"An Interview with Ben Bajarin About Apple, AI, and Compute" — @benthompson
Interview with Ben Bajarin, CEO/Principal Analyst at Creative Strategies (semiconductor-focused market research, co-host of The Circuit podcast). Third Bajarin appearance (prior: Sep 2024, Oct 2025). Covers WWDC 2026, Apple's cloud-AI architecture, Nvidia's PC chip, agentic CPUs in the data center, Intel's reversal, and the compute capacity shortage.
Why this is in the vault
Direct feed for two active RDCO threads: the chip-fab/memory capital-cycle thesis (this is one of the densest supply-demand discussions Stratechery has run this year) and hyperscaler compute economics (a genuinely new claim about why Google/Amazon are shifting internal workloads off their own ASICs). Pulled from the full email body via Gmail; no web reconstruction needed.
The core argument
WWDC = platform re-architecture, not features. Bajarin's read: Apple spent the keynote on "boring" plumbing (CPU scheduler, a rebuilt system index) because the 2024 Apple Intelligence promises required re-architecting the platform first. Thompson's framing: Apple is "perfecting a 2024 AI story" — the Siri demos that work are fancy search, and that is what consumers actually want.
Consumers won't pay for agents. Thompson's recurring thesis, restated hard: enterprises pay for productivity, consumers pay for entertainment (Netflix, not Dropbox Carousel; OpenAI as "the modern Dropbox"). Bajarin partially dissents via Jobs to Be Done — Siri as a consumer "control plane" that hires out small outcomes — but both agree the endgame is human-in-the-loop, not autonomous consumer agents. Both think Apple shipping good-enough on-device AI makes OpenAI's device-level consumer ambitions "very, very difficult, if not nil."
Apple's cloud architecture: Gemini base + Google Cloud + Nvidia. On-device Apple Foundation Models are Apple's own; the cloud Pro model is base Gemini with Apple post-training. Inference runs on Nvidia GPUs (both lean "largely H100s — nothing Apple does needs a GB200") in Google Cloud, with Intel as confidential-compute head node. Thompson's take: Apple Silicon in the cloud is effectively dead; Nvidia-on-GCP preserves portability to AWS/Azure, whereas TPUs would lock Apple to Google. The ~$1B/yr Gemini deal makes sense once you see Google landed a massive GCP customer.
The ASIC lock-in mechanic (the standout claim). Bajarin, from Creative Strategies' customer-chain work: Google is deliberately moving internal workloads onto GPUs (including SpaceX/neocloud capacity) so it can offer scarce TPU capacity to third parties at priority pricing — because once a customer's stack runs on TPUs (or Trainium, where Amazon mirrors the play), "you're stuck." The capacity crunch is being used as a customer-acquisition weapon for custom silicon. Thompson called it mind-blowing; it reframes the Google–SpaceX GPU deal as TPU go-to-market, not GPU shortage relief.
Agentic inference brings back the CPU. "Humans click, agents swarm" (Jeetu Patel, Cisco): agent orchestration is CPU work, so expect dedicated CPU racks beside GPU racks (Nvidia Vera, Arm selling chips, AMD/Qualcomm following). Concurrency-per-megawatt ("cores per megawatt") becomes the metric. Bajarin: CPU inference is already happening more than people realize because GPUs are consumed by training. Software moats (maybe even CUDA) erode when agents can rewrite your software cheaply.
The capacity shortage was seeable and everyone missed it. TSMC slowed capex growth post-ChatGPT; memory makers underbuilt for years and new capacity takes ~2 years to arrive. The industry's overcapacity scars beat the Samsung/Morris Chang lesson (the two most valuable fab franchises were built by investing into downturns). Bajarin's estimate: TSMC could have used "five more foundries" and still not met demand. Consequence: TSMC created its own competition — Intel will get foundry customers ("a matter of when, not if"), and Intel's EMIB advanced packaging (mix-and-match tiles, capacity where CoWoS is rationed) is becoming a multi-billion business. Supply-demand does not balance "before 2030," possibly 2035.
