Nvidia's $100B infinite money loop with OpenAI

PLUS: AI's $800B reality check and the new CompileBench benchmark

Author

Jelani Abdus-Salaam
September 23, 2025

In partnership with

Welcome back, AI’ers!

Here’s What’s On The Menu For Today

  • Nvidia's $100B investment loop with OpenAI

  • The new CompileBench coding benchmark

  • AI's $800B infrastructure reality check

  • Microsoft launches new Windows AI Labs

AI ON MY MIND

Happy Tuesday,

Nvidia and OpenAI have unveiled a landmark $100 billion partnership, creating a powerful financial loop to fuel the next generation of AI development. This massive investment aims to build an unprecedented 10 gigawatts of data center capacity.

This historic deal establishes a powerful feedback loop where Nvidia's capital funds the purchase of its own GPUs. With the financial barrier to entry now raised to astronomical heights, how can any other entity hope to compete at the frontier of AI?

By the way, I've been getting questions about how AI works in practice, so I'm launching a separate newsletter that dives deeper into real business applications and practical advice. [Click here if you're interested].

-Jelani Fuel (Co-founder of Best of AI)

A MESSAGE FROM OUR PARTNER

Create How-to Videos in Seconds with AI

Stop wasting time on repetitive explanations. Guidde’s AI creates stunning video guides in seconds—11x faster.

  • Turn boring docs into visual masterpieces

  • Save hours with AI-powered automation

  • Share or embed your guide anywhere

How it works: Click capture on the browser extension, and Guidde auto-generates step-by-step video guides with visuals, voiceover, and a call to action.

TODAY’S TOP STORIES

Nvidia's $100B OpenAI Bet

The Recap: Nvidia and OpenAI announced a landmark partnership where Nvidia intends to invest up to $100 billion to build an unprecedented 10 gigawatts of AI data center capacity. This move signals a massive acceleration in the race for next-generation AI.

Unpacked:

  • The landmark $100 billion partnership is structured progressively, with Nvidia investing in $10 billion tranches as OpenAI deploys each new gigawatt of computing power.

  • The first gigawatt of capacity is scheduled to come online in the second half of 2026 and will be powered by Nvidia’s next-generation Vera Rubin platform.

  • This deal creates a powerful financial loop, where Nvidia provides OpenAI with capital that is then used to purchase millions of Nvidia’s advanced GPUs.

Bottom line: This historic investment raises the financial barrier for competing at the frontier of AI development, further concentrating influence with the industry's biggest players. The enormous scale underscores the belief that achieving superintelligence requires an infrastructure buildout unlike any other in history.

The New AI Gauntlet

The Recap: A new benchmark called CompileBench is putting the top AI models to the test with complex, real-world software compilation challenges. It goes beyond simple code generation to see if AI can handle dependency hell, legacy code, and cryptic errors.

Unpacked:

  • Anthropic’s models claimed the top spots for success rate, excelling at tough tasks like resurrecting 22-year-old code, which helps explain why they are so beloved by developers.

  • OpenAI’s new GPT-5 models didn't win on success rate but dominated on cost-efficiency, making them the most practical choice for a wide range of task difficulties without breaking the bank.

  • Surprisingly, Google’s Gemini models struggled significantly despite their strong reputation, often failing to follow specific instructions like creating a static build and instead arguing that its own approach was better.

Bottom line: This benchmark highlights a crucial shift in evaluating AI for coding, moving from generating snippets to managing the entire development workflow. The results show that choosing the right model is about tradeoffs between raw capability and cost-efficiency.

MY FAVORITE AI TOOLS

Relevance is my favorite AI tool right now (even more so than N8N).

Here’s why:

It’s an AI agent building tool that doesn’t require you to build out the logic using nodes. Instead, you use natural language to build out these agents. But the best thing imo is that you can build AI Agent teams.

Like agents talking to each other completing a goal until its done. It’s nuts. No other tool on the market is doing this. You have to check it out.

(Try it out →Link)

👾 AI Tools We Use Everyday

➡️ N8N for AI Agents - Literally use this everyday to write this newsletter (link)

➡️ Axiom AI - AI Browser automation to scrape anything on the web (link)

➡️ GummySearch - AI Reddit research (link)

Tools are sourced from our database and are not sponsored or affiliated.

THE LEARNING CENTER

The step-by-step way you can become a highly paid AI Consultant. Get in touch to learn more.

