đź’» Microsoft Takes On Nvidia

Plus OpenAI adds $1B in one month and AI creates 200K new founders

Reading time: 5 minutes

🗞️In this edition

  • Microsoft Launches Maia 200 to Reduce Nvidia Dependence

  • Sponsored: Picsart Creative APIs: One Hub. 20+ AI Models. Better Cutouts. 30% OFF.

  • OpenAI API Business Adds $1 Billion in One Month

  • Davos Says AI Makes Starting Businesses Easier for Non-Coders

  • In other AI news –

    • EU Investigates Grok for Creating Sexual Images of Children

    • Oracle CEO Says All AI Models Miss Private Data

    • AI Drove $262 Billion in Holiday Shopping Sales

  • 4 must-try AI tools

Three moves this week that reveal how fast things are shifting.

One company just launched something to cut ties with its biggest supplier. Another quietly built a billion-dollar business nobody talks about. And a third trend is creating thousands of new businesses that all look exactly the same.

Each story shows something about control. About reducing dependence. About building invisible infrastructure. About democratization that might not be what it seems.

The patterns are getting clearer when you look at who's trying to own what.

This edition looks at the infrastructure plays happening while everyone focuses somewhere else.

What's happening:

Microsoft launched its second-generation AI chip called Maia 200 on Monday. The chip is already running in data centers in Iowa, with Arizona coming next.

Built using TSMC's 3-nanometer process, Maia 200 delivers "efficient performance per dollar" according to Microsoft. The chip is designed for large-scale AI workloads and reasoning models.

Microsoft says Maia 200 offers 30% better performance than comparable chips from Amazon and Google at the same price. It can connect up to 6,144 chips together for massive computing power.

The company is already designing Maia 300, the next version. Microsoft's superintelligence team led by Mustafa Suleyman will use Maia 200, along with Microsoft 365 Copilot and Azure cloud services.

Why this is important:

Microsoft is following Amazon and Google by building its own chips. All three want to reduce dependence on Nvidia, which controls the AI chip market and charges premium prices.

This doesn't mean ditching Nvidia entirely. Microsoft still needs Nvidia chips for training large models. But for running AI after training, Microsoft can use cheaper in-house chips.

The timing matters. Nvidia chips are expensive and hard to get. When you're spending billions on AI infrastructure, not controlling your own hardware becomes a business risk.

Microsoft waited years while Amazon and Google built their own chips. Now they're playing catch-up with a second-generation chip while competitors are several generations ahead.

We’re kicking off the year with major upgrades to Picsart Creative APIs. Our API suite for building AI-powered image and video workflows directly into your product, pipelines, and internal tools. Plus, for a limited time, you can get 30% OFF.

What’s new in the Creative APIs:

  • Remove Background API is upgraded with a new model for noticeably cleaner results and sharper edges

  • AI Providers Hub: choose from 20+ models to generate images, stickers, logos, and more

  • 4 new services to expand your end-to-end image workflow capabilities

If you’re building AI-powered creative, e-commerce image pipelines, or automated content ops, Picsart Creative APIs are the fastest way to level up output quality without adding workflow complexity.

👉 Claim 30% OFF and start building today.

What's happening:

"People think of us mostly as ChatGPT, but the API team is doing amazing work!" Altman wrote on X. The API lets other companies embed OpenAI's models into their own products.

Many high-profile startups depend on OpenAI's API as core infrastructure. Perplexity uses it for AI search. Harvey, a legal tech startup, uses it to help lawyers with research and drafting.

The API business is growing faster than ChatGPT despite massive computing costs. OpenAI burned through $8 billion in 2025 and faces $1.4 trillion in infrastructure commitments.

Why this is important:

OpenAI is becoming infrastructure, not just a product. The API makes them more like Amazon Web Services than like a consumer app.

Adding $1 billion in one month shows enterprise adoption is outpacing consumer use. Businesses pay for reliable access to models. Consumers pay $20 monthly or use it free.

