OpenAI to build its Own Chips

Hello Readers👀 

Curious about the biggest AI moves today? You’re in the right place. In today’s edition, we have:

⚙️ OpenAI and Broadcom Team Up to Build 10GW of Custom AI Accelerators

🧠 Improve Your Ads Before You Spend

⚖️ California Becomes First State to Regulate AI Companions

🔨Tools you cannot miss

⚙️ OpenAI and Broadcom Team Up to Build 10GW of Custom AI Accelerators

OpenAI just announced a strategic collaboration with Broadcom to co-develop and deploy 10 gigawatts of AI accelerator infrastructure, a massive step toward building the hardware foundation for next-generation intelligence.

The Shift:

1) Building Next-Gen AI Hardware Together - OpenAI will design the accelerators and systems, embedding its learnings from frontier models directly into the hardware. Broadcom will lead manufacturing and deployment, equipping racks with Ethernet, PCIe, and optical connectivity to support large-scale AI workloads.

2) Deployment Timeline and Scale - The first racks will go online in late 2026, with full rollout targeted by 2029 across OpenAI’s data centers and partners. The 10GW infrastructure aims to handle rising global AI demand while improving cost efficiency and scalability for OpenAI’s expanding ecosystem.

3) Strategic Shift Toward Custom Chips - This move positions OpenAI alongside Google and Amazon, both of which design their own AI chips. By developing custom accelerators, OpenAI gains greater control over performance, cost, and availability, potentially reducing dependence on Nvidia, though it continues partnerships with Nvidia and AMD.

The deal marks OpenAI’s most ambitious hardware bet yet, signaling a shift from model-building to infrastructure ownership. If successful, it could redefine AI’s compute economics, turning OpenAI from a model leader into a full-stack AI powerhouse.

^{TOGETHER WITH SUPERHUMAN}

Go from AI overwhelmed to AI savvy professional

AI keeps coming up at work, but you still don't get it?

That's exactly why 1M+ professionals working at Google, Meta, and OpenAI read Superhuman AI daily.

Here's what you get:

• Daily AI news that matters for your career - Filtered from 1000s of sources so you know what affects your industry.

• Step-by-step tutorials you can use immediately - Real prompts and workflows that solve actual business problems.

• New AI tools tested and reviewed - We try everything to deliver tools that drive real results.

• All in just 3 minutes a day

Join 1M+ pros

🧠 Improve Your Ads Before You Spend

Most teams wait until ads go live to see what works. That’s too late. With Neurons, you can test every creative before launch using neuroscience-backed predictions, so you optimize impact, not guess it.

✅ How to use it:

1. Upload your creative – image or video

2. Get your Impact Score instantly – 1 to 10, based on attention, recall, and emotion

3. See what to fix – with science-backed tips that boost ROI

Brands using it have doubled conversions and lifted CTRs by 73% all before spending a dollar.

👉 Try Neurons and test every creative before it runs.

⚖️ California Becomes First State to Regulate AI Companions

California has made history with Senate Bill 243 (SB 243), the nation’s first law regulating AI companion chatbots. Signed by Governor Gavin Newsom, it establishes new safety, transparency, and accountability standards for companies like OpenAI, Meta, and Character AI, marking a turning point in AI governance.

The Shift:

1) Strong Safety Rules for AI Companions - The law mandates age verification, suicide-prevention systems, and content safeguards to protect minors and vulnerable users. Chatbots must offer break reminders, avoid sexual or romantic interactions with minors, and cannot pose as healthcare professionals.

2) Accountability and Enforcement - Violations can trigger fines of up to $250,000 per offense, especially for harmful deepfakes or noncompliance with safety protocols. Companies must also share their crisis management statistics and prevention actions with California’s Department of Public Health for state monitoring.

3) AI Disclosure and Annual Reporting - If users might mistake an AI chatbot for a human, the platform must clearly disclose it’s artificial. Developers are required to submit yearly reports to the Office of Suicide Prevention, detailing safeguards against self-harm and how incidents are handled.

Taking effect January 1, 2026, SB 243 sets a national precedent for responsible AI oversight, ensuring child protection and ethical transparency in chatbot design. Together with California’s AI transparency bill (SB 53), it positions the state as the policy leader in shaping safe and accountable AI use across industries.

🤳AI Nugget of the Day

🔨AI Tools for the Shift

🚀 Aiso – Get your brand visible in ChatGPT conversations with AI-powered visibility optimization.

💡 Content Dreamer AI – Generate fresh, relevant content ideas daily so you can focus on building in public and engaging your audience.

📚 Turbo-Learn – Your AI-powered learning assistant for creating concise notes and clear, structured diagrams instantly.

🎨 TattooCoverUp.AI – Upload your existing tattoo photo, input your idea, and get stunning AI-generated cover-up design options.

🎬 1MoreShot – Create professional music videos for Suno, Udio, and any AI-generated music with perfect sync and stunning visuals.

🚀Quick Shifts

💻 Nvidia’s DGX Spark goes on sale October 15th. The $3,999 “personal AI supercomputer” delivers a petaflop of performance, 128GB memory, and 4TB storage, putting data center-level AI power on your desktop for researchers and developers alike.

🖼️ Microsoft AI unveils MAI-Image-1, its first in-house image generator. The new model promises faster, photorealistic results and ranks in LMArena’s top 10, marking Microsoft’s shift from reliance on OpenAI to self-built creative AI systems.

💬 Slack is turning Slackbot into an AI assistant. The upgraded Slackbot can now search, summarize, plan, and coordinate tasks across channels, turning it into a personalized AI companion that helps manage projects and meetings directly inside Slack.

🤖 Salesforce launches Agentforce 360 ahead of Dreamforce 2025. The platform introduces Agent Script for smarter, conditional AI behavior, Agentforce Builder for end-to-end agent deployment, and deep Slack integrations.

🍌 Google’s Nano Banana image model is going platform-wide. The AI editor will now power NotebookLM Video Overviews, Google Photos, and Lens in Search, bringing faster, higher-quality image generation and seamless editing across Google’s ecosystem.

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