Kuo Calls Apple Chip Stockpile Rumor Unlikely Given Tight TSMC Planning Cycle

A billion dollars’ worth of unfinished Apple chips, supposedly sitting in a TSMC warehouse because nobody could find enough memory to finish them. That’s the rumor that spread through supply chain circles this week. According to analyst Ming-Chi Kuo, it doesn’t hold up.

Kuo, who covers Apple’s supply chain for TF International Securities, addressed the claim directly on August 10. The story, as it circulated online, held that TSMC was sitting on 2-nanometer processor wafers built for Apple, worth roughly $1 billion, that couldn’t be packaged into finished chips because of a shortage of DRAM.

What the Apple chip stockpile rumor actually claimed

Apple’s newer Silicon chips pair the processor with memory in the same package, so a shortage on the memory side can, in theory, stall the whole assembly line even if the processors themselves are ready. The rumor leaned on that logic, plus one piece of real evidence: TSMC’s Q2 2026 earnings call, where executives mentioned rising “inventory days” tied to the ramp-up of 2-nanometer production.

That single data point was enough for the theory to take hold. A billion-dollar number is the kind of thing that travels fast, especially when it’s layered onto a supply crunch people already know is real.

Why Kuo isn’t buying it

Kuo’s pushback comes down to how Apple and TSMC actually plan their orders. Apple books its processor volumes roughly three months ahead of time, and those orders are sized against the memory Apple already expects to have on hand. That’s the mechanism that would need to fail for a billion-dollar stockpile to exist in the first place.

“It’s hard for such dramatic developments to occur,” Kuo wrote, framing the scenario as one that would require a serious breakdown in a relationship built specifically to avoid that kind of mismatch.

He’s not denying that Apple has a memory problem. Kuo says his own research shows Apple’s 2026 hardware shipment forecasts have already been downgraded because of the DRAM shortage. What he rejects is the idea that the fallout looks like a billion dollars in half-finished chips waiting on a shelf.

Reading TSMC’s earnings call correctly

Kuo’s strongest argument is about what “inventory days” actually measures. TSMC’s number covers finished goods, raw materials, spare parts, and work-in-progress items together, not a single line item for stalled Apple chips. Reading one broad accounting metric as proof of one specific customer’s supply problem skips several steps.

There’s also a pattern here that isn’t new. TSMC’s inventory days tend to climb whenever it ramps a new process node, simply because early production runs carry more buffer stock while yields stabilize. The 2-nanometer node is in that ramp-up phase now, which alone explains part of the shift Kuo describes.

And Apple isn’t TSMC’s only 2-nanometer customer. AMD and MediaTek are both in the mix too, so a genuine packaging bottleneck tied to memory shortages would show up as a shared problem across TSMC’s client list, not something isolated to Apple.

What this means going forward

None of this erases the memory crisis itself. Apple has reportedly been exploring supply sources that come with their own political complications, a sign that the DRAM shortage is squeezing the company enough to consider workarounds it might otherwise avoid. The shortage is real, and it’s already affecting how much hardware Apple can ship this year.

What Kuo’s response does is draw a line between a supply chain under real strain and a specific viral number in Apple chip stockpile rumor that doesn’t match how Apple and TSMC actually operate together. The two companies have spent years building an ordering process precisely to prevent this scenario, and a three-month planning cycle tied to actual memory availability doesn’t leave much room for a billion dollars in orphaned wafers to pile up unnoticed.