India Launches Global AI Summit as Software Stocks Reel from AI Disruption Fears

Software stocks continued their steep descent on February 16, with the iShares Expanded Tech-Software Sector ETF (IGV) down more than 20% for 2026 and falling over 8% in February alone. Citi analysts stepped in, arguing the sell-off has gone too far and flagging buying opportunities across many names — even as CNN and other outlets warned the so-called "AI scare trade" may not be finished.

The rout intensified earlier in the month after Anthropic released new agentic tools for legal, finance, and marketing workflows. ServiceNow tumbled nearly 7% (down 28% year-to-date), Salesforce dropped about 7% (down 26% YTD), and Intuit fell nearly 11% (down more than 34% YTD). A wave of AI-powered startup launches in insurance, freight, and wealth management amplified fears that SaaS incumbents face structural disruption.

Wall Street is divided. JPMorgan Chase sees potential for a rebound, pointing to solid fundamentals and what it calls an "overly bearish" outlook on disruption. Goldman Sachs and Wedbush named Microsoft — the worst-performing Magnificent Seven stock in 2026 — as a top buy. For investors, the key question is whether AI will complement or cannibalize the software businesses that have powered market returns for the past decade.

February 2026 is shaping up to be the most competitive single month in AI model history, with seven major releases from the world's leading AI labs all expected to land simultaneously: Google DeepMind's Gemini 3 Pro GA, Anthropic's Sonnet 5, OpenAI's GPT-5.3, Alibaba's Qwen 3.5, Zhipu AI's GLM 5, DeepSeek v4, and xAI's Grok 4.20. The convergence is being called a "Model Rush" by industry analysts.

The timing is no accident. February sits between CES and MWC, making it a prime window for companies to establish market position before spring. What makes this rush particularly significant is that open-source contenders (Qwen 3.5, DeepSeek v4) are clashing head-on with closed commercial models — a dynamic that reflects how dramatically open-source performance has improved since mid-2025. Three of the seven models come from Chinese AI companies, underscoring how the US-China AI competition has moved firmly to the model layer.

For developers and enterprises, the simultaneous releases are expected to accelerate API price compression and expand the range of cost-effective deployment options. Analysts note that each lab is under pressure to differentiate — whether through reasoning improvements, coding capabilities, multimodal features, or pricing — rather than simply releasing a larger model than the competition.

A wave of state-level AI legislation advanced across the US last week, with chatbot regulation emerging as the dominant theme of 2026's legislative session. Bills in Virginia and Washington passed their respective state senates, while similar measures cleared committees in Utah, Arizona, and Hawaii. Six additional states introduced new chatbot-related bills, according to Troutman Pepper's weekly legislative tracker.

Washington's HB 1170 — a disclosure and provenance bill modeled on California's AI Transparency Act — passed the full House on February 13. It would require AI content providers to offer users a provenance detection tool and include both manifest and latent disclosures in AI-generated content. Tennessee's Senate also passed a bill barring AI systems from advertising themselves as qualified mental health professionals.

The activity comes amid ongoing federal tension: President Trump signed an executive order in December 2025 that challenges the enforceability of state AI laws and signals a preference for a unified federal framework. Colorado's landmark AI Act, the first comprehensive state AI law in the US, had its effective date pushed back to June 30, 2026. With dozens of bills moving simultaneously across nearly two dozen states, legal experts say businesses face mounting compliance complexity regardless of how the federal preemption question resolves.