Did Apple Get AI Spending Right?, Microsoft & OpenAI’s New Reality, Where’s Stargate?

Episode Summary

Executive Summary: The episode centers on Apple’s unusually low AI CapEx, arguing it may be either a costly miss or a shrewd wait-and-see bet depending on how AI evolves. MG Siegler and the host also unpack Microsoft and OpenAI’s revised relationship, which removes the AGI clause and opens OpenAI to other clouds, and they conclude that Stargate has shifted from a grand buildout narrative to a looser umbrella for infrastructure deals.

Main Topics: Apple’s AI CapEx divergence (Priority: 5/5): Apple is spending far less on AI infrastructure than peers like Microsoft, Google, Amazon, and Meta. The discussion weighs whether Apple’s restraint reflects strategic discipline, lack of ambition, or a smart bet that AI will commoditize before Apple needs to commit heavily. Apple’s long-term AI strategy and optionality under new leadership (Priority: 5/5): The arrival of John Ternus is framed as a possible turning point: Apple may loosen its cash discipline, spend more on R&D and CapEx, and make acquisitions to catch up in AI while still leveraging its hardware advantage. Microsoft–OpenAI deal reset (Priority: 5/5): The new agreement removes the AGI clause and allows OpenAI to run models on any cloud provider, while Microsoft retains important first-look and commercial protections. This is treated as a major but incomplete loosening of the partnership. Stargate’s evolution into a broader infrastructure umbrella (Priority: 4/5): What began as a headline-grabbing $500 billion buildout is described as increasingly looking like a catch-all term for many separate infrastructure arrangements, with OpenAI relying more on partners than building everything itself. AI infrastructure economics and market dynamics (Priority: 4/5): The conversation contrasts the massive spend of hyperscalers and model companies with the uncertain returns, raising questions about whether any single firm can win by spending more, or whether open source and on-device AI will blunt the advantage of giant infrastructure investments. Future device platform battle (Priority: 4/5): The speakers debate whether AI will consolidate around a single dominant interface, potentially threatening Apple if it doesn’t control the models, or whether the iPhone and on-device AI will remain the central computing hub.

Key Arguments: Apple may be making a rational bet by avoiding the AI CapEx arms race because frontier models and compute could commoditize, reducing the value of massive upfront spending. Apple’s hardware ecosystem, especially the iPhone, could still become the primary AI delivery platform even if Apple doesn’t own the best models. The biggest risk to Apple is that AI becomes the core interface for computing, leaving Apple dependent on external model providers and weakening its control over the user experience. Microsoft and OpenAI’s updated deal suggests both sides wanted more separation: OpenAI gains cloud flexibility, while Microsoft protects its position through commercial rights and first access. The removal of the AGI clause is symbolically important but not the whole story; Microsoft still has significant leverage through Azure-first provisions and revenue-sharing terms. Stargate appears less like a single grand project and more like a marketing and coordination layer over multiple partner-led data center deals. OpenAI’s shift toward partner-owned infrastructure offloads financial risk onto Oracle, SoftBank, NVIDIA, and neo-clouds, making those companies more exposed to OpenAI’s growth and cash flow. Open-source models remain an important counterforce, but they may be falling behind the frontier faster than many expected, limiting the likelihood that Apple can simply wait for commoditization without consequences.

Data Points: Apple 2025 CapEx: $9–10 billion - Projected Apple capital expenditures for the year, contrasted with far larger AI infrastructure spending by peers. Meta CapEx expectation: $145 billion - Used as a comparison point to show the scale gap between Apple and AI-heavy peers. Amazon and Google CapEx guidance: ~$190 billion each - Referenced as the latest guidance indicating continued massive infrastructure investment. Microsoft CapEx: Comparable to Amazon and Google at the high end - Described as spending at roughly the same scale, though exact figure was not specified in the transcript. OpenAI–Microsoft access window: Until 2032 - Microsoft retains access to OpenAI IP through this year under the revised arrangement. OpenAI revenue-share change: Microsoft no longer pays revenue share to OpenAI - The revised agreement removes Microsoft’s obligation to share revenue back to OpenAI. OpenAI revenue-share obligation: OpenAI still pays revenue share to Microsoft - Even when using AWS or potentially other clouds, OpenAI remains financially tied to Microsoft. Stargate headline commitment: $500 billion - Initial public framing of the infrastructure initiative announced with major political fanfare. Initial Stargate funding commitment: $100 billion - Described as the portion initially expected to be committed out of the gates. Apple buybacks/dividend policy: Potentially slower under Ternus - Bloomberg reporting suggested slower or less frequent buybacks and dividend hikes may be considered. Core AI adoption stat: 95% of AI initiatives fail - Used in sponsor copy, not as a substantive argument in the discussion.

Pivotal Quotes: "Everyone thinks AI is going to be a core bit of technology that they each need to control, lest they be beholden to someone else." — MG Siegler: Explaining why companies like Microsoft are spending aggressively to avoid dependence on outside AI providers. "The iPhone is probably going to be the central computing hub, and maybe even more so." — MG Siegler: Arguing that Apple may still win the AI era because the iPhone is already the dominant device in consumers’ hands. "This is a big moment. Apple has a famous policy in place to try to be, you know, cash neutral... They are sending a signal." — Alex Kantrowitz: Discussing the Ternus-era financial shift that could allow more AI-related spending and acquisitions.

Implications: Apple may be setting up for a delayed but larger AI push, while OpenAI and Microsoft are redefining their partnership to maximize flexibility. For the industry, AI leadership now looks less like a single model race and more like a battle over infrastructure, distribution, and device control.