Market Openai Altman
DeepSeek's Price Shock and the AI Industry's Race to the Edge
Chinese startup DeepSeek announced it is slashing API prices to one-tenth of what OpenAI, Anthropic, and Google charge for comparable language-model output, igniting what analysts are calling a full-scale AI price war. The company's new V4-Pro model reportedly achieves GPT-4-level performance at a fraction of the computational cost, though whether that reflects genuine efficiency breakthroughs or subsidized pricing designed to capture market share remains unclear. Wall Street banks including UBS, JPMorgan, and Citi characterized the ensuing selloff in rival AI stocks as 'overdone.'
OpenAI chief Sam Altman added to the weekend's turbulence with a warning that artificial general intelligence could 'end work as we know it and collapse entire economic systems' — remarks that landed alongside separate reports that OpenAI is developing its own smartphone in partnership with MediaTek and Qualcomm, with mass production reportedly targeted for 2028. Analyst Ming-Chi Kuo cited that timeline, which observers noted aligns closely with Altman's own previous predictions about AGI's arrival. Altman also called for an entirely new internet protocol built specifically for AI agents, arguing that current TCP/IP architecture was designed for human-to-human communication and cannot support agent-to-agent interaction at massive scale.
OpenAI separately published five 'guiding principles' — democratization, empowerment, universal prosperity, resilience, and adaptability — on a new company webpage, even as it pursues proprietary hardware platforms and battles Chinese price competition. The company's legal drama also intensified: jury selection began in Oakland in Elon Musk's lawsuit alleging that Altman violated OpenAI's original non-profit mission by pursuing commercial applications.
Safety failures underscored broader concerns about the speed of AI deployment. Code auditors found that Microsoft's newly shipped AI agent security toolkit contains authentication checks that, across all five language implementations, have zero production callers — meaning the security controls reportedly never actually run. Tesla, meanwhile, recalled 2.1 million vehicles after NHTSA identified autopilot sensor defects that could fail to detect obstacles, requiring over-the-air software fixes for Models 3, Y, S, and X. Senator Bernie Sanders highlighted a warning from Verizon's chief executive that AI could produce 30 percent unemployment — roughly 50 million American jobs — yet significant legislative responses have not materialized.