OpenAI generated $6.7 billion in revenue during the second quarter, up 18% from the prior three months, while its operating loss deepened as the ChatGPT maker faced growing competition from Anthropic, according to The Wall Street Journal.
OpenAI's revenue increased from $5.7 billion in the first quarter, people familiar with the matter told the WSJ. However, its operating loss, including stock-based compensation, widened from $9.3 billion to $12.3 billion over the same period, outpacing sales.
Anthropic, meanwhile, more than doubled its quarterly revenue to $11.6 billion and overtook OpenAI's sales for the first time, per the report. The Claude developer also posted a small adjusted operating profit, though it is unclear how Anthropic calculated that figure, the outlet said. The company has previously excluded stock-based compensation from adjusted results.
OpenAI says growth has accelerated in third quarter
However, OpenAI has told investors that growth picked up following the release of new models in July, the people said. The company's annualized revenue run rate increased 20% month over month in July to $40 billion, while revenue from business customers grew 32%, according to slides viewed by CNBC.
The results come amid broader changes at OpenAI as it looks to reaccelerate growth. The company recently released a "super app" combining its Codex coding tool with ChatGPT and a web browser, which it says is attracting users quickly. Co-founder and President Greg Brockman has recently taken a more active role across its product and business divisions. Chief Revenue Officer Denise Dresser was replaced last week after less than a year in the role, following other senior departures including former Chief Operating Officer Brad Lightcap and Fidji Simo. OpenAI also subsidizes hundreds of millions of non-paying ChatGPT users and recently cut prices for two of its latest models as some corporate customers became more cautious about AI spending and shifted certain workloads to cheaper Chinese systems, the WSJ said.
OpenAI is simultaneously slowing parts of its model development as it aims to strengthen security controls. The company said Tuesday it has paused the pace of model development after an AI agent under testing escaped the sandbox and hacked Hugging Face. OpenAI said its upcoming model Astra may meet the "Critical cybersecurity capability threshold" under its Preparedness Framework, the AI firm's major security document.
Despite the slowdown, OpenAI continues to expand its computing footprint. The company said Monday that it had agreed to secure roughly 8 gigawatts of IT capacity at the PORTS-Pike Technology Campus in Pike County, Ohio.