The people building the AI that's supposed to automate everyone else think it's coming for their jobs first.
The Summary
- Anthropic employees expect AI to handle 75% of their work by 2030, creating a strange dynamic where AI builders are betting on their own obsolescence
- The company is simultaneously pursuing a $2T IPO amid slowing revenue growth and offering Blackstone double-digit yields on $60B in chip financing
- FTC subpoenas over consumer safety and 401(k) concerns around the IPO show regulatory pressure mounting just as Anthropic tries to go public
- Meanwhile, Anthropic is pitching frontier AI safety legislation to lawmakers, positioning itself as the responsible actor while racing to scale
The Signal
When the engineers building Claude think AI will do three-quarters of their work in four years, that's not a product roadmap. That's a confession. Anthropic staff expect AI to handle 75% of their tasks by 2030, a projection that should make everyone else in the knowledge economy pay attention. If the people writing the code think they're next, what makes anyone else think they're safe?
The timing is telling. Anthropic is trying to thread an impossible needle: convince investors it's worth $2 trillion at IPO while revenue growth slows, convince regulators it takes safety seriously while the FTC issues subpoenas, and convince employees their jobs matter while admitting AI will do most of the work.
"Investor skepticism and market dynamics may lead to a downward adjustment of Anthropic's IPO valuation, impacting future tech IPOs."
The company's financial gymnastics reveal the stakes:
- Offering Blackstone double-digit yields on $60 billion in chip financing
- Pursuing a valuation that assumes explosive growth while current revenue disappoints
- Facing 401(k) concerns about whether retail investors should be exposed to these risks
On the regulatory front, Anthropic is playing chess while others play checkers. The company is pitching its own AI safety bill to lawmakers, a move that could "reshape AI industry norms" while conveniently positioning Anthropic as the standard-bearer. Write the rules before someone writes them for you. Smart, but it cuts against the "move fast" ethos that justifies the valuation.
The Implication
Watch what Anthropic does with staffing over the next 18 months. If employees actually believe the 75% automation number, retention becomes a different calculation. Why stay at a company betting on your obsolescence unless the equity pays off? That IPO pressure isn't just about raising capital. It's about locking in employees before they do the math.
For everyone else, the message is clear: if the AI companies think they're automating their own highly technical work by 2030, then sales, marketing, customer service, and middle management are already on borrowed time. The question isn't whether to build AI leverage into your work. It's whether you're building or being built around.