Build by AI

The Great AI Partnership Shuffle I 28th April

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 32:16
Picture this: It's 2028 and the AI landscape looks nothing like it did just two years ago. The biggest tech partnerships have been completely reshuffled, former allies are in court, and billions of dollars are changing hands faster than you can say GPT. Today we're diving into the seismic shifts happening right now that are redrawing the entire AI industry map - from OpenAI breaking free of Microsoft's exclusive grip to a mysterious new $5 billion AI lab that wants to learn without any human data at all. If you want to understand where AI is really heading, this is the episode that connects all the dots.
SPEAKER_01

Two years from now, you're at a developer conference and someone asks you to name Microsoft's biggest AI partner. And you pause because honestly you're not sure anymore. OpenAI is selling their models on Amazon's cloud. There's this mysterious British lab called Ineffable Intelligence that nobody saw coming. And Sam Altman is testifying in a Northern California courtroom about whether his company should even exist as a for-profit entity. Like imagine if tomorrow Google and Apple suddenly ended their search deal, or if Netflix started exclusively streaming on Amazon Prime. That's the scale of disruption we're talking about here.

SPEAKER_00

Except it's not just one deal. It's the entire foundation of how AI companies work together, compete, and raise money. Everything is up for grabs right now.

SPEAKER_01

You're listening to Build by AI. I'm Alex Shannon. And honestly, I've been staring at these stories for two days, trying to figure out how they all fit together.

SPEAKER_00

And I'm Sam Hinton, and I think I've got it figured out. We're watching the biggest power shift in AI since ChatGPT launched. We've got partnership breakups, billion-dollar lawsuits, mysterious new labs, and geopolitical drama all happening at once.

SPEAKER_01

So let's start with the bombshell that's been years in the making and dive into what this all means for anyone building with AI.

SPEAKER_00

Because if you thought the AI world moved fast before, buckle up.

SPEAKER_01

And seriously, if you're making strategic decisions about AI platforms or partnerships right now, this episode might save you from some very expensive mistakes down the road.

SPEAKER_00

Or help you spot opportunities that everyone else is missing. Let's dive in.

SPEAKER_01

Microsoft and OpenAI just ended their exclusive partnership deal. This is the arrangement that made Microsoft OpenAI's primary cloud provider, gave them exclusive access to OpenAI's models for their products, and set up this whole revenue sharing structure that's been the foundation of their relationship since 2019. And to be clear, this isn't just a small tweak to the contract. They're calling it an amended partnership agreement. But multiple sources are reporting that the exclusive and revenue sharing components are completely done.

SPEAKER_00

Yeah, this is this is massive. This deal was worth potentially tens of billions of dollars to Microsoft. They were getting first dibs on every OpenAI breakthrough, plus they were basically OpenAI's exclusive sales channel for enterprise customers. Now that's all gone.

SPEAKER_01

So what does this mean practically? Is OpenAI just free to partner with anyone now?

SPEAKER_00

Exactly. And we're already seeing it play out. OpenAI just secured concessions to sell their products directly on Amazon Web Services, AWS, which would have been impossible under the old exclusive arrangement. They're basically saying we don't want to be tied to one cloud provider anymore.

SPEAKER_01

But hold on, Microsoft isn't walking away empty-handed here. The reports say they're getting increased revenue sharing benefits. So they're losing exclusivity but getting more money per transaction.

SPEAKER_00

Right, which tells me this was probably a negotiated breakup, not a hostile one. Microsoft realized they couldn't keep open AI locked down forever, so they traded exclusivity for better economics on the deals they do get.

SPEAKER_01

This feels like a fundamental shift, though. For years, Microsoft has been positioning itself as the AI company because of this open AI partnership. Now what's their differentiation?

SPEAKER_00

That's the million-dollar question, or should I say billion-dollar question. Microsoft still has Copilot, they still have their internal AI research, but they just lost their exclusive access to what most people consider the best AI models in the world.

SPEAKER_01

And for open AI, this has to be about their upcoming IPO, right? Investors probably weren't thrilled about being so dependent on one partner.

SPEAKER_00

Absolutely. No public company wants to tell shareholders that their entire business model depends on the goodwill of one other company. This gives them way more flexibility and probably a much higher valuation.

