OpenAI Hires Former Meta AR Executive to Lead Hardware

TMTPOST -- OpenAI, the artificial intelligence (AI) star behind ChatGPT, is foraying into the hardware business.

Credit：Meta

Caitlin "CK" Kalinowski joined in OpenAI to lead robotics and consumer hardware, the former head of Meta Platforms' augmented reality (AR) glasses announced Monday. "In my new role, I will initially focus on OpenAI's robotics work and partnerships to help bring AI into the physical world and unlock its benefits for humanity," Kalinowski said in a LinkedIn post.

As a veteran of virtual reality (VR) and AR industry, Kalinowski used to lead Oculus VR’s hardware for more than nine years before the startup was acquired by Meta in 2014 for $2 billion. She served as head of AR glasses hardware in Meta since March 2022 and oversaw the creation of Orion, Meta’s most advanced AR glasses to the date. With 70 degree field-of view, the glasses run all of the hand tracking, eye tracking, simultaneous localization and mapping (SLAM), and specialized AR world locking graphics algorithms while the app logic runs on the puck to keep the glasses as lightweight and compact as possible.

Kalinowski could work with her old boss Jony Ive, the legendary iPhone and iMac designer who left his full-time role at Apple five years ago, on a new AI hardware. CNBC reported in September that Ive partnered with OpenAI to develop and AI-powered device. Ive confirmed that month he was building a hardware product with OpenAI, describing it as “a product that uses AI to create a computing experience that is less socially disruptive than the iPhone.”

Besides the latest recruitment, OpenAI has revived its robotics division earlier this year after disbanding the relevant research team in 2020. It was said in June that new job listings showed OpenAI is looking for research engineers to join a robotics team to train multimodal large language models (LLMs) to “unlock new capabilities for our partners' robots.”

Earlier this week, OpenAI’s further bets on robotics was confirmed. Physical Intelligence, a robot startup based in San Francisco, said Monday it has raised $400 million at a $2.4 billion post-money valuation. Investos participated in the funding included Amazon founder Jeff Bezos, OpenAI, Thrive Capital, Lux Capital and Bond Capital.

Expansion into robotics and hardware suggested OpenAI’s efforts to transfer from the present nonprofit form to a for-profit model in face of huge losses these years.

OpenAI will remain unprofitable until 2029 and its could incur losses up to $14 billion in 2016, nearly tripling this year’s expected loss, according to an analysis of data from OpenAI financial reports seen by The Information in Ocotober. It is reported that OpenAI expected to make an annual profit of $14 billion in 2029 after a total of $44 billion losses between 2023 and 2028. The aforementioned losses exclude stock compensation, which is one of OpenAI’s biggest expenses, according to the report. It said the company recorded stock compensation of $15 billion in the first half of year, likely equivalent to the revenue in the same period.

The report highlights how costly the AI development OpenAI delivered with massive investment and operational costs. Based on the reported financial reports, OpenAI projected it will spend more than $200 billion through the end of the decade, excluding stock compensation. Around 60% to 80% of its expected annual spending will go toward either training or running the AI models.

The financial reports also showed how fast OpenAI has been burning money. The report commented OpenAI cash burn pace seems far less than previsously thought as it burned through about $340 billion in from January to June this year, leaving it a cash of $1 billion on the balance sheet before its latest funding round. However, the reports suggested the cash burn could accelerate significantly in the next couple of years. OpenAI calculated net losses of $3 billion for the first half of the year, while its steep net losses mainly result from expenses like stock compensation and computing costs, which don’t flow through its cash statement.