Tech trillionaire Elon Musk has made a dramatic reversal in his stance on Anthropic, publicly praising the AI company as 'obviously the current leader in the field' and acknowledging he was 'clearly wrong about Anthropic'. In a reply on X, Musk wrote that he was 'clearly wrong about Anthropic' and acknowledged they are 'obviously currently the leader in AI'. The remarks mark a complete turnaround from Musk, who has been a vocal critic of Anthropic over the years, including his most sharp criticism in February after the company raised $30 billion at a $380 billion valuation. As per Mint, Musk predicted that 'No company has released a model as good as Mythos/Fable and they will undoubtedly have Mythos 2 ready soon'. This U-turn comes after years of public spats where Musk called Anthropic 'hypocritical,' 'misanthropic and evil,' and an inevitable loser in the AI race. In January, he joked that Anthropic's 'fate is to be misanthropic' and the following month accused the company of stealing training data 'at massive scale' and called it 'smug, sanctimonious and hypocritical.' The timing of this reversal is particularly significant as it coincides with Musk's own AI venture pushing deeper into the model race, with SpaceXAI launching Grok 4.5 this week.
The financial ties between the two rivals have intensified significantly, with Anthropic's valuation soaring to $1.2 trillion on secondary markets, according to Caplight data, representing a 550% surge year-over-year and eclipsing OpenAI's secondary market standing of $908 billion. The company confidentially filed for a U.S. initial public offering in June, just weeks after a Series H funding round valued it at $965 billion. In May 2026, Anthropic struck a deal to lease the full capacity of SpaceX's Colossus 1 facility in Memphis — over 220,000 Nvidia GPUs and more than 300 megawatts of power — at approximately $1.25 billion per month through May 2029, valuing the arrangement at over $40 billion. Musk emphasized this commitment despite the fact that SpaceXAI launched Grok 4.5 on July 8, a model positioned as faster and far cheaper than Anthropic's recent launches. The arrangement supports inference for Claude models and includes discussions on future orbital data centers.
The latest competitive landscape shows SpaceXAI's Grok 4.5 priced at $2 per million input tokens and $6 per million output tokens, significantly undercutting Anthropic's Claude Opus 4.8 which costs $5 for input and $25 for output. OpenAI's top-tier GPT-5.6 Sol model is priced at $5 for input and $30 for output tokens, while even OpenAI's lowest-cost Luna model matches Grok 4.5's output pricing at $6 but charges $1 for input. On benchmark tests, Grok 4.5 scored 53% on DeepSWE 1.1, trailing Claude Opus 4.8, GPT-5.5, and Fable 5, but achieved 64.7% on SWE Bench Pro, ahead of GPT-5.5. Musk acknowledged that Anthropic's Fable model remains 'definitely better than Grok 4.5' but argued most users do not require Fable-level capability for everyday tasks. He also teased that the next Grok release could bring a 'step-change improvement' as SpaceXAI applies lessons learned from engineering problems at Tesla, SpaceX, Neuralink, and the Boring Company. Independent analysis has ranked Grok 4.5 as the 'top non-Anthropic model', though three Anthropic models remain ahead at the top of the list.
In a surprising development, Musk has pledged to never cut off Anthropic's access to SpaceX compute resources in a way that hurt them badly, even as direct rivals. This commitment addresses speculation that SpaceXAI, SpaceX's AI segment, could terminate Anthropic's lease on the Colossus 1 data center in Memphis. As per Mint, Musk explained his stance by pointing to how he has treated competitors across his businesses, noting that 'Tesla open sourced its patents and we made the Supercharger network available to all competitors, even though we could have made it a walled garden.' He also said 'SpaceX competes with other satellite operators without raising prices or imposing unfair terms' and added 'Even my worst enemies can attack me on this platform.' Musk argued he does not believe in using his businesses or platform to unfairly disadvantage rivals, contrasting this approach with his previous criticism of Anthropic's alleged 'left-wing bias' and 'woke' training methods. The reversal reflects a broader industry reality where even the fiercest rivals now struggle to dismiss Anthropic's frontier position, with Musk's words suggesting that the Claude maker currently sets the benchmark that every competitor in the AI sector must chase.
The flurry of announcements underscores an AI industry in the midst of fundamental reshaping, with pricing pressure from both Chinese open-source models and aggressive moves by U.S. labs compressing margins for premium closed-source providers like Anthropic. Meanwhile, the race to embed AI agents into enterprise workflows has intensified, with companies betting that the next phase of growth lies not in better chatbots but in autonomous systems capable of executing complex, multi-step business tasks. SpaceX shares (SPCX) extended their losing streak to three straight sessions, falling nearly 1% on the day and 8% for the week as investors weighed the Grok 4.5 launch against Musk's acknowledgment of Anthropic's lead. Retail sentiment on Stocktwits remained in bearish territory, reflecting the ongoing competitive dynamics in the rapidly evolving AI landscape. SpaceX debuted on the Nasdaq in June, while Anthropic is expected to go public later this year.
