OpenAI and Anthropic Slash Prices as Enterprise Quantum Spending Hits $300M
OpenAI and Anthropic have entered a price war, releasing cheaper models as Chinese AI rivals gain ground on their trillion-dollar ambitions, according to the Financial Times. The US groups are cutting API costs to retain market share against challengers offering competitive performance at a fraction of the price.
Enterprise users spent a total of $300 million on quantum computing in 2025, outranking the combined spending of research labs and governments for the first time, according to BCG data reported by the Wall Street Journal. Many companies are pouring millions into quantum programs to gain a competitive edge and prepare for future encryption threats.
Google will now allow users to remove the visible watermark from its AI generations, though turning off this setting will not affect the invisible benchmarks used to identify an AI-generated file, per The Verge. The change gives creators more flexibility while preserving Google's internal tracking across AI outputs from Gemini and other models.
French startup Kog is going deeper to squeeze more inference out of GPUs, arguing that the idea GPUs are poorly suited for agentic workflows may be a misconception, according to TechCrunch. The company is optimizing GPU utilization to push more agentic performance from existing hardware rather than relying on custom silicon.
Meta CEO Mark Zuckerberg's Instagram rolled out a new wordmark this week that has left users baffled, as it no longer clearly spells "Instagram" and appears to read "Instagzam," reports The Verge. Instagram reportedly felt the previous 13-year-old logo looked dated, but the redesign has sparked widespread confusion across its 2 billion monthly active users.
A pro se litigant suspecting the court of using AI injected hidden prompts in legal filings to try to win his case, according to Reuters. The judge warned that self-represented litigants are increasingly using chatbots incorrectly and getting desperate, raising concerns about AI manipulation in courtrooms handling thousands of filings annually.
Sources: