Roko's Basilisk
Posts
When OpenAI Calls Google for Backup

When OpenAI Calls Google for Backup

Plus: Boardroom alliances, blockbuster sales, and your AI weekend starter pack.

Roko's Basilisk
June 13, 2025

Here’s what's on our plate today:

🤝 OpenAI and Google’s surprise partnership rewrites the rules of AI.
📊 Would you trust Big Tech rivals with your cloud and your data?
🧰 3 tools to level up your weekend hustle
🎮 Nintendo smashes sales records, Klarna’s AI hotline, and Australia bets on AI.
🚀 Make music with Suno, remix AI art, and chat with a wellness bot.

Let’s dive in. No floaties needed…

143,596 founders, investors and leaders read this.

Founders and leaders who read Open Source CEO end up with 2.3x the cerebral horsepower of those who don’t.

Ok, we cannot actually prove that, but we think it’s about right. What we do know is that 140k+ readers from Google, TikTok, OpenAI and Deel love our deep dive business content.

Subscribe for free to see what it’s all about.

^{*This is sponsored content}

The Laboratory

What the OpenAI-Google deal says about the future of computing

Every time you ask ChatGPT or any other chatbot a question, they are able to answer based on their training. But have you ever wondered where these models actually run, and how they manage to process an endless stream of queries without crashing?

Every AI model like ChatGPT, Claude, and Gemini is powered by large language models that are trained on massive amounts of data. This data could include billions or even trillions of parameters, which makes their training and inference, generating responses, computationally intensive. Since not everyone can own computer systems capable of housing and running these models, AI companies rely on cloud computing.

Cloud providers offer specialized AI hardware (e.g., NVIDIA A100/H100 GPUs, Google TPUs) that is far more powerful than a laptop or smartphone. Companies rent computing power from cloud providers, instead of purchasing physical servers, to bring down their costs. This brings us to a rare instance where two rivals, OpenAI and Google, signed a deal, putting aside their rivalry in the AI domain.

The deal between OpenAI and Alphabet’s Google comes as the ChatGPT-maker looks to meet its growing needs for computing capacity.

Under the terms of the deal, OpenAI plans to use the Google Cloud service as it works on bigger, more capable AI models. While this may sound straightforward, it could have far-reaching implications for both Google and OpenAI, and for other players operating in the AI space. Let us take a closer look.

Microsoft loosens its grip on OpenAI

Since 2019, Microsoft has had arrangements with OpenAI that gave the Windows maker exclusive rights to build new computing infrastructure for OpenAI. However, in January 2025, the company announced it had approved OpenAI’s ability to build additional compute capacity for research and training models.

Soon after the announcement, OpenAI announced a joint venture with Oracle and Japan's SoftBank Group to build up to $500 billion of new AI data centers in the United States. The deal was announced as part of the "Stargate" initiative, aimed at helping the United States stay ahead of China and other rivals in the global AI race.

OpenAI, earlier this year, also signed a deal with CoreWeave, an artificial intelligence startup backed by Nvidia. Under the terms of the deal, CoreWeave will provide AI infrastructure to OpenAI while boosting CoreWeave ahead of its US IPO.

For reference, CoreWeave provides access to data centers and high-powered chips for AI workloads, mainly supplied by Nvidia. It competes against cloud providers such as Microsoft's Azure and Amazon's AWS.

However, the most interesting deal came when two rivals in the AI arena chose to partner for cloud computing.

A strategic pact between rivals

The deal between OpenAI and Google comes at a time when cloud computing has seen a substantial increase in demand, in large part due to the rise of generative AI. In 2024, revenues in cloud computing grew by $60 billion, or 22%, as AI companies looked to pay-as-you-go models to minimize infrastructure costs and scale resources as needed.

The deal had been under discussion for months and will give the ChatGPT-maker access to Google’s cloud service to meet its growing needs for computing capacity.

The deal will also allow Google to earn substantial revenue from OpenAI by communicating via CoreWeave or directly, which could give it a leg up in the race for cloud profitability against AWS and Azure, which combined control more than 50% of the market.

The deal will also allow Google to expand the use of its in-house chip known as Tensor Processing Units, or TPUs, which the company has historically reserved for internal use. With increased usage, Google looks to prove the prowess of its chips and win more customers, including Big Tech players like Apple, and startups like Anthropic and Safe Superintelligence, both of which are OpenAI competitors launched by former OpenAI leaders.

The deal, then, is mutually beneficial for both companies. However, this deal could have larger implications for other Big Tech companies as well.

How this deal could shake Nvidia

The AI industry currently relies heavily on Nvidia’s products for machine learning and data center workloads. The AI wave helped push Nvidia’s data center revenue in 2025, reaching 115.2 billion U.S. dollars, a 142% increase compared to the previous year.

Currently, the company’s H200 and H100 GPUs have been the favored chips to help power AI, sourced by the likes of Microsoft, ByteDance, and Tesla for AI developments.

However, with Google designing and deploying its own TPUs, Nvidia's dominance in the AI chip market faces new challenges. It is estimated that Google may be able to run its AI compute at a fraction of the cost when compared to chips from Nvidia. Google’s deal with OpenAI signals that this cost-performance trade-off is now compelling enough to attract the biggest players. Delays in deliveries and reports of performance issues in Nvidia’s Blackwell chips could further compound the chip maker's troubles.

In such a space, OpenAI and Google’s willingness to work together could spell trouble for Nvidia as it demonstrates that, for the right price and performance, even the biggest AI labs are willing to work on alternative hardware stacks.

Compute is the new AI battleground

The partnership between OpenAI and Google Cloud underscores a fundamental reality of the generative AI era: compute trumps competition. As AI models grow ever larger and more complex, access to high-performance cloud infrastructure becomes not just a business decision but a strategic imperative. The line between competitor and collaborator is blurring, and survival in the AI arms race increasingly depends on who can deliver the fastest, most cost-effective compute, regardless of brand allegiances.

For OpenAI, whose appetite for compute has outpaced what even Microsoft Azure can reliably supply, the move to tap into Google’s TPU-rich infrastructure through CoreWeave is both practical and prescient. It allows the company to de-risk its operations, gain flexibility, and avoid overreliance on a single provider.

For Google, opening up its prized TPUs to a competitor marks a calculated bet: that infrastructure scale and chip performance can become a lucrative, independent business, regardless of who owns the AI models riding on top.

The arrangement also sends ripple effects through the broader ecosystem. Microsoft may need to reassess its platform lock-in strategy, while Amazon AWS, long a leader in cloud, now faces pressure to compete on specialized AI hardware. And for Nvidia, the deal is a potential inflection point. While its GPUs still dominate the market today, the growing maturity and cost-effectiveness of alternatives like TPUs could erode that position, especially if other major AI labs follow OpenAI’s lead.

TL;DR

OpenAI and Google just inked a surprise cloud deal—putting rivalry aside to supercharge ChatGPT with Google’s hardware.
Microsoft loosens its grip on OpenAI, while CoreWeave and other cloud partners jump in.
The partnership shakes up Nvidia’s AI chip dominance, as Google bets big on its own TPUs.
The real AI arms race? It’s about computing, not just code.

Welcome to the human-verified internet. Bring your brain. And maybe a watermark…

Quick Poll

🗳️ Would you trust two AI rivals to collaborate on your cloud computing?

Stats of the Week

Airlines Sold 160 Million+ Passenger Records to DHS
A new report revealed that major airlines handed over data on more than 160 million passengers to U.S. Homeland Security, fueling privacy and surveillance concerns.
Meta Sues “CrushAI” for AI Nude Ads—App Had 500K+ Installs
Meta filed a lawsuit against the creators of an “AI nudify” app that ran deceptive ads and saw over half a million installs before being banned.
OpenAI-Google Partnership Could Boost Google’s AI Revenue by 15%
Analysts estimate Google’s cloud business may see a 15% bump in AI-driven revenue in 2025, thanks in part to its new partnership with OpenAI.

Global hiring made easy—quality talent, big savings.

Why limit your hiring to local talent? Athyna gives you access to top-tier LATAM professionals who bring tech, data, and product expertise to your team—at a fraction of U.S. costs.

Our AI-powered hiring process ensures quality matches, fast onboarding, and big savings. We handle recruitment, compliance, and HR logistics so you can focus on growth. Hire smarter, scale faster today.