Adept, which is building an AI model that can turn text commands into computer actions, raised $350M in a Series B, sources say at a $1B+ post-money valuation
Context & Ripple Effects
Adept came out of stealth less than a year ago with a $65M round led by Greylock and Addition built around one thesis: an AI model that doesn't just answer questions but executes any software process a user describes in text. The new $350M Series B at a reported $1B+ post-money valuation is investors pricing that thesis as a category rather than a research bet.
The raise lands amid a broader funding surge for founder-led AI labs and agents — Inflection AI's $1.3B round followed months later, and agent-focused startups like Cognition's Devin and 11x.ai have since drawn comparable or larger valuations — making Adept an early marker of how quickly capital consolidated around 'AI that acts' after the chatbot wave.
First-order effects
- Adept gains the balance sheet to fund the compute-heavy training its text-to-action models require, extending runway well beyond what the $65M stealth round supported.
- Greylock and Addition's early stakes are now marked against a $1B+ post-money valuation, roughly an order of magnitude above the seed, giving the firm a flagship agentic-AI position.
Second-order effects
- Competing labs with marquee-founder pedigrees — Inflection under Mustafa Suleyman most visibly — face a funding benchmark set at nine figures per round, pressuring every frontier-adjacent startup to scale raises or concede the field.
- Application-layer agent builders such as 11x.ai, which sells AI sales reps, now have both a potential foundation-model supplier in Adept and a signal that the platform layer beneath them is being capitalized fast enough to become a dependency risk.
Third-order effects
- If text-to-action agents work as described, software consumption shifts from users clicking through applications to models operating them, positioning whoever owns the action layer — not the app vendors — as the new intermediary across enterprise workflows.
- That same layer automates routine screen-based knowledge work directly, feeding the documented pressure on formulaic and entry-level professional tasks rather than displacing whole job categories at once.
The trend: Venture capital is rotating from conversational AI toward agentic systems that execute software tasks on a user's behalf, with Adept's Series B an early, outsized data point in that rotation.