Everything Elon Musk posted on X today about Grok, Starlink, and more

Elon Musk spent another busy day on X, weighing in on everything from self-driving cars and Starship launches to World Cup predictions and the future of AI. Buried in the memes and politics were some important updates about Tesla autonomy, Starlink’s growth, and xAI’s Grok models.

Trial lawyers, self-driving cars, and Tesla’s robotaxi push

Musk reacted strongly to a post claiming that U.S. trial lawyers are lobbying against self-driving cars because they are “too safe” and would reduce the number of car accident lawsuits. The post cited estimates that U.S. auto insurance payouts total roughly $180–220 billion per year, a huge pool of money that supports a large personal injury industry.

When a commenter highlighted the $220 billion figure and suggested it explains why some groups oppose autonomous vehicles, Musk replied with a simple “Wow,” signaling his view that financial incentives are a major barrier to adopting safer self-driving systems.

On the product side, Musk amplified a clip of a test track filled with Tesla “cyber cabs” — a fleet of Cybertruck-based robotaxis. While there were no new specs, the footage reinforces Tesla’s push toward a dedicated robotaxi service built on its Full Self-Driving (FSD) technology.

Starship, Dragon, and Starlink’s insane growth

On the SpaceX front, Musk shared a throwback video from 2014 showing the interior of the Dragon Version 2 spacecraft. In the clip, he explains how the cockpit was designed to be clean and simple, with touchscreens for most functions and physical buttons reserved for critical emergency controls. That design later helped restore America’s ability to launch astronauts from U.S. soil.

Looking forward, Musk confirmed that the next Starship flight test, the 13th overall, is now targeting a launch attempt on Friday after a recent delay. SpaceX is continuing its rapid test cadence as it iterates on Starship’s design.

He also highlighted just how quickly Starlink has evolved. A new generation of Starlink satellites now carries about 1,024 GB of bandwidth each, compared with 96 GB in the previous generation — roughly a tenfold jump in a single design cycle. Commentators framed this as the kind of leap you only see in industries that are still aggressively innovating.

Another post laid out Starlink’s growth timeline:

• In 2019, the first dedicated Starlink launch carried 60 satellites.
• In 2020, Starlink entered wider public beta.
• In 2024, SpaceX launched its first direct-to-cell satellites and began commercial satellite texting.
• By 2025, direct-to-cell expanded across more carriers, devices, and markets.
• By March 31, 2026, SpaceX had about 9,600 Starlink satellites in low Earth orbit and roughly 10.3 million subscribers across 164 countries and territories.

Direct-to-cell services were already providing satellite texting and voice to about 7.4 million monthly unique devices in around 30 countries. The post described Starlink as a new global communications layer connecting homes, businesses, aircraft, ships, governments, rescue missions, and ordinary smartphones. Musk responded by praising the “amazing” SpaceX team.

Grok predicts the World Cup and doubles down on forecasting

The 2026 FIFA World Cup final between Spain and Argentina dominated global attention, and Musk live-posted throughout the match on X, sharing stats like ball possession, shots on target, and more. After a long scoreless stretch, Spain finally scored, and Musk congratulated the new champions.

What really caught his attention, though, was a recap of how xAI’s Grok model had predicted Spain as the likely winner months before the tournament began. According to the post, Grok combined betting odds, prediction markets, and supercomputer forecasts back in April and concluded that Spain had the highest probability of winning. When Spain ultimately lifted the trophy, Musk highlighted this as an example of Grok’s strength in prediction and forecasting, commenting that “Grock is good at predictions.”

This fits into Musk’s broader vision of AI not just as a chatbot, but as a tool for probabilistic reasoning across noisy, real-world signals — a theme that also runs through xAI’s broader strategy, explored in more depth in our look at Musk’s role in the global AI power struggle.

Grok 4.5 and the rise of Grok Build

Musk also amplified several updates about Grok 4.5, xAI’s latest flagship model, and the growing Grok Build ecosystem.

Grok 4.5 tops a new coding benchmark

On the Vulcan Bench coding benchmark, Grok 4.5 reportedly achieved a 91.3% score, solving 21 out of 23 real-world software tasks across five programming languages. The post claimed that Grok 4.5 outperformed competing models like Claude “Fable 5” and GPT “5.6 Soul,” while also being more cost-efficient.

While independent verification and broader benchmarks still matter, Musk’s decision to highlight these results shows how xAI is positioning Grok as a serious coding assistant and developer tool — not just a conversational model. For more context on how Grok has evolved to this point, see our breakdown of Grok 4.3 and xAI’s model roadmap.

Grok Build: from idea to deployed app in one place

Musk also promoted “Grok Build,” described as a complete environment for building, deploying, and managing real-world applications around the Grok model. Instead of just generating code, Grok Build aims to integrate the entire development stack into a single workflow.

The platform’s plugin marketplace connects Grok with popular tools across the software lifecycle:

• Development and monitoring: Replit-style environments, Sentry for monitoring, Chrome DevTools for debugging.
• Infrastructure and hosting: Cloudflare for edge and networking, Railway for hosting.
• Data and storage: Supabase, MongoDB, and Neon for databases; Firecrawl for web data.
• Design and payments: Figma for UI/UX design and Stripe for payment processing.

