23/02/2026
Anthropic just wiped $285 billion off the stock market.
Claude Cowork launched industry-specific agents for legal, sales, and finance. Wall Street panicked. The enterprise software sector bled. Here are the numbers:
• Atlassian: -47% YTD (down 80% from 2021 highs)
• Salesforce: -25% YTD (45% below 52-week high)
• Workday: Price target slashed from $285 to $230
Morgan Stanley named Claude Cowork specifically as the catalyst. Not "market headwinds." A specific product from a specific company.
Here's why:
The incumbents built their empires on rigidity. Atlassian charges per seat for Jira workflows that haven't changed in a decade. Salesforce locks you into annual contracts for CRM fields you can't modify without a consultant. Workday takes 6 months to implement because their codebase predates the iPhone.
AI agents just made all of that optional.
Task boards are dead.
Not deprecated. Not "enhanced with AI." Dead.
Why would any business pay $15/seat/month for a Kanban board when an AI agent can custom-code a workflow perfectly matched to their actual process in 20 minutes? Not a template nor a plugin. Rather, a system built specifically for how your team works.
This is what we've been building at Tech Horizon Labs.
Mission Control: our open-source agent orchestration platform — doesn't bolt AI onto legacy project management. The entire way we interact with computers has changed. If it doesnt have a purpose in the age of ai we cant use it. The buttons and cards you know and love have changed forever. Why do you need a task board. Now ai fills it our for us and we dont need to see it. Your agents don't live inside someone else's task board. Rather, they run your operations directly.
OpenClaw — our open-source AI infrastructure — runs locally on your hardware or connects to the models you choose. No vendor lock-in. No data leaving your jurisdiction. No praying the API doesn't change next quarter.
We architected both from the ground up for one purpose: maximum customisation.
• Local AI for sensitive data (contracts, HR, financials)
• Cloud AI for heavy lifting (research, analysis, generation)
• Agent swarms that hand off tasks based on context
• Custom workflows that evolve as your business does
While the SaaS giants scramble to add "AI features" to their 20-year-old codebases, we built agent-native infrastructure.
The Australian SMBs we work with don't care about Atlassian's market cap. They care that their invoices get processed, their leads get followed up, and their compliance documentation gets generated without them touching it.
They want systems that fit their business. Not businesses that fit the system.
The $285B question: Are you retrofitting AI into legacy workflows? Or building operations that were only possible once AI became infrastructure?
I know where we're placing our chips.