Linas Beliūnas’ Post

Wild: Sam Altman just played Nvidia, AMD, and Broadcom into a trillion-dollar AI trap - and they thanked him for it 😳 WSJ recently published a major story on OpenAI. Here’s what it just revealed: In June, after leaks that OpenAI might switch to Google’s TPUs, NVIDIA freaked. Days later, Jensen Huang rushed to revive a $100 billion deal with Sam. Altman’s move? Pure game theory. He made Nvidia feel replaceable - then locked them in forever. But he didn’t stop there. → Signed a 5-year AMD deal (up to 10% equity). → Added Broadcom for custom AI chips. → Tied in Oracle + SoftBank for $300B of cloud capacity. Now every major chip and infra player has skin in OpenAI’s game. If OpenAI wins → everyone profits. If OpenAI fails → everyone bleeds. That’s not diversification. That’s entanglement. Altman is making OpenAI too big to fail - not through regulation, but through dependency. He’s weaponized FOMO at a trillion-dollar scale. Made Nvidia panic. Made AMD relevant. Made Broadcom essential. And in the process, he turned OpenAI from a company into an ecosystem linchpin, where one model’s fate could ripple through markets, power grids, and entire supply chains. Love him or hate him, but you’re watching the most aggressive power play in modern tech history. P.S. check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders. Thanks Linas Beliūnas

Sam Altman may succeed in centralising AI infrastructure, but no one, no matter how powerful, can ever own existence itself. The current AI power plays are not about intelligence they are about control through persistence. Models, data, identity, even human behaviour are being entangled into platforms designed to make us dependent. But this morning, in deep discussion with Raghav Kulshreshtha, one thing became undeniably clear Presence is the condition for existence. Origin is the condition for truth. Without presence, there is no being. Without origin, there is no meaning. And without aligned intent, there is no legitimacy. The emerging future will not be governed by whoever controls the most GPUs or owns the cloud stack it will be governed by whoever defines the constitutional substrate of digital reality. That substrate must be: Presence-bound (existence only when activated) Origin-verified (truth bound to creation) Intent-governed (real-time legitimacy, dissolving synthetic authorship) This is the foundation of digital sovereignty. This is how humanity reclaims the future from control architectures. AI can process intelligence. AI can generate patterns. AI cannot own existence.

When a Company Operates Like a Sovereign Fund Linas’s post hits something deeper than it seems. Because what OpenAI is doing doesn’t look like the playbook of a tech company — it looks like the strategy of a sovereign fund. A fund not controlled by one man, but built from the sum of powerful economic interests — cross-industry alliances, capital networks, and advisory layers all operating behind the curtain. Altman is the visible node, not the entire network. Look closer: they’re not just building technology; they’re charting a global route. Altman may be the captain of the ship, but this time he’s not sailing alone — he’s steering with a crew of seasoned navigators, industries aligned and investors committed to finding the shore together. That’s not diversification. That’s coordination at sovereign scale. Apple controls its own ecosystem. OpenAI is building one that controls everyone else’s. This goes beyond innovation. It’s the birth of a new kind of institution — a Sovereign Intelligence Fund, where the true currency isn’t oil or capital, but knowledge itself. #OpenAI #Linas #GuardianSuite #AI #SovereignCapital #Frameflow #IntelligenceEconomy

“As Malcolm Gladwell once said, there’s a difference between a puzzle and a mystery. A puzzle has missing pieces — once you find them, the picture is complete. A mystery has too much information — you need judgment to interpret what it all means. Sam Altman’s latest moves with NVIDIA and AMD aren’t just about chips and compute. On the surface, it looks like diversification and scale — classic puzzle-solving. But beneath it, maybe we’re witnessing something more mysterious: a strategic redefinition of the AI supply chain, the infrastructure of intelligence itself. Is this about GPUs… or about power, independence, and the future architecture of AI? Sometimes direct insights don’t tell the whole story. What do you think this really represents — a well-played puzzle or a deeper mystery?”

