Image

Elon Musk’s Attempts at AI Are Falling Apart

In the AI world, there are what the tech scholar Kate Crawford has called the “Great Houses of AI.” These are Microsoft, Amazon, Google, and Meta — giant tech monopolies which happen to also be four of the six top US corporations by market value.

Then there are the lesser houses, the lower fiefdoms squabbling over the crumbs that fall from the big kids’ table. This is where we find Elon Musk’s xAI. Though the world’s richest man has pumped billions of dollars into his pet AI project, it seems all the money in the world can’t buy customers — or even respect, for that matter.

According to the Wall Street Journal, new data shows the growth of xAI’s two-year old chatbot, Grok, has stagnated. Though Musk succeeded in giving Grok a temporary boost after integrating the thing into his social media site X, formerly Twitter, new monthly downloads have fallen from over 20 million in January to just 8.3 million in April.

And those are just the individual users. According to a survey of over 260,000 AI users by the research firm Recon Analytics, the number of people actively paying for Grok barely budged in the last year: it’s up from a meager 0.173 percent of X users in 2025 to a virtually unchanged 0.174 percent today.

While hooking paying customers remains a challenge for most of the AI industry, Grok stands out as a particularly terrible performer. In the same survey, more than 6 percent of respondents admitted they paid to use OpenAI’s ChatGPT, the WSJ shared.

There’s a clear reason for that lukewarm conversion rate. According to AI benchmark site LiveBench, xAI’s current model of Grok lags far behind AI models by companies like Google and OpenAI in reasoning and coding tasks — in fact, it’s even being beat by lightweight open-source models from China like Kimi and DeepSeek.

And on OpenLM’s Chatbot Arena, for another reality check, Grok currently ranks below OpenAI, Google, and numerous models of Anthropic’s large language model, Claude.

Still, diagnosing Grok’s troubles on benchmarks alone may be a fool’s errand. As we’ve seen with those aforementioned Chinese models, the most popular AI model is not always the most powerful — sometimes, it really comes down to vibes.

And as engineer and tech investor Ben Pouladian told the WSJ, the vibes around Grok are atrocious. “OpenAI is Coke, Anthropic is Pepsi and Grok is RC Cola,” he observed, “I never really saw people drinking it.”

More on Grok: Grok Just Issued a Brutal Beatdown to Elon Musk

The post Elon Musk’s Attempts at AI Are Falling Apart appeared first on Futurism.

Releated Posts

Anthropic Reports User to the Police

With mounting pressure on AI labs to curb the many real-world harms their products cause consumers, it’s never…

Oct 6, 2026 2 min read

Meta Insiders Convinced Muse Is Going to End Up Leaking the Bank Accounts and Email Archives They’re Vacuuming Up From Users

With the release of Meta’s buzzy Muse AI agent, there’ve been plenty of concerns over the amount of…

Oct 6, 2026 3 min read

Meta Is Kicking Gay Bars Off Its Apps While Allowing Videos of Graphic Child Abuse

Meta’s platforms have become an abominable cesspit of AI slop, ranging from violent child abuse to ultra-patriotic engagement…

Oct 6, 2026 3 min read

Our Solar System Is Unstable and Doomed for Destruction

Scientists believe that the massive ball of hydrogen and helium at the center of our star system was…

Oct 6, 2026 3 min read