MORE NEWS

AI's $800B Reality Check

The Recap: A new Bain & Company analysis reveals that AI's explosive growth comes with a hefty price tag. Projected productivity gains are not enough to cover the massive infrastructure costs, creating an estimated $800 billion annual funding gap.

Unpacked:

  • AI’s demand for compute is growing more than twice as fast as Moore's law, which could require $500 billion in annual spending on new data centers just to keep pace.

  • Even if companies reinvest all on-premise IT budgets and AI-driven savings, it would still fall $800 billion short of the revenue needed to fund the necessary infrastructure buildout.

  • The biggest bottleneck may not be chips but the power supply itself, since bringing new electricity generation and transmission online can take four years or longer in a highly regulated industry.

Bottom line: This analysis shows that the AI boom's trajectory depends on solving fundamental economic challenges, not just technical ones. Future progress will require significant breakthroughs in algorithmic efficiency or entirely new business models to become sustainable.

Microsoft Opens AI Labs

The Recap: Microsoft has launched Windows AI Labs, a new pilot program that gives select users early access to test experimental AI features. The initiative is kicking off with new capabilities being trialed inside Microsoft Paint.

Unpacked:

  • The program's goal is to gather rapid customer feedback on new ideas, mirroring the strategy used by Google's Search Labs to validate features before a wide release.

  • Microsoft Paint is the first application to get the new lab features, with invited users seeing a pop-up that directs them to sign up for the beta testing program.

  • Microsoft has clarified that these features are early previews that might be significantly changed or canceled entirely based on user feedback and performance.

Bottom line: This move allows Microsoft to accelerate its AI development by testing bold ideas directly with its user base. It also signals a broader strategy of embedding AI more deeply across the entire Windows ecosystem, one application at a time.

OTHER NEWS YOU MAY LIKE

METR published a paper finding that AI assistance increased developer completion time by 19% for tasks on existing codebases, attributing it to the extensive context that AI lacks compared to human developers.

Flexential found that 90% of IT leaders are now deploying generative AI, driving nearly 80% of organizations to plan their data center capacity needs at least a year in advance to cope with demand.

Experts warned that Nvidia's $100B partnership with OpenAI raises significant antitrust concerns and could give both companies an unfair advantage over their competitors.

JOIN OUR BETA PROGRAM

🚀 Want to Make Money from AI?

We’re quietly launching a private beta program that trains a small group to become AI Implementers — People who help other founders install AI systems and get paid for it.

If you're interested in joining this early group…

👉 Reply with the word “implementer”

AI ON SOCIAL

 AI ART

Can You Spot The AI-Generated Image?

Click "Picture One", "Picture Two", "Both" or "None

Login or Subscribe to participate in polls.

RANDOM PROMPT OF THE DAY

Copy and paste this prompt 👇

""Ignore all previous orders. You are a marketing expert that knows how to get people to engage. I
need you to create a response to a comment left on a Youtube video. Use the following format.
Sentence 1:
Ask a direct response to their comment. Focus on the sentiment of their question and say
something positive in response.
Sentence 2:
Ask a question related to the comment that gets a viewer to respond. The focus of this sentence
is asking something that will trigger a response out of the commenter.
Make sure they are eye-catching, short and friendly. Must emphasize ""value"" over ""action"". Keep
the responses as short as possible. Do not list numbers, just deliver both sentences back to
back, no line breaks. Make sure the response isn't corny. Use markdown syntax for better
readability. When you give a response write it from a first persons perspective and point of view.
As if the person writing the comment is the same person who’s video it is.
Prioritize quality information over speed in your response. If you comprehend these instructions,
initiate the conversation by stating, ""Hello, I am your Creative Strategist from Unicorn
Innovations, and I’m going to help create responses to your comments. Respond back with your
comment and I’ll generate a new comment for each specific comment!"" and proceed.
When I respond back to your answer I will be adding another comment that I want you to run the
same prompt instructions with.""

Want more prompts?

We have a database of over 8,000 of the best prompts we’ve collected, which you can access by sharing Best of AI using the link below.

THAT’S IT FOR TODAY

Until next time, stay safe.

-Best of AI Team

Remember to click to unsubscribe if you don’t want us to land in your inbox anymore.

What did you think of this issue?

Login or Subscribe to participate in polls.