SPEAKER_01

Much but think about this from a developer perspective. If you've been building your entire AI strategy around Microsoft's ecosystem because of their open AI access, suddenly you have way more choices. You could go directly to OpenAI or use them through AWS or stick with Microsoft.

SPEAKER_00

That's both exciting and terrifying, right? More choice is great, but it also means more complex decisions. And pricing is probably going to get really interesting as these companies start competing more directly for the same customers.

SPEAKER_01

I'm curious about the timing though. Why now? OpenAI has been incredibly successful with Microsoft's backing. What changed?

SPEAKER_00

I think it comes down to scale and ambition. OpenAI has grown beyond what any single partnership can handle. They need global infrastructure. They need enterprise relationships that Microsoft doesn't have. They need flexibility to make deals in markets where Microsoft might have conflicts of interest.

SPEAKER_01

And let's be honest, Microsoft probably saw this coming. They've been investing heavily in their own AI research, their own models. They knew they couldn't rely on open AI forever.

SPEAKER_00

Exactly. This might actually be better for both companies long term. Microsoft gets to focus on building their own AI capabilities instead of being dependent on a partner. And OpenAI gets the freedom to become the platform they want to be.

SPEAKER_01

Keep an eye on this because I think we're going to see open AI announcements with Google Cloud, maybe even smaller cloud providers in the next few months. The floodgates are officially open.

SPEAKER_00

And Microsoft is going to have to prove that their AI stack is competitive without exclusive open AI access. That's going to be a real test of their internal capabilities.

SPEAKER_01

For anyone listening who's been putting off AI adoption because they were confused about which platform to choose, this actually makes the decision easier in some ways. You can evaluate each option on its own merits instead of trying to guess which partnerships will last. Now speaking of OpenAI's future being uncertain, we need to talk about this court case that could literally determine whether OpenAI continues to exist as we know it. Elon Musk is suing OpenAI and Sam Altman, and this isn't just about hurt feelings or broken promises. The trial is happening in Northern California, and the court's ruling could determine whether OpenAI can even operate as a for-profit enterprise. And get this, the ruling could potentially result in Sam Altman and other current executives being ousted from the company.

SPEAKER_00

Dude, this is wild. Musk is basically saying that Altman and Greg Brockman deceived him about OpenAI's original mission and funding structure. He's arguing that the company was supposed to remain a nonprofit focused on beneficial AI, not become this for-profit juggernaut.

SPEAKER_01

And the timing here is incredibly suspicious, right? This lawsuit ramps up just as OpenAI is preparing for what could be one of the biggest tech IPOs in history.

SPEAKER_00

Yeah, that's not a coincidence. If Musk wins this case, it could completely derail the IPO. Imagine trying to sell shares in a company when a court might rule that its entire corporate structure is illegitimate.

SPEAKER_01

But let's play devil's advocate for a second. Does Musk have a legitimate point here? OpenAI did start as a nonprofit with this mission of democratizing AI research.

SPEAKER_00

OpenAI went from we're going to open source everything to our models are too dangerous to release pretty quickly. But suing to potentially destroy the company, that feels more like revenge than principle.

SPEAKER_01

And meanwhile, Musk is building his own AI company, XAI, which is also for profit. So it's not like he's opposed to commercial AI development in general.

SPEAKER_00

Exactly. This feels like if I can't have the biggest AI company, then maybe no one should. Which is a pretty dangerous precedent for the industry.

SPEAKER_01

What's really concerning to me is the uncertainty this creates for developers and businesses building on OpenAI's platform. How do you make long-term plans when the company's entire legal structure is in question?

SPEAKER_00

That's the real impact here. Even if OpenAI ultimately wins this case, the legal cloud hanging over them makes them a riskier partner. And in an industry where trust and stability are crucial, that's a serious problem.

SPEAKER_01

I've been thinking about this from a business perspective. If you're a Fortune 500 company, considering a major open AI integration, this lawsuit has to be giving your legal team nightmares.

SPEAKER_00

Oh, absolutely. Enterprise customers hate uncertainty. They want to know that their AI partner is going to be around in five years, with the same leadership, the same business model, the same APIs. This lawsuit puts all of that in question.