Is AI software or graphics? Bajarin's framing for the cycle-length question: if AI is like internet/client software, it gets "good enough" and infrastructure spend plateaus (bearish). If it's like graphics — an industry that has never had enough compute, chasing fidelity for 30 years, "tokens and pixels are the same thing" (Jensen's insight) — the buildout runs decades. Bajarin leans graphics.
Two notable bear-case asides from Thompson: (1) memory — 40 years of never optimizing for memory means huge low-hanging fruit for demand-side optimization "to the detriment of the memory providers"; (2) hyperscaler-as-meter dominance could leak if token generation migrates to the edge/on-prem (OpEx vs CapEx logic for enterprises with token budgets).
Sponsor/bias scan
No third-party sponsors; Stratechery is subscription-funded. Standard podcast/Stratechery Plus plugs in footer — not structurally biasing. Perspective note: the closing segment is a friendly promo for The Diligence Stack, Creative Strategies' new $300/mo research Substack. No commercial relationship with Stratechery disclosed, so not a sponsor flag — but Bajarin's authority claims ("vast number of conversations with CIOs," "we hear in the customer chain") are also the sales pitch for his paid product. His customer-chain claims (e.g., the TPU lock-in mechanic) are plausible, sourced from non-public client work, and unverifiable from here — treat as informed analyst signal, not confirmed fact.
Mapping against Ray Data Co
- Chip-fab/memory capital-cycle thesis (strong, corroborating + one new risk). The capacity discussion supports the founder's Phase 2 placement: underbuilt supply, ~2-year lag for memory capacity, "not balanced before 2030." Samsung/Morris Chang invest-through-downturn history is the pattern the thesis rides. New bear-case input to log against the memory leg: Thompson's demand-side optimization point — astronomical token/memory prices create incentives to finally optimize memory usage, which could blunt memory-maker pricing power before new capacity lands. Doesn't change the phase read (supply stays tight either way near-term), but belongs in the thesis risk register. The software-vs-graphics question is effectively the thesis's terminal-value variable: graphics-like demand = longer Phase 2/3 runway.
- Hyperscaler capex tracking (strong). The TPU/Trainium lock-in mechanic gives the EDGAR capex pulse a new interpretive lens: hyperscaler capex + third-party GPU deals (SpaceX/neoclouds) may be funding ASIC go-to-market, not just internal capacity. Worth carrying into the next /investing-edgar-watch brief. Also: watch CPU-rack buildout (Vera, Axion, Arm) as a new capex line item distinct from GPU spend.
- Fable 5 access-cliff (medium, contextual). The interview's consumer-vs-enterprise split is the strategic logic behind Anthropic's June 23 move: agentic inference is metered enterprise/pro compute, not a subscription consumer good (Thompson: "I'm not going to pay Anthropic $50 per million tokens" for a kitchen agent). RDCO — an always-on autonomous COO agent — is precisely the swarm-class workload the metering model targets. Confirms the framing in the Fable 5 tiers note; no new facts that change the decision.
- Squarely / Apple distribution (weak, honestly). No App Store/distribution content. Tangentially: the rebuilt system index + contextual Siri could eventually change how users find content in apps, and Apple's on-device AFM models are a developer surface Squarely could someday use — but nothing actionable here.
- phData consulting (medium-weak). "CIO token budgets" and edge token generation as cost optimization is a real enterprise conversation per Bajarin — useful vocabulary for phData client work on AI cost architecture, nothing more specific.
Related
- [[2026-06-10-stratechery-fable-5-anthropic-alignment-ai-tiers]] — the access-cliff decision this interview's consumer/enterprise split contextualizes
- [[2026-06-08-stratechery-nvidia-vs-tpu-broadcom-miss]] — Nvidia-vs-ASIC framing that the TPU lock-in mechanic extends
- [[2026-01-26-stratechery-tsmc-risk]] — the "TSMC Brake" piece Thompson cites as his prior on underbuilt capacity
- [[2026-03-17-stratechery-interview-jensen-huang-nvidia]] — Jensen's tokens-as-pixels worldview, source of the "AI is graphics" frame