The API revenue is predictable and high-margin. ChatGPT experiments with ads and pricing tiers. The API just charges per use at scale.

This explains why Altman keeps saying OpenAI isn't primarily ChatGPT anymore. The invisible backend business is bigger than the consumer-facing product everyone knows.

Comments from the editor:

Altman announced this days after saying OpenAI will test ads in ChatGPT. That timing isn't random. He's saying "don't worry about ChatGPT monetization, look at the API."

The API burns money too. OpenAI burned $8 billion last year. $1 billion in monthly additions doesn't solve $1.4 trillion in commitments. It just makes the story sound better.

"People think of us mostly as ChatGPT" is Altman managing perception. Most people do think of ChatGPT. That's the product consumers use. The API is invisible unless you're a developer.

Perplexity and Harvey building on OpenAI's models means OpenAI controls their infrastructure. If OpenAI raises prices or restricts access, those companies are stuck. That's the real business model.

What's happening:

World Economic Forum experts at Davos said AI is making entrepreneurship easier for people without technical skills. Tools like Jasper for writing, Canva's AI for design, and no-code builders like Bubble let anyone create businesses without coding.

This matters most for underrepresented groups. Women and minority entrepreneurs who traditional venture capital ignores are using AI to start companies without outside funding.

LinkedIn shows a 20% jump in users calling themselves "entrepreneur" or "founder" over the past year, linked to AI adoption. Young professionals facing uncertain job markets see startups as safer than employment.

Over 200,000 tech and finance jobs were cut in 2023. Many laid-off workers from Google and Meta are using AI to launch businesses in areas like personalized education and sustainable fashion.

Why this is important:

"Democratizing" entrepreneurship sounds good until you realize everyone is using the same AI tools pulling from the same data. That creates thousands of similar businesses, not diverse ones.

The New York Times warned that AI-generated content is flooding markets and devaluing human creativity. When algorithms decide what businesses to build, they copy what already exists.

This isn't empowerment if the AI does the creating. People become managers of machines, not entrepreneurs with original ideas. That's a different kind of business owner.

The 20% LinkedIn increase could mean real entrepreneurship or people calling themselves founders while AI does their work. There's a difference between building something and supervising an algorithm.

  • EU Investigates Grok for Creating Sexual Images of Children – The European Commission opened a formal probe after users generated 3 million sexualized images in days using prompts like "put her in a bikini," potentially including child abuse material that Malaysia and Indonesia already banned.

  • Oracle CEO Says All AI Models Miss Private Data â€“ Larry Ellison argues ChatGPT, Gemini, Grok, and Llama are trained only on public internet data and don't know your company's accounts, customers, or contracts—making them commodities without real business value.

  • AI Drove $262 Billion in Holiday Shopping Sales – Salesforce says AI influenced 20% of all retail sales as traffic from ChatGPT and Perplexity jumped 693%, but most impact came from customer service chatbots, not shopping assistants.

👩🏼‍🚒Discover mind-blowing AI tools

  1. Bot Butcher - An anti-spam tool designed to protect your inbox from unwanted messages

  2. TTcare - An AI-powered mobile app that helps pet parents monitor the health of their dogs and cats

  3. Alter AI - A tool that transforms amateur selfies into professional-looking headshots

  4. Convai - A platform to create NPC characters and assets for games

The real story is always about who controls the foundation.

Building chips to escape supplier lock-in. Growing an API business that makes others dependent. Spreading tools that create uniform businesses instead of diverse ones.

None of these are about innovation. They're about control, dependency, and leverage. About who owns the infrastructure everyone else builds on top of.

What matters now is whether people notice they're becoming dependent before it's too late to switch. Or whether control gets locked in while everyone celebrates democratization.

If you have any feedback for us, please reply and let us know how we did. We're always looking to improve.

How did we like this version?

Login or Subscribe to participate in polls.

Interested in featuring your services with us? Email us at [email protected]