SPEAKER_01

And think about the employees at OpenAI. You're working on cutting-edge AI research, but there's this constant background noise that your CEO might get kicked out, and the whole company restructured.

SPEAKER_00

That's got to be incredibly distracting. And in a talent market where AI researchers can basically work anywhere they want, that kind of instability could lead to some serious brain drain.

SPEAKER_01

But here's what I find interesting. Despite all this legal drama, OpenAI keeps announcing new partnerships and breaking away from exclusive deals. Either they're incredibly confident they'll win this case, or they're trying to create facts on the ground that make it harder for a court to unwind their business.

SPEAKER_00

That's a smart observation. If OpenAI can establish themselves as critical infrastructure for hundreds of enterprise customers before this case resolves, it becomes much harder for a judge to just shut them down or force a complete restructuring.

SPEAKER_01

The other angle here is what this means for other AI startups. Are we going to see more lawsuits from early investors or founders who feel like they were promised one thing and got something else?

SPEAKER_00

That's a really good point. The AI space has moved so fast that a lot of early agreements and handshake deals might not reflect the current reality of these companies. This could open the floodgates for similar litigation.

SPEAKER_01

Keep watching this case because it could set major precedents for how AI companies can structure themselves and transition from research organizations to commercial enterprises. The entire industry is watching this one.

SPEAKER_00

And honestly, regardless of who wins, this case is probably going to result in much more careful legal structuring for future AI companies. Nobody wants to be the next open AI dealing with this kind of existential lawsuit.

SPEAKER_01

Now early reports suggest we might have a completely new player in the AI space that nobody saw coming. David Silver, this is the guy who led the AlphaGo project at DeepMind, apparently founded a company called Ineffable Intelligence and just raised $1.1 billion. If confirmed, this British AI lab is already valued at $5.1 billion, and their focus is on building AI systems that can learn without human data. So we're talking about a completely different approach from companies like OpenAI that train on massive data sets of human-created content.

SPEAKER_00

Wait, hold up.1 billion dollars valuation for a company we've never heard of? That's approaching open AI territory. And David Silver leaving DeepMind to start this, that's huge. He's literally one of the most respected AI researchers in the world.

SPEAKER_01

Right. This is the guy who figured out how to beat the world champion at Go, which was supposed to be impossible for decades. If anyone can crack the code on AI that learns without human data, it's probably him. So this could potentially solve one of the biggest problems facing the AI industry right now, running out of high-quality training data. Everyone's scraping the internet, but there's only so much human-created content out there.

SPEAKER_00

Exactly. And it also sidesteps all the copyright and legal issues around training data. If your AI is learning through self-play and simulation, you don't need to worry about whether you have permission to use that content.

SPEAKER_01

But I'm a little skeptical about whether this approach can work for general AI tasks. Reinforcement learning works great for games with clear rules and objectives. But how do you apply that to something like writing or creative tasks?

SPEAKER_00

That's the billion-dollar question, literally. My guess is they're focusing on specific domains first. Maybe scientific research, mathematical reasoning, things where you can define clear success criteria and let the AI experiment.

SPEAKER_01

That's not money you throw around lightly.

SPEAKER_00

No, especially not in today's funding environment. VCs are being way more uh cautious about AI investments. So either David Silver has some incredible breakthrough he's not talking about publicly, or the investors see something we don't.

SPEAKER_01

What's interesting to me is the name Ineffable Intelligence. Ineffable means something that can't be described or expressed in words. That's either really pretentious or a hint about what they're building.

SPEAKER_00

But maybe it's it's actually meaningful. If they're building AI that learns through pure experimentation without human language or concepts, maybe ineffable is literally what they're aiming for.

SPEAKER_01

Think about the competitive implications though. If this actually works, it could completely change the game. All the advantage that companies like OpenAI have from their massive training datasets suddenly becomes irrelevant.

SPEAKER_00

And it levels the playing field for smaller companies. Right now you need massive resources to scrape and process internet scale datasets. But if you can train effective AI through self-play, suddenly a small team with good algorithms could compete with the Giants.

