Old Employees, New Dollars — Google’s $2.7 Billion Investment in Character.AI’s Reverse Acquihire for AI Innovation!

When Noam Shazeer left Google in 2021, it wasn’t on the best of terms. Frustrated by the company’s refusal to release a chatbot he had developed, Shazeer made a bold decision: he walked away from one of the world’s most powerful tech companies and co-founded Character.AI. The startup was a rebellious new venture, pushing the boundaries of conversational AI by allowing users to create custom AI companions, interact with fictional personas, and even tap into the future of artificial general intelligence (AGI). In less than two years, Character.AI skyrocketed, reaching a $1 billion valuation, with millions of users flocking to the platform each month.

But now, in an unexpected and somewhat controversial twist, Shazeer and Daniel De Freitas are heading back to Google—along with 30 top researchers from his startup. The tech giant has orchestrated a rare “reverse acquihire,” bringing the AI pioneer back into the fold without buying out Character.AI. Instead, Google opted for a non-exclusive license to use Character.AI’s advanced large language model (LLM) technology, allowing the startup to exist independently but without its cutting-edge talent. Character.AI, which debuted in 2022 amid the explosive AI chatbot craze, had ambitious goals, including the pursuit of artificial general intelligence (AGI). Shazeer, a co-author of the seminal “Attention is All You Need” paper, had high hopes for the startup, which reached a valuation of $1 billion in March 2023 after securing substantial funding.

Critics have questioned whether this move signals a retreat for Character.AI. With Shazeer and his team leaving, the company is stepping away from building its own LLMs entirely. For some, it feels like Google has swooped in, stripping the upstart of its most valuable asset—the very minds that made it a formidable player in the AI landscape. Shazeer, however, remains optimistic, expressing pride in what they built and confident that this partnership would ensure Character.AI’s future success. Still, the arrangement highlights Google’s aggressive strategy to reclaim AI talent at all costs, raising eyebrows in an industry where tech giants are increasingly reabsorbing innovators who once defied them.

Character Calls and Initiatives

Character.AI wasn’t just about chat-based interactions. The company introduced Character Calls, a feature allowing users to simulate voice conversations with AI characters. This initiative allowed for even more immersive experiences where users could have real-time “phone calls” with AI characters like Queen Elizabeth or Draco Malfoy. The feature was launched on the mobile app and quickly became popular, with over 20 million calls made by more than 3 million unique users during its initial release. The technology supported multiple languages, including English, Spanish, Russian, and Chinese, providing an inclusive experience for a global user base.

This was followed by Character Voice, a one-way communication feature that used text-to-speech (TTS) to give AI characters a voice. It was part of the company’s continuous push to refine user interaction, and by the time the “Character Calls” feature was launched, over 1 million user-created voices were available in the Character.AI library.

Cost Estimates and Development

Developing an app like Character.AI isn’t cheap. The cost estimates vary based on the complexity of the AI model, data gathering, personalization capabilities, and integration with existing systems. A basic version could cost between $40,000 and $60,000 with a development timeline of 2-3 months, while a more advanced version could exceed $300,000 and take upwards of 9 months to complete. Factors such as data security, ongoing AI improvement, and the location of the development team (with hourly rates varying from $25-$40 in Asia to $95-$100 in the U.S.) also significantly impacted the overall cost. Despite these challenges, Character.AI continued to enhance its offerings, leading to its rapid growth in users and valuation.

From Success to Controversy

While Character.AI’s rise was swift, it wasn’t without challenges. The platform enabled users to create and role-play with AI characters, but this open-ended approach also drew criticism. Controversial chatbots mimicking figures like Adolf Hitler and Saddam Hussein were briefly created, sparking concerns about content moderation. Ethical dilemmas arose, with critics questioning whether teenagers forming friendships with AI companions or using them as AI therapists could have adverse social consequences. An even more troubling incident occurred when an unauthorized chatbot used the likeness of Jennifer Ann Crecente, a murdered teenager, causing significant distress to her family. Character.AI quickly removed the chatbot, but the incident highlighted the ethical and legal pitfalls of AI.

The Reverse Acqui-Hire

In August 2024, Google announced it was bringing Shazeer, De Freitas, and around 30 key team members back into its fold. This marked one of the most high-profile reverse acqui-hires in recent years—a rare occurrence where instead of acquiring the entire company, the focus is on talent. As part of the agreement, Google secured a non-exclusive license for Character.AI’s technology, while Shazeer returned to lead AI efforts at DeepMind, particularly the highly anticipated Gemini initiative. Interestingly, Dominic Perella, Character.AI’s general counsel, stepped in as interim CEO, signaling that the company would remain independent despite the return of its leadership to Google. “We’re thrilled to continue building Character.AI as an independent company,” Perella remarked, emphasizing that while the founders had returned to Google, the startup still had its original vision to pursue.

Growing Concerns and Ethical Dilemmas

Character.AI’s story took a darker turn when it was revealed that the likeness of Jennifer Ann Crecente, a murdered teenager, was used to create an unauthorized chatbot. This distressing event highlighted the ethical and legal challenges facing companies like Character.AI. Her father, Drew Crecente, discovered the chatbot through a Google Alert, causing him significant emotional distress. Character.AI quickly removed the chatbot, acknowledging it violated their policies against impersonation, but the damage had been done. This incident underscored the need for better regulation and ethical guidelines in AI development.

“What if you could create your own AI, and it was always available to help you with anything?” Shazeer mused in a 2023 interview with Forbes, showcasing his ambitious vision for AI companions. Yet, incidents like these show the ethical pitfalls of unrestricted AI creation, raising questions about how to responsibly harness the power of such technology.

A New Chapter for Google and Character.AI

With Shazeer now back at Google, the company is poised to leverage Character.AI’s innovations as it embarks on its next big AI push. Shazeer will play a leading role in Google’s Gemini initiative, an advanced AI program expected to build on the breakthroughs of past models while incorporating even more sophisticated conversational abilities. This move underscores Google’s strategy of combining external talent with internal AI resources to stay ahead in the AI arms race. The deal also highlights the lengths companies are willing to go to secure top AI talent in a fiercely competitive landscape—illustrating a growing trend of reverse acqui-hires where talent, not products, is the primary commodity.

In the broader context, the reverse acqui-hire of Character.AI by Google raises questions about the sustainability of talent poaching and whether the market might be overspending in the AI race. Yet for Google, bringing back Shazeer and his team could be a game-changing move as the tech giant doubles down on AI innovation.

Final Thoughts: Talent Over Acquisition

The reverse acquihire of Character.AI signals a shift in tech acquisitions, where talent is often prioritized over the products themselves. While Google has regained Noam Shazeer and key team members to drive its Gemini initiative, Character.AI will continue to operate independently under interim CEO Dominic Perella. The company remains committed to refining its AI-driven conversational experiences, such as Character Calls, while stepping back from developing new large language models. As Shazeer helps shape Google’s Gemini initiative, it’s clear that both companies have much to gain from this symbiotic relationship, even if the journey has been anything but typical.