Huang Tells Wall Street Software Selloff Misses AI’s True Value Equation
Nvidia CEO argues markets fundamentally misread how AI agents will consume enterprise software, challenging the narrative behind a $2 trillion sector rout.
Nvidia CEO Jensen Huang declared the software sector selloff a market misjudgment hours after reporting fiscal Q4 revenue of $68.13 billion, arguing that agentic AI will amplify rather than replace enterprise software demand.
Speaking to CNBC on 26 February following Nvidia’s earnings beat, Huang rejected fears that AI agents will cannibalize the Enterprise Software industry. The S&P 500 software and services index has shed 23% since the start of 2026, with JPMorgan analysts estimating approximately $2 trillion in market capitalization erased amid concerns that large language models could render traditional SaaS offerings obsolete.
The Tool Use Argument
Huang’s central thesis contradicts the prevailing investor narrative. According to CNBC, he argued AI agents will function as tool users rather than tool replacers, stating that “Agentic AI will be intelligent software that uses these tools on our behalf.” He cited ServiceNow, SAP, and Salesforce as platforms that agents will leverage rather than displace, noting that companies will augment biological employees with “hundreds of thousands of digital employees,” each requiring software licenses.
The perspective carries weight given Nvidia’s position at the infrastructure layer. Hyperscalers—comprising Alphabet, Amazon, Meta, Microsoft, and Oracle—accounted for over 50% of Nvidia’s $62.3 billion data center revenue in Q4, according to CNBC. Combined capital expenditure from these customers is projected to approach $700 billion in 2026, a 60% increase year-over-year per Morningstar data.
“I think the markets got it wrong. These agentic AI will be intelligent software that uses these tools on our behalf and help us be more productive.”
— Jensen Huang, Nvidia CEO
Market Reaction and Pushback
Not all analysts accept Huang’s framing. Dan Niles of Niles Investment Management told CNBC that “there’s some real companies that are going to go to zero in the software space,” predicting database and cybersecurity firms will prove most resilient. TheStreet reported forward earnings multiples for the software sector have collapsed from 39 times a year ago to 21 times currently.
The selloff accelerated following Anthropic’s release of legal contract review plugins for Claude AI in early February, which investors interpreted as proof of concept for AI-driven software displacement. Bloomberg characterized the rout as driven by “fear of AI displacement,” with SaaS companies bearing the brunt.
| Metric | Result | vs. Estimate |
|---|---|---|
| Revenue | $68.13B | +$1.92B |
| Data Center Revenue | $62.3B | +75% YoY |
| Q1 Guidance | $78B (±2%) | +$5.4B |
| EPS (adjusted) | $1.62 | Beat |
Enterprise AI Deployment Reality
Enterprise adoption data presents a nuanced picture. TechRepublic reports 78% of organizations now use AI in at least one function, but only 8.6% have deployed AI agents in production environments. Deloitte survey data indicates 74% of companies report advanced AI initiatives met or exceeded ROI targets, with approximately 20% seeing returns above 30%.
Yet scaling remains problematic. Lucidworks research shows only 6% of organizations have fully implemented agentic AI, with most enterprises stuck between exploratory pilots and production deployment. According to ETR, enterprises are reducing AI license counts while increasing total spending—a shift toward “targeted access” and “higher-value capabilities” rather than broad-based rollout.
The software selloff dubbed the “SaaSpocalypse” began in early February after a wave of disappointing earnings and new AI capabilities from Anthropic. ServiceNow shares have tumbled 28% year-to-date, Salesforce is down 26%, and Intuit has lost 34%. The iShares Expanded Tech-Software ETF (IGV) has declined 31% from its high, with valuations compressing to levels not seen since the pandemic.
The Visibility Advantage
Huang’s confidence stems from direct customer engagement. Nvidia’s Sovereign AI business—government and national AI infrastructure projects—more than tripled year-over-year to exceed $30 billion, driven by deployments in Canada, France, the Netherlands, Singapore, and the UK, according to Nvidia CFO commentary. CFO Colette Kress stated the company has “visibility” to $500 billion in Blackwell and Rubin revenue from calendar 2025 through 2026.
Analysts at HSBC issued a note arguing “software is already eating AI,” predicting “the lion’s share of value” will be generated within the software sector as agentic AI kicks off in 2026. Wedbush analysts characterized the selloff as pricing in “disruption that, on current evidence, is not coming in the form the market fears.”
What to Watch
The validity of Huang’s thesis will be tested in coming quarters as enterprises move from AI pilots to production deployments. Key indicators include software companies’ ability to transition to outcome-based pricing models rather than per-seat licensing, the velocity of agentic AI adoption beyond the current 8.6% production deployment rate, and whether enterprise AI spending consolidates around incumbent platforms or fragments across new entrants.
Nvidia’s Q1 guidance of $78 billion—well above the $72.6 billion analyst consensus—signals continued infrastructure buildout. Whether that translates to software consumption at scale or reflects infrastructure overbuilding ahead of demand remains the trillion-dollar question. Goldman Sachs strategist Ben Snider warned this may represent “the end of the beginning” of software declines, drawing parallels to industries like newspapers where disruption caused prolonged valuation compression before stabilization.
For now, the market faces a binary choice: believe the CEO whose customers are spending $700 billion annually on AI infrastructure, or trust the valuation reset that has wiped $2 trillion from software market caps. The resolution will determine whether 2026 marks a generational buying opportunity or the start of a structural repricing.