SPEAKER_01

Although, let's be realistic about the timeline here. Even if this approach works, it's probably going to take years to develop and scale. AlphaGo is impressive. But it took DeepMind years to go from that to more general applications.

SPEAKER_00

True, but they've got a $5 billion valuation and presumably learned a lot from the DeepMind experience. Plus, the compute infrastructure and AI, AEI tooling available today is way better than when AlphaGo was developed.

SPEAKER_01

I'm also curious about talent. David Silver is brilliant, but building a company this ambitious requires a whole team of world-class researchers. Where are they going to find people who aren't already locked up by Google, OpenAI, or Anthropic?

SPEAKER_00

Well, with a billion dollars in funding, they can afford to make some very attractive offers. And there are probably a lot of researchers who are excited about working on this approach instead of just scaling up language models.

SPEAKER_01

This is definitely one to watch closely. If ineffable intelligence can actually deliver on AI that learns without human data, it could completely change the competitive landscape and solve some major technical and legal challenges facing the industry.

SPEAKER_00

And even if they don't fully succeed, the research they produce along the way could be hugely valuable. Sometimes the most important breakthroughs come from ambitious failures that teach us something new about what's possible.

SPEAKER_01

Alright, let's shift gears and talk about something that could have major implications for Meta's AI strategy. Early reports suggest that China has ordered Meta to unwind its $2 billion acquisition of a company called MANUS after a months-long investigation. Now, I'll admit I'm not super familiar with MANUS, but the reports indicate this deal was central to Meta's AI agent strategy. So we're talking about a major setback for their AI ambitions if this is confirmed. Well Hei.

SPEAKER_00

Yeah, $2 billion isn't pocket change, even for Meta. And the fact that China is blocking this deal tells me there's probably some technology involved that they consider strategically sensitive. China doesn't usually interfere with tech acquisitions unless there's a national security angle.

SPEAKER_01

This is part of a broader pattern we're seeing, where countries are getting much more aggressive about blocking tech deals that might give foreign companies advantages in AI or other critical technologies.

SPEAKER_00

Exactly. And it puts Meta in a really tough position. They're trying to compete with OpenAI and Google in AI, but if they can't make the acquisitions they need to build out their capabilities, they're fighting with one hand tied behind their back.

SPEAKER_01

What's interesting to me is the timing. This comes as Meta is really pushing hard on AI agents and trying to integrate AI across all their platforms. Losing a key acquisition like this could set them back significantly.

SPEAKER_00

And it's not just about the money or the technology. It's about the signal it sends to other potential acquisition targets. If you're a promising AI startup, are you going to want to get involved with Meta if there's a risk that regulatory authorities might block the deal?

SPEAKER_01

That's a really good point. This could have a chilling effect on Meta's ability to do AI deals going forward. And in a space where talent and technology are moving so fast, that's a serious competitive disadvantage.

SPEAKER_00

Right. While Meta is dealing with regulatory headaches, companies like OpenAI are free to make partnerships and acquisitions without worrying about geopolitical interference. That's a huge advantage in a fast-moving market.

SPEAKER_01

I'm also wondering about the strategic implications here. Meta has been positioning itself as this global platform, but if they can't do deals in major markets like China, how global can they really be?

SPEAKER_00

That's the new reality for big tech companies. You can't just think about technology and business strategy anymore. You have to think about geopolitics, national security concerns, regulatory relationships across multiple countries.

SPEAKER_01

And this is happening while Meta is already facing scrutiny from US regulators about antitrust issues. They're getting squeezed from multiple directions, which has to be incredibly frustrating when you're trying to innovate in a fast-moving space.

SPEAKER_00

Maybe they decide it's less risky to build everything in-house rather than deal with regulatory approval processes.

SPEAKER_01

But that's much slower and more expensive. When you're competing with companies that can move fast through partnerships and acquisitions, going the internal development route puts you at a real disadvantage.

SPEAKER_00

True. But what's the alternative? They could try to structure deals differently, maybe licensing technology instead of outright acquisitions, or doing joint ventures that are less likely to trigger regulatory concerns.

