Apple’s Asa & Microsoft MAI: The AI power shift begins

In partnership with

Dear Reader,

What if every Apple Store employee could instantly explain the iPhone’s camera tricks like a pro, and your favorite AI-powered apps were running on Microsoft’s own brainpower—not OpenAI’s?

That’s not a vision for 2030; it’s unfolding right now.

Flipped.ai’s weekly newsletter is read by thousands of AI professionals, entrepreneurs, and decision-makers, and today we’re unpacking two major shifts in AI that you probably haven’t seen coming:

• 🍏 Apple’s Asa AI chatbot: A private, in-house assistant that’s transforming retail training and enterprise workflows.

• 🖥️ Microsoft’s MAI initiative: Its first homegrown AI models, a bold step toward independence from OpenAI.

Whether you’re building AI systems, running teams, or just curious about the next wave of innovation, this week’s updates show how fast the AI race is changing—from retail counters to massive cloud clusters.

Before we dive in, a quick thank you to our sponsor, Cartograph.

The Key to a $1.3 Trillion Opportunity

A new trend in real estate is making the most expensive properties obtainable. It’s called co-ownership, and it’s revolutionizing the $1.3T vacation home market.

The company leading the trend? Pacaso. Created by the founder behind a $120M prior exit, Pacaso turns underutilized luxury properties into fully-managed assets and makes them accessible to the broadest possible market.

The result? More than $1B in transactions and service fees, 2,000+ happy homeowners, and over $110m in gross profit to date for Pacaso.

With rapid international growth and 41% gross profit growth last year alone, Pacaso is hitting their stride. They even recently reserved the Nasdaq ticker PCSO.

The same VCs that backed Uber, eBay, and Venmo also backed Pacaso. Join them as a Pacaso shareholder before the opportunity ends September 18.

Invest While You Still Can

_{Paid advertisement for Pacaso’s Regulation A offering. Read the offering circular at invest.pacaso.com. Reserving a ticker symbol is not a guarantee that the company will go public. Listing on the NASDAQ is subject to approvals.}

Apple’s Asa: AI that trains the trainers

Apple has rolled out Asa, a new AI chatbot that’s not for customers, but for its own retail employees. Asa is embedded into SEED, Apple’s in-house training app, and acts like a “sales coach in your pocket,” ready to deliver:

• Instant product knowledge: Employees can ask Asa about specs, features, or pricing of Apple’s latest devices.

• Sales strategies: Asa helps staff handle objections, upsell, and explain complex tech in simple terms.

• Customer insights: The bot offers tips on connecting with different customer personas, enhancing in-store experiences.

This tool is more than just a Q&A assistant—it’s Apple’s way of standardizing training globally across its 500+ stores, ensuring every employee speaks the same Apple language.

Years in the Making

Apple’s push into AI training tools isn’t new:

• In 2023, reports from The Verge revealed Apple was quietly testing AI-powered customer support bots.

• By 2025, Asa’s launch proves Apple has refined its internal AI to a level that rivals tools like ChatGPT or Gemini.

• Apple also recently introduced AI-driven chat features in the Apple Support app for direct customer help, showing a broader shift toward AI-first services.

Privacy as a differentiator

Apple knows privacy is its superpower. While competitors like Microsoft and Google lean on cloud models, Apple is:

• Keeping sensitive data locked down with on-device AI where possible.

• Giving enterprises control over AI access on company devices.

• Positioning itself as the “privacy-first” AI provider, which could be a strong selling point for industries like healthcare or finance.

What’s Next for Asa?

Insiders hint at exciting possibilities:

• Augmented reality training: Imagine an AR demo showing how to set up the Vision Pro right in the store.

• Hyper-personalized coaching: Asa could adapt training recommendations based on each employee’s sales history.

• Integration into customer interactions: While Asa is internal-only now, Apple could eventually use this tech to guide customers directly in-store.

Apple’s strategy is classic Apple: start small, nail the experience, and scale globally.

Microsoft MAI: From OpenAI partner to rival

Microsoft just announced its first AI models fully built in-house under its Microsoft AI (MAI) team:

1. MAI-Voice-1: A speech generation model with natural tone, expressive voices, and lightning-fast processing.

2. MAI-1-preview: A text-based LLM designed for instruction-following and conversation, currently ranked 15th on LM Arena—above GPT-4.1 Flash.

This is a massive shift. For years, Microsoft was OpenAI’s infrastructure backbone, investing billions and embedding GPT models in Copilot. Now, it’s signaling independence, reducing reliance on OpenAI, and competing head-to-head.

Under the hood

• Training scale: MAI-1-preview was trained on 15,000 NVIDIA H100 GPUs. While smaller than xAI’s Grok (~200k GPUs) or GPT-5 (~200k GPUs), it’s a strong debut.

• Efficiency first: MAI-1 uses a mixture-of-experts (MoE) approach, activating only parts of the model per query to save compute and speed up responses.

• Voice tech breakthrough: MAI-Voice-1 can generate a full minute of audio in under a second on a single GPU, making it one of the fastest voice models available.

Deployment and access

• MAI-Voice-1 is already powering Copilot Daily, Podcasts, and Labs, letting users create custom voices and storytelling experiences.

• MAI-1-preview is being tested in LM Arena, where users can pit it against models like Gemini 2.5 and Claude 3.5.

• Early API access is limited, but this will expand as Microsoft gathers real-world feedback.

The strategic play

Microsoft’s move is about control, cost, and scale:

• By building its own models, Microsoft reduces dependence on OpenAI, which became a competitive tension point after leadership changes in 2023.

• Vertical integration lets Microsoft tune AI for Windows, Teams, Office, and Azure; its biggest strengths.

• Expect a rapid iteration cycle as Microsoft scales its next-gen GB200 clusters.

This is Microsoft saying: “We’re not just the infrastructure provider. We’re here to lead the AI race.”

What does this means for the AI ecosystem?

Apple and Microsoft are taking very different approaches:

Both approaches point to one clear trend:

For AI professionals, this means:

• More specialized AI roles in enterprise and retail sectors.

• A shift toward custom LLMs instead of generic models.

• Faster innovation cycles as Big Tech races to differentiate.

Closing thoughts

This week’s moves from Apple and Microsoft mark a turning point in AI’s evolution. Apple is bringing AI deeper into its ecosystem, quietly reshaping employee training and enterprise services. Microsoft is no longer just OpenAI’s biggest partner, it’s becoming a serious AI powerhouse in its own right.

The race is no longer about a single “best model.” It’s about who can deliver AI that feels invisible, personal, and deeply integrated into our tools and workflows.

If these launches are any sign, 2025 will be the year AI stops being a separate product and becomes the core of how we work, sell, and create.

Stay tuned, next week, we’ll dive into more AI updates that are rewriting the rules faster than ever.

Want to get your product in front of 75,000+ professionals, entrepreneurs, decision-makers, and investors around the world? 🚀

If you are interested in sponsoring, contact us at [email protected].

Thank you for being part of our community, and we look forward to continuing this journey of growth and innovation together!

Best regards,

Flipped.ai Editorial Team