IT Sector Update : Making sense of the Anthropic product launches by Motilal Oswal Financial Services Ltd
A slew of product releases, mostly from Claude, have unsettled markets as AI moves deeper into workflows
February has seen a series of product releases (mostly by Anthropic), threatening to automate not just software engineering but also workflows in myriad streams such as legal, financial data sets, cybersecurity, insurance broking, wealth management, and legacy modernization. The reaction from the markets (and to a few doomsday blog posts) to this has been quite sharp (see our report dated 13th Feb’26: Indian IT services: Assessing the narrative shock); as shown in stocks from most of these verticals have seen deep drawdowns in the month of February (IT services being one of the biggest losers, registering its worst MoM dip since the dot-com era – refer to Exhibit 2). We take stock of these product announcements in this report and also present the erosion of value in key industries over the past one month. Our analysis of API calls for Claude and token usage for OpenAI reveals two key things: 1) software engineering is ground zero for AI invasion – 50% of all API calls target software engineering, and 2) AI is currently being used only by cloud-first/AInative enterprises. Of the top 20 token users for OpenAI, 90% are new-age companies. This indicates that AI is easier to deploy in greenfield environments, and still difficult to deploy at scale for enterprises with legacy burdens. Of course, AI models keep improving exponentially.
What did Claude launch in Feb’26 ?
* In Feb, Anthropic announced a series of product upgrades (Exhibit 1) across legal workflows, coding, cybersecurity, and legacy modernization within a span of three weeks. Each release targeted a specific billable knowledge-work layer.
* During 30th Jan- 3rd Feb: Anthropic launched Cowork with 11 plugins focused on routine legal drafting, compliance checks, and review workflows. LegalZoom fell ~29%, Thomson Reuters ~27%, and Wolters Kluwer ~25% as markets read this as AI entering recurring legal billing pools. The pressure spilled into SaaS - Workday (~25%) and ServiceNow (~22%) declined amid concerns that AI could begin to absorb SaaS-led workflow within enterprise software layers.
* During 5 th -17th Feb’26, Anthropic released Opus 4.6 and Sonnet 4.6, improving reasoning, coding capability, and long-context execution. Sonnet became the faster, lower-cost default model. The signal was that Claude is positioning itself as an enterprise-grade execution layer rather than a consumer-grade chatbot. Software stocks corrected broadly during this period, and the iShares Expanded Tech-Software Sector ETF saw one of its sharpest drawdowns since 2008.
* On 20th Feb, Claude Code Security demonstrated the ability to scan full codebases, detect vulnerabilities, and suggest fixes. CrowdStrike and SailPoint declined ~15-18%. The concern was that AI-led vulnerability detection may pressure parts of traditional security tooling.
* On 23rd Feb, Claude showcased COBOL modernization capability, including system analysis and migration assistance for legacy stacks. IBM has fallen ~23% till date. This expanded the debate from coding productivity to consulting-led legacy modernization revenue.
Brownfield and greenfield AI-implementation
* Our analysis of API calls for Claude and token usage for OpenAI reveals two trends. First, software engineering is ground zero for AI adoption – about 50% of Claude API calls relate to coding and development tasks. Second, OpenAI token usage remains concentrated among cloud-first and AI-native companies. Among the top 20 token users for OpenAI, ~90% are new-age firms.
* This tells us that AI deployment today is easier in Greenfield environments. Start-ups can redesign workflows around AI from day one. There are fewer legacy systems, fewer approval layers, and faster iteration cycles.
* Large enterprises operate differently. Most run on systems built over 20–30 years, with applications that are layered, integrated, and customized. In such brownfield environments, deploying AI at scale requires integration with legacy stacks, data cleanup, and governance alignment.
* About 60–80% of enterprise IT budgets still go toward maintenance. This means AI-led productivity gains often depend on prior modernization. Without addressing legacy complexity, scaling AI beyond pilot use cases becomes difficult.
* While models continue to improve, scaled enterprise-wide usage outside software engineering is still evolving. As depicted in Exhibit 3, adoption across other functions such as sales, finance, BI, customer service, and e-commerce remains in low single digits.
* Overall, we believe the AI model is advancing quickly, but deployment at scale remains uneven. AI works well in Greenfield settings today; broad-based brownfield transformation will take longer.
For More Research Reports : Click Here
For More Motilal Oswal Securities Ltd Disclaimer
http://www.motilaloswal.com/MOSLdisclaimer/disclaimer.html
SEBI Registration number is INH000000412
