration: “AI is increasing output, but also increasing tech debt.” Vibe coding—high-level prompts yielding apps—feels fast. Reality bites on scale. Large codebases confound AI; context eludes it. “AI doesn’t understand the intent,” notes an MIT study via X posts. Seniors orchestrate, juniors audit. James Bessen’s report confirms: AI grabs tasks, but new products boost demand. No job apocalypse.
Private credit exposes risks. Over $330 billion in software loans mature soon, just as AI threatens moats. Funds like those tied to ailing SaaS face losses. WSJ flags the hidden scale. Oracle cuts software jobs for AI redundancy, naming Hilary Maxson CFO amid the purge. Layoffs hit: Oracle Bengaluru axes veterans with 15+ years. X buzzes with pleas for referrals.
Survivors adapt. AI-first models win. ServiceNow thrives. CrowdStrike accelerates revenue. Engineers evolve into builders, validators, architects. Code commoditizes; judgment doesn’t. Wall Street’s bet? Bargains abound. But execution decides. AI won’t destroy software. It will remake it. Firms that harness agents endure. Laggards fade. The rout ends when revenue stories trump fear.
AI’s Assault on Software: Bargain Hunters Bet on Survivors Amid $2 Trillion Rout first appeared on Web and IT News.