Google offered a chosen group of researchers at its AI division DeepMind stock worth millions of dollars per person in a bid to retain its brightest minds amid interest from rival OpenAI.
The offer comes even as Google lays off thousands of workers in other departments like engineering and hardware to focus on AI, The Information reported, citing inside sources.
Also read: OpenAI Lures Top Google AI Devs With $10 Million Pay Packages
AI talent war escalates
OpenAI has been targeting senior Google AI researchers, enticing them with huge pay packages worth $5 million to $10 million per year, mostly in stock. The ChatGPT creator specifically aimed to lure staff who developed Google’s large language model, Gemini.
OpenAI promised its target employees shares in the company and access to high-tech, including AI accelerator chips, to run tests. In November, when the offers were made, Google did “not appear to match” them, according to earlier reports.
Now, the Mountain View, California-based tech giant is fighting back. Select researchers at DeepMind, the division at the center of Google’s AI ambitions, “have received large grants of restricted stock worth up to millions of dollars per person,” The Information wrote.
“Leaders of the unit have been increasingly using a little-known advanced compensation program to compete with OpenAI’s eye-popping offers,” the tech publication said, quoting one person with knowledge of the matter. The report did not mention specific figures.
While OpenAI touted access to superior computing resources as a key bargaining tool, some researchers believe Google still has an advantage after making huge investments in data centers and its own micro-chips.
On the other hand, OpenAI relies on AI hardware from Microsoft, which invested $13 billion in the firm. OpenAI CEO Sam Altman admitted Google will likely continue to have an edge in computing until later this year, when it gets more capacity from Microsoft.
Tech companies are battling to dominate the artificial intelligence industry, and talent has become the latest battleground amid a shortage of skilled workers. Apparently, the talent war is two-way.
OpenAI has successfully recruited top personnel from Google, including Jiahui Yu, a key figure in the development of Gemini. Meanwhile, Matt Wiethoff, who built ChatGPT’s code interpreter feature, crossed the floor to Google from OpenAI in October.
According to data from LeadGenius and Punks & Pinstripes, OpenAI has hired around 93 people who have worked for Google and Meta before. As of February last year, about 59 former Google employees and 34 former Meta staff worked for OpenAI.
OpenAI is paying research engineers in its superalignment team between $245,000 and $450,000 per year, plus “generous equity,” per an earlier job ad.
Rivalry everywhere
Microsoft is developing an AI chip called Athena, which will compete with Nvidia’s H100 GPU and can be used as a cloud solution for developers. The company is also working with Advanced Micro Devices on its latest AI chip, MI300X.
Google and Amazon are building their own AI chips, too, all part of a strategy for self-sufficiency and to reduce dependence on Nvidia.
Google has been using and improving its own AI chip, known as the Tensor Processing Unit (TPU), since 2015. The chip has been used in several apps, including Google Street View text processing and Google Photos, as well as to train its current AI models.
ChatGPT’s success has spawned massive competition pitting big tech companies such as Google, Microsoft, and lately, Elon Musk’s xAI, all seeking to assert their dominance in the fledgling AI sector. Musk recently launched the AI chatbot Grok to compete with OpenAI’s ChatGPT, Google’s Bard, and others.
and then