The idea is that you can ask Grok to help you design, build, test, debug, connect infrastructure, deploy, and manage production apps without constantly switching between separate tools. A new “summarize” command for instant session summaries and Grok 4.5 as the default model (with high, medium, and low reasoning modes) are part of this push.

Musk encouraged users to “try Grok Build,” signaling that xAI wants developers to see Grok not just as an AI assistant, but as the backbone of a full-stack, agentic development environment.

AI in law: Kimi K3 vs. Claude

Another AI-related post Musk reacted to compared performance on hard, autonomous legal tasks. On a benchmark from legal AI company Harvey, the model “Kimi K3” reportedly scored 26.7%, about double the 14.2% reported for Claude “Fable.” A commenter joked that Claude must have been “distilled” from Kimi, and Musk responded with amusement.

While the numbers themselves are modest, the key takeaway is that specialized benchmarks are emerging for complex, domain-specific work like law. Models are being evaluated not just on generic reasoning, but on their ability to handle real legal workflows — drafting, analysis, and argumentation with minimal human guidance. That’s the same direction xAI is pushing Grok in, with a focus on real-world tasks rather than just chat.

Big-picture AI optimism: “Intelligence will be free”

Musk also engaged with several posts about the long-term impact of AI on society. One user wrote that “Intelligence will be free,” suggesting a future where access to powerful AI systems is ubiquitous and low-cost. Musk replied, “Everything will be,” implying a broader vision of post-scarcity driven by AI, robotics, and abundant energy.

He shared a historical comparison: in 1927, a three-minute phone call across the Atlantic cost about $75 — over $1,000 in today’s money. Today, you can video chat with someone on the other side of the planet for free. For Musk, this is a preview of what AI could do to the cost of knowledge, labor, and digital services.

In another post, he echoed the idea that a “population of 1 billion builders is coming,” referring to a future where billions of people can create software, products, and companies with the help of AI tools. Musk added that it will be “way more than a billion,” suggesting that AI will turn nearly everyone into a kind of creator or engineer.

Supporters, critics, and political flashpoints

Beyond pure tech, Musk’s feed also featured a mix of praise, criticism, and political controversy.

On the supportive side, a radio guest discussed Musk’s work on AI and technology — from Tesla robotaxis and Grok’s potential role in curing diseases, to Neuralink helping disabled patients and the geopolitical race to stay ahead of China in AI. Musk responded with a heart emoji, signaling appreciation.

He also agreed with a post predicting that AI will make “intelligence free” and that we’re heading toward a world where scarcity is dramatically reduced. These posts align with Musk’s recurring themes: using AI and automation to expand human capability and improve quality of life on a global scale.

On the critical side, actor Mark Ruffalo shared a meme portraying Musk as indifferent to global poverty and urged him to “course correct” and use his wealth to help people rather than “barreling down” a path of chaos. Another user responded by pointing out Ruffalo’s own net worth, properties, and ongoing high-paying roles, calling him hypocritical for lecturing others while sitting on tens of millions of dollars. They also argued that global hunger can’t be solved by a single billionaire writing a check, and that Musk’s charitable giving already exceeds Ruffalo’s net worth many times over. Musk replied with a brief “Yeah,” siding with the criticism of Ruffalo.

Musk also reposted or commented on a range of politically charged topics, including immigration, crime, self-defense laws in Europe, and debates over how history and race are taught in schools. These posts reflect his broader use of X as a platform for culture-war commentary alongside tech updates.

Light moments: world records, forests, and vegan memes

Amid the heavy topics, Musk kept some lighter content in the mix. He reacted with “Wow” to a post about Scottish runner Josh Kerr breaking the mile world record in 3:42, an average speed of 16.2 mph — a pace most people couldn’t hit even briefly, let alone sustain for a full mile.

He also joined in on a joke about a photo from George Washington National Forest, where a user captioned the scene with a playful “The beacons are lit” reference, and Musk responded with laughter. Another meme showed a deer peacefully hanging out with humans, poking fun at how some vegans imagine animals live in the wild; Musk replied “100,” indicating strong agreement with the joke.

Grok Imagine and long-form video generation

To close out the day, Musk shared a showcase of “Grok Imagine,” highlighting its new ability to generate longer-length videos from prompts. While details were sparse, the demo suggests xAI is expanding beyond text and code into multimodal capabilities, competing with other text-to-video tools in the broader AI ecosystem.

Longer, more coherent AI-generated video is a key frontier for creative tools, marketing, simulation, and even training data — and Musk’s promotion of Grok Imagine signals that xAI wants a seat at that table as well.

What Musk’s X activity tells us about his AI strategy

Across all of these posts, a few themes stand out. Musk is doubling down on AI as a full-stack capability: Grok for reasoning and prediction, Grok Build for end-to-end app development, Grok Imagine for video, and Starlink as the global connectivity layer underneath it all. He’s also framing AI as the engine of a future where intelligence, and eventually many goods and services, become effectively free.

At the same time, he’s using X to shape narratives around regulation, safety, and culture — from self-driving cars and trial lawyers to education and political ideology. For followers trying to understand where Musk is taking xAI and his broader tech empire, days like this on X offer a real-time window into both the products he’s pushing and the worldview behind them.