Good morning, AI enthusiast.

Elon Musk is adding more fuel to the summer AI race with the surprise announcement of Grok 4. The model is skipping version 3.5 and will arrive shortly after the July 4th holiday, with a sharp focus on coding.

With claims of 'unmatched' intelligence and a strategic focus on developers, xAI is setting high expectations. The big question is whether Grok 4's specialized coding abilities will be enough to challenge the established players and lure developers into its ecosystem.

In today’s AI recap:

• How to automate generating short form video scripts (watch here)

• Elon Musk announces Grok 4

• OpenAI turns to Google for chips

• Meta's $29B AI data center plan

• Washington's new AI power play

Elon Musk Announces Grok 4

The Recap: Elon Musk revealed that xAI is skipping Grok 3.5 and will release its next flagship model, Grok 4, shortly after the July 4th holiday. The new version is being positioned as a powerful, specialized coding model aiming to compete at the frontier of AI.

Unpacked:

• Musk noted the model needs one more training run specifically to enhance its capabilities as a specialized coding model.

• The jump from Grok 3 straight to 4 signals that xAI considers this a major upgrade, not just an incremental update.

• The xAI team is building high expectations, with one researcher claiming its intelligence will be unmatched and designed to rival other upcoming frontier models.

Bottom line: xAI's sharp focus on coding capabilities shows a clear strategy to attract developers and technical professionals. This announcement intensifies the summer AI competition, promising users another powerful tool in an increasingly crowded field.

OpenAI Turns to Google for Chips

The Recap: In a major strategic shift, OpenAI has started using Google's custom TPU chips to power its products. This move marks the first time the company has meaningfully used non-Nvidia hardware, signaling a new chapter in the AI hardware race.

Unpacked:

• The primary drivers for OpenAI are reducing the high cost of inference computing and diversifying its hardware supply chain away from its heavy reliance on Nvidia and Microsoft.

• This creates a surprising partnership between direct AI competitors, though reports indicate Google is not renting its most powerful TPUs to OpenAI.

• The deal is a significant win for Google Cloud, validating its TPUs as a viable alternative and potentially loosening Nvidia's tight grip on the AI chip market.

Bottom line: This decision introduces serious competition into a market long dominated by a single player. Increased rivalry among chip providers could ultimately drive down costs and accelerate innovation across the entire AI industry.

AI Training

This AI System Makes Viral Reels For Me (n8n free template)

The Recap: In this video, I’m going to break down how we’ve built an AI short form video generation machine that repurposes long form content from The Recap’s daily newsletter, and generates bite-size short form videos that can be shared across Instagram, TikTok, and YouTube shorts.

P.S We also launched a free community for AI Builders looking to master the art and science of building AI Automations — Come join us!

Meta's $29B AI Infrastructure Bet

The Recap: Meta is seeking to raise a massive $29 billion from private capital firms to fund an enormous build-out of its artificial intelligence data centers. This move underscores the immense capital required to compete at the frontier of AI development.

Unpacked:

• The company plans to structure the deal with $26 billion in debt and $3 billion in equity, engaging with firms like Apollo Global Management, KKR, and Carlyle.

• This raise supports a broader strategy, as Mark Zuckerberg previously announced plans to spend up to $65 billion this year to expand the company's AI infrastructure.

• The aggressive infrastructure build-out is designed to strengthen Meta's position against rivals like OpenAI and Google in the race for AI leadership.

Bottom line: This massive capital raise shows that building frontier AI is now a game of immense infrastructure and financial muscle, not just algorithms. It also signals Meta's deep commitment to providing powerful platforms and tools for the entire ecosystem.

Where AI Experts Share Their Best Work

Join our Free AI Automation Community

Join our FREE community AI Automation Mastery — where entrepreneurs, AI builders, and AI agency owners share templates, solve problems together, and learn from each other's wins (and mistakes).

What makes our community different:

• Real peer support from people building actual AI businesses

• Complete access to download our automation library of battle-tested n8n templates

• Collaborate and problem-solve with AI experts when you get stuck

Dive into our course materials, collaborate with experienced builders, and turn automation challenges into shared wins. Join here (completely free).

The White House's AI Energy Plan

The Recap: The Trump administration is preparing executive orders to tackle the enormous energy demands of the U.S. artificial intelligence industry. The plan aims to clear major infrastructure and regulatory hurdles to maintain America's competitive edge in the global AI race.

Unpacked:

• AI's rapid expansion is driving the first significant increase in U.S. power demand in decades, with consumption from data centers projected to grow more than thirtyfold by 2035.

• The proposals include fast-tracking grid connections for new power projects and offering federal land managed by the Department of Defense or Interior for data center construction.

• To accelerate development, the administration is also considering a single, nationwide Clean Water Act permit for data centers to replace the current state-by-state approval process.

Bottom line: This move frames AI's power requirements not just as a logistical challenge, but as a core issue of national strategic importance. By proactively addressing the energy bottleneck, the U.S. aims to secure the foundational infrastructure needed to lead global AI innovation.

The Shortlist

Apple introduced STARFlow, an image generation model family that revives a 'forgotten' technique called Normalizing Flows to efficiently create high-resolution images in latent space.

Audos raised $11.5M in seed funding for its AI-powered platform designed to help non-technical entrepreneurs launch and scale businesses.

Reddit launched Reddit Community Intelligence, a new suite of AI-powered advertising tools that analyze conversations to provide brands with real-time trends and surface user-generated content in ads.

MrBeast shuttered his new AI thumbnail generator tool after widespread backlash from creators who argued that it devalued the work of human artists.

Signing off,

David, Lucas, Mitchell — The Recap editorial team