Google is making a calculated bet on the future of voice—and it’s doing it by hiring people, not companies.
In a low-key licensing agreement, Google DeepMind has brought on Alan Cowen, the CEO of Hume AI, along with roughly seven senior engineers, according to people familiar with the deal. Financial terms were not disclosed.
The move underscores a growing belief inside top AI labs: voice is becoming the most important interface for artificial intelligence—and emotional awareness is the missing layer.
Why Google wants Hume’s talent
Hume AI has built its reputation around emotionally intelligent voice systems. Its models are trained to recognize tone, mood, and subtle emotional cues in real speech, using conversations annotated by human experts. That work is costly, slow, and difficult to replicate at scale.
By bringing Cowen and his team into DeepMind, Google gains years of specialized research without acquiring the company outright. At Google, the team will help integrate voice and emotion-sensing capabilities into next-generation AI models, sources say.
Cowen’s background is unusually well-suited to the task. He holds a PhD in psychology and has positioned Hume AI at the intersection of behavioral science and machine learning—an area Big Tech increasingly sees as strategic.
Hume AI stays independent
Despite the talent shift, Hume AI isn’t shutting down. The company says it will continue supplying its technology to multiple frontier AI labs, not just Google.
Hume AI expects to generate about $100 million in revenue in 2026 and has raised $74 million in funding so far, according to investors. Andrew Ettinger, an investor and longtime executive, is stepping in as CEO and says new voice models will launch in the coming months.
“Voice is going to become a primary interface for AI,” Ettinger said in an interview. “That’s clearly where this is headed.”
The competitive backdrop
Google’s move lands in the middle of an intensifying voice AI arms race. OpenAI already offers a highly realistic voice mode in ChatGPT, while Amazon, Apple, and Meta are all pushing deeper into conversational AI.
Google is also partnering closely with Apple, with Google Gemini expected to power parts of a future version of Siri. Emotional intelligence could become a key differentiator in how natural those assistants feel to users.
A familiar Silicon Valley playbook
The deal follows a pattern that’s becoming more common: licensing agreements that function like acquisitions, minus the regulatory friction.
In recent years, Google, Microsoft, Amazon, and Meta have all struck arrangements that allow them to absorb high-value AI talent without formally buying companies. Regulators are taking note. The Federal Trade Commission has said it plans to scrutinize these “aqui-hire” style deals more closely.
Still, for now, they remain an efficient way for Big Tech to move fast in an increasingly crowded AI market.
Why it matters
For users, emotionally aware voice AI could mean assistants that respond with more nuance—detecting frustration, urgency, or confusion and adjusting accordingly. For businesses, especially customer support, it could translate into smoother interactions and better outcomes.
For Google, the message is clear: the next phase of AI won’t just be smarter. It will sound more human—and know when you’re annoyed.
Google isn’t waiting for the voice revolution to arrive. It’s hiring the people who’ve already been building it.