697: The Chart Is Terrifying

What It Covers

Apple announces across-the-board price increases of 10–60% on Macs, iPads, and Vision Pro, driven by an unprecedented DRAM and NAND storage crisis fueled by AI data center demand. RAM prices are projected to reach 900% above 2023 levels by 2027. The episode also covers Apple's PCC server strategy, Tesla CarPlay developments, and new US sunscreen ingredient approval.

Key Questions Answered

• •RAM Crisis Timeline: DRAM prices are projected to hit 900% above 2023 baseline levels by end of 2027, according to Wall Street Journal data. This is not a short-term disruption — new fabrication capacity takes three to five years to come online, meaning 2026 and 2027 are forecast to be worse than 2025. Consumers and businesses should plan purchasing decisions around a multi-year elevated pricing environment, not a near-term correction.
• •Apple Price Increase Specifics: Average Mac and iPad price increases land at $258 and 21.5% across the product line. The Apple TV jumped 54% from $129 to $200 despite no hardware update in 1,346 days. BTO RAM and SSD upgrades took the hardest hits — an eight-terabyte MacBook Pro SSD option now costs $3,000, and a maxed-out MacBook Pro configuration crossed $10,000, up from roughly $7,200 previously.
• •Buy Now or Wait Strategy: Three ATP listeners saved $400–$800 by purchasing MacBooks in May 2025 based on earlier price-increase warnings. For anyone currently planning a Mac or iPad purchase, prices are unlikely to decrease — Apple historically does not reverse price increases even when component costs normalize. Buying refurbished or used Apple Silicon hardware now, before secondary market prices adjust upward, represents a practical cost-mitigation approach.
• •AI Memory Conflict for Apple: Apple's on-device AI strategy requires more RAM per device with each generation, yet RAM is the exact component experiencing the most severe price inflation due to AI data center demand. The m5 generation devices are expected to ship with higher RAM configurations to run Apple's Foundation Models locally. This creates a compounding cost problem where Apple's own AI ambitions directly worsen the component pricing pressure affecting its hardware margins.
• •Apple PCC Server Viability: Apple built custom Apple Silicon servers running a proprietary CloudOS for Private Cloud Compute, but those servers use M2 Ultra chips, now multiple generations behind current NVIDIA data center hardware. Apple has since partnered with Google to run PCC workloads on Google infrastructure. The custom server program is unlikely to remain viable long-term as Google, AWS, and Microsoft operate purpose-built AI data centers with current-generation NVIDIA GPUs at scale Apple cannot match.
• •Tesla CarPlay Route Sharing: Apple introduced a new CarPlay API called Route Sharing at WWDC 2026, which passes trip data as geographic coordinate arrays to the vehicle. This directly addresses the previously reported technical blocker preventing Tesla from supporting CarPlay — Tesla's supervised self-driving system required knowledge of the intended navigation route to execute automatic lane changes. If Tesla implements this, it could pressure other CarPlay holdouts like Rivian and GM to follow.
• •AI Chatbot Memory Reliability: Current AI memory features across ChatGPT, Siri, and similar tools store remembered facts in background text files appended to system prompts. These memory stores have no guaranteed long-term persistence, can be truncated, compacted, or lost across updates, and should not be trusted for durable information storage. For anything requiring reliable recall — locations of stored items, preferences, instructions — use a dedicated notes app like Apple Notes or Bear, and explicitly instruct the AI agent to write to that app.