And costs are still on the rise.
Big Spender
OpenAI, the Elon Musk-founded research firm behind ChatGPT, is apparently hemorrhaging money on the game-changing chatbot that put it on the map.
People familiar with the company’s losses confirmed to The Information that OpenAI spent upwards of $540 million last year while developing its widely-used chatbot — including funds it used to poach talent from the likes of Google.
The report highlights just how expensive it is to run and maintain the popular AI tool.
And OpenAI’s costs are still on the rise, something that could turn the company into “the most capital-intensive startup in Silicon Valley history,” as CEO Sam Altman suggested earlier this week, according to the report.
Rising Costs
The Information‘s latest figure is roughly in line with Fortune‘s reporting in January that revealed the breakdown of the company’s 2022 expenses, which totaled $544.5 million: “$416.45 million on computing and data, $89.31 million on staff, and $38.75 million in unspecified other operating expenses.”
These costs would have been amassed before OpenAI struck a multi-year, multi-billion dollar deal with Microsoft at the beginning of this year.
In April, Dylan Patel, chief analyst at consulting firm SemiAnalysis, also told The Information that he estimated it costs $700,000 per day to run ChatGPT due to computing costs.
Money Talks
At the end of last year, Reuters reported that OpenAI CEO Sam Altman has set the company’s revenue bar very high in an investor pitch, with estimates that the firm could make $200 million this year and $1 billion next year.
Compared to the $30 million OpenAi made in revenue last year, according to Fortune, that figure seems almost impossibly high.
All the same, The Information also reported that Altman has privately suggested that OpenAI could raise $100 billion as it moves towards creating human-level AI or “artificial general intelligence” — an admission which, if true, says a lot about the CEO’s priorities, if nothing else.
More on OpenAI: Ex-OpenAI Safety Researcher Says There’s a 20% Chance of AI Apocalypse