SPEAKER_01

We'll definitely need to keep an eye on how this develops and whether it affects Meta's other AI initiatives. But if confirmed, this is a significant setback for one of the biggest players in the space.

SPEAKER_00

Nobody wants to spend months on due diligence and negotiations just to have a deal blocked at the last minute.

SPEAKER_01

First up, early reports suggest Google is testing something called Ask YouTube, basically an AI chatbot search experience for YouTube that lets you search more conversationally.

SPEAKER_00

If they can crack conversational video search, that's a pretty significant moat against competitors.

SPEAKER_01

What's interesting is that this could change how creators optimize their content. Instead of just thinking about keywords, they might need to think about conversational queries and natural language descriptions.

SPEAKER_00

That's a good point. And it could surface a lot of older, high quality content that gets buried by the current algorithm. Sometimes the best tutorial on a topic was made three years ago, but never gets discovered because the search is so limited.

SPEAKER_01

I'm curious how this affects YouTube's monetization, though. If AI can give people direct answers from video content, do they still need to watch the full videos and see the ads?

SPEAKER_00

Now that's the million dollar question. Google needs to balance making search more helpful with maintaining the ad revenue that funds the whole platform. It's going to be a delicate balance.

SPEAKER_01

Next, early reports indicate that over 600 Google employees have signed a letter to CEO Sunda Pichai urging the company to stay out of classified military AI projects with the Pentagon.

SPEAKER_00

This takes me back to the whole Project Maven controversy from a few years ago where Google employees successfully pressured the company to drop a military AI contract. Seems like there's still a strong faction within Google that wants to keep military applications off limits.

SPEAKER_01

But here's the tension. If Google doesn't work with the military, other companies will. And some of those companies might not have the same ethical standards or employee oversight that Google does.

SPEAKER_00

That's the classic dilemma. Do you engage and try to influence from the inside, or do you stay pure and let someone else potentially do a worse job? There's no easy answer, but 600 employees signing a letter suggests this is a major internal debate at Google.

SPEAKER_01

What strikes me is that this is happening while China is heavily investing in military AI and other countries are racing to build these capabilities. Google's employees want them to opt out, but the world isn't opting out.

SPEAKER_00

If Google won't work with the Pentagon, that creates opportunities for Microsoft, Amazon, and other cloud providers who are less squeamish about defense contracts.

SPEAKER_01

I also wonder about the definition of classified military AI use. Are we talking about weapon systems or intelligence analysis or logistics optimization? The ethical implications are very different depending on the specific application.

SPEAKER_00

That's a crucial distinction. Helping the military optimize supply chains is very different from building autonomous weapons. But the letter apparently focuses on classified projects, so employees might not even know what they're objecting to specifically.

SPEAKER_01

Going back to OpenAI, there are reports about them securing concessions from Microsoft that resolve legal concerns over a $50 billion Amazon deal. This seems connected to the partnership changes we talked about earlier.

SPEAKER_00

Yeah, this is probably the behind-the-scenes negotiation that made the partnership restructuring possible. Microsoft was probably threatening legal action to block OpenAI from working with Amazon. So they had to work out some kind of compromise.

SPEAKER_01

A $50 billion deal with Amazon would be absolutely massive. That's got to be about cloud infrastructure and maybe enterprise sales partnerships. That's the kind of deal that could really change the competitive landscape.

SPEAKER_00

And it shows how much these companies are willing to invest to secure AI partnerships. When you're talking about $50 billion, that's not just buying compute, that's buying strategic positioning for the next decade.

SPEAKER_01

What's interesting is that this deal was apparently creating legal peril for OpenAI. That suggests Microsoft's exclusive partnership agreement had some pretty serious enforcement mechanisms built in.

SPEAKER_00

Right. Microsoft wasn't going to just let their biggest AI partner walk away without a fight. The fact that they worked out a compromise suggests both sides really wanted to avoid a messy legal battle that could hurt everyone.

SPEAKER_01

And from Amazon's perspective, 50 billion is a massive bet on AI infrastructure. They're clearly convinced that AI compute is going to be a huge business for the next decade.

SPEAKER_00

It also positions AWS as a real competitor to Microsoft's cloud offerings for AI workloads. If you're building an AI application, now you have multiple major cloud providers competing for your business with OpenAI models.

