In a significant shift for the AI industry leader, OpenAI's first open-source model in years will take longer than expected to reach the public. CEO Sam Altman announced on June 10 that the release would be postponed beyond June, explaining that the research team had achieved "something unexpected and quite amazing" that would be "very worth the wait."
The delay comes at a pivotal moment for OpenAI, which earlier this year acknowledged being on the "wrong side of history" regarding open-source AI development. Since transitioning from its non-profit origins to a more closed, proprietary approach in 2019, the company has faced growing competitive pressure from open-source alternatives.
Chinese AI startup DeepSeek has emerged as a particularly disruptive force, with its R1 model demonstrating comparable capabilities to OpenAI's offerings at a fraction of the development cost. By January 2025, DeepSeek had overtaken ChatGPT as the most downloaded free app on Apple's App Store in the US, sending shockwaves through Silicon Valley and prompting a strategic reassessment across the industry.
Meta has also made significant inroads with its Llama family of open-source models, which surpassed one billion downloads in March 2025. CEO Mark Zuckerberg has positioned Meta as "the standard-bearer for open-source AI" with plans to invest over $60 billion in AI development this year alone.
For enterprise customers, the appeal of open-source models extends beyond cost considerations. The ability to run models locally addresses persistent concerns around data sovereignty, vendor lock-in, and regulatory compliance, particularly in sectors like healthcare, finance, and government where data privacy requirements have limited cloud-based AI adoption.
As AI technology continues its rapid evolution, the conversation has expanded beyond technical capabilities to include questions of responsible development, accessibility, and sustainability. With OpenAI reportedly spending $7-8 billion annually on operations while projecting losses of $5 billion this year, the economic viability of different AI development approaches has become a central industry concern.