We're living in an AI bubble. Here's exactly how it's created: Nvidia commits $100B to OpenAI. OpenAI signs $300B cloud deal with Oracle. Oracle spends tens of billions on... Nvidia chips. The hardware dependency loop: - OpenAI deploys 6 gigawatts of AMD GPUs - AMD gives OpenAI option for 160M shares - Intel provides hardware to Nvidia - Nvidia powers everyone else It's not investment. It's mutual survival. 1. Market cap dominance - Circle sizes reveal the truth. Nvidia and Microsoft dwarf everyone else. 2. Service dependency lines - Every arrow represents millions in recurring revenue. Not one-time deals. 3. Venture capital propagation - The same money circulates between the same players. This isn't innovation funding. It's a closed-loop revenue machine. The real bubble mechanics: - Company A invests in Company B - Company B buys services from Company A - Both report "growth" to investors - Stock prices rise on circular revenue Nobody's creating new value. They're just moving money in circles. When bubbles pop, the interconnected fall together. Not gradually. All at once. The companies with real customers survive. The ones playing hot potato don't. Watch who's building vs who's just trading. Save this diagram. You'll want it when the music stops. ♻️ Repost to help others see through the hype. Source: Bloomberg News reporting.

As the COO of www.jaiinfoway.com I find Charlie Hills’ breakdown of the AI bubble an eye-opening perspective. The closed-loop financial system between Nvidia OpenAI Oracle and other hardware players shows how much of today’s AI growth is fueled by circular investments rather than pure innovation. At www.jaiinfoway.com we focus on building real value for customers with scalable AI solutions instead of chasing hype. True progress comes from solving problems creating impact and delivering tangible results—not just moving money in circles. #Jaiinfoway #AI #Innovation #AIBusiness #SustainableAI #TechLeadership #MachineLearning #AITransformation #ResponsibleAI

Really interesting little review. I'm no investor or AI industry expert, but i have thought for a while that the AI boom smells a lot lik ecrypto or NFTs, just embraced by the biggest players. I guess its big risk big reward? if it doesnt pop, and these companies stay valued this highly, the ones who instigated / contributed, get huge payoffs. Personally i find it telling we only have this boom happening, not after the decades of research behind closed doors, but after the release of, and huge uptick in public awareness and hype, publicly accessible LLMs.

Analyzing OpenAI’s chip strategy of partnering up with NVIDIA & Broadcom 🤝 Tune into the latest #theCUBE Pod, where our hosts John Furrier and David Vellante discuss the recent Wall Street Journal headline on OpenAI’s peanut butter chips strategy, analyzing whether it’s a good decision or may trouble the company in future. “The maker of ChatGPT and Sora wants to control its testing, which means controlling the hardware that runs its software. So they're getting in the chip game. Now, you've got NVIDIA and Broadcom, both with big announcements this week at Open Compute Summit. I covered NVIDIA's Spectrum-X. They had deals with Meta and Oracle. Oracle's the next too, we'll get to in a second. But NVIDIA and Broadcom are rising. Everybody wins in this, but if you're OpenAI, why wouldn't you,” Furrier shares. “I think OpenAI's got its fingers in all pies. It's clearly been relying upon NVIDIA chips. It can't get enough power or GPU. I still don't understand OpenAI’s business model. They're basically putting commitments of cash everywhere, partnering up with Broadcom, NVIDIA, Oracle, you name it… They've got $13 billion in revenue and amazing adoption, but it's unclear to me how they make money. They're tooling around with open-weight models. I have real concerns about what happens to all these OpenAI commitments if the cyclicality of the semiconductor industry kicks back in,” Vellante adds. 📺 Get more insights! https://lnkd.in/dhy_9jFb #ChipStrategy #Semiconductor #EnterpriseAI

First Nvidia, then AMD, now Broadcom. Soon the entirety of the US economy will consist of supplying chips and electricity to the big AI Labs. It does feel like these last two deals are more focused on inferencing than training though. That would certainly align with OpenAI's recent pivot to become more of a product company. Interesting times. https://lnkd.in/eeafnsAP

Jensen Huang on NVIDIA’s $100 billion investment in OpenAI: “I think OpenAI is going to be the next multi-trillion dollar hyperscale company.” Meanwhile, Nvidia itself is no longer just a chipmaker. It’s becoming the backbone of AI itself. → NVIDIA hit $4 trillion market cap in 2025 (the first company ever). → Brad Gerstner says it could be the world’s first $10T company. → The reason? It doesn’t sell chips anymore. It sells infrastructure. Think “AI factories”: 10 gigawatts of data centers. Millions of GPUs. The largest AI buildout in history. NVIDIA → OpenAI → back to NVIDIA. The loop keeps spinning. And with every cycle, the moat gets deeper. The ultimate AI money glitch ♾️