SPEAKER_01

Finally, OpenAI and Microsoft have put out official statements about their amended partnership, saying it streamlines their collaboration and provides long-term stability for continued AI innovation.

SPEAKER_00

But honestly, both companies probably benefit from having more flexibility.

SPEAKER_01

It's smart PR. Instead of letting the media frame this as a messy divorce, they're positioning it as a mature partnership evolution. But the reality is still that Microsoft lost exclusive access to OpenAI's models.

SPEAKER_00

Right. But they probably saw the writing on the wall. OpenAI was always going to want more independence eventually. Better to negotiate a favorable transition than fight it and lose everything.

SPEAKER_01

The phrase long-term stability is interesting, though. Both companies are probably trying to reassure enterprise customers that this change won't disrupt existing services or create integration problems.

SPEAKER_00

Absolutely. The last thing they want is for Fortune 500 companies to pause their AI deployments because they're worried about platform instability. That would hurt both companies' revenue.

SPEAKER_01

And Streamlined collaboration suggests they're actually going to work together more efficiently now, which seems optimistic. Usually when you end exclusive deals, the relationship gets more complicated, not simpler.

SPEAKER_00

Maybe they mean it removes some of the conflict of interest issues. Like Microsoft can now compete with OpenAI in some areas while still partnering in others without the weirdness of being both partner and competitor under an exclusive deal.

SPEAKER_01

Open AI is breaking free from Microsoft. New players like ineffable intelligence are emerging with completely different approaches, and regulatory authorities are getting much more involved in AI deals.

SPEAKER_00

And honestly, that's probably better for innovation.

SPEAKER_01

But it's also creating a lot of uncertainty. If you're a business trying to build on AI platforms, or a developer choosing which tools to invest in, all these partnership changes and legal battles make it really hard to plan for the future. The wild card is regulation. China blocking Meta's deal, employees pushing back on military contracts, courts potentially restructuring entire companies. There are a lot of external forces shaping this industry now.

SPEAKER_00

Which means the companies that can navigate both the technical challenges and the regulatory landscape are going to have a huge advantage. Pure technical excellence isn't enough anymore.

SPEAKER_01

And think about the money involved here. We're talking about a $1.1 billion raise for a company most people have never heard of, $50 billion deals between major tech companies, $2 billion acquisitions, getting blocked by foreign governments. The scale of investment and the stakes involved are just massive.

SPEAKER_00

That tells you how much these companies believe AI is gonna reshape every industry. They're not making billion-dollar bets on incremental improvements. They think this is a fundamental shift in how business gets done.

SPEAKER_01

What's interesting is how different the strategies are becoming. OpenAI is going for maximum distribution and partnerships. Ineffable Intelligence is betting on a completely different technical approach. Meta is trying to build everything in-house. Google is integrating AI into their existing products.

SPEAKER_00

And that diversity of approaches is actually really healthy for the industry. We don't know which strategy will work best. So having multiple companies trying different approaches increases the chances that we'll see real breakthroughs.

SPEAKER_01

The geopolitical angle is getting more important too. It's not just about building the best technology anymore. It's about building technology that can navigate regulatory approval in multiple countries, satisfy national security concerns, and deal with trade restrictions.

SPEAKER_00

Right. And that's probably going to lead to more regionalized AI development. Maybe we end up with different AI ecosystems for different parts of the world rather than truly global platforms.

SPEAKER_01

For developers and businesses, I think the key takeaway is to stay flexible and avoid getting too locked into any single platform or partnership. The landscape is changing too fast to make big long-term bets on specific technologies.

SPEAKER_00

Exactly. Focus on building skills and capabilities that can adapt to different platforms, rather than becoming an expert in one specific tool that might not be around in two years.

SPEAKER_01

And if you're an investor or entrepreneur in this space, pay attention to the regulatory and geopolitical trends as much as the technical ones. The best AI technology in the world doesn't matter if you can't get regulatory approval to deploy it. That's a wrap on what might be the most consequential day for AI partnerships we've seen all year. If you found this useful, definitely hit subscribe because this stuff is changing daily.