IPOs

Blip 2.0++, OpenAI IPO now 2027, ‘RAMAgeddon’ reaches Apple. AI-RTZ #1130

Web Desk June 27, 2026

0 2 5 minutes read

Blip 2.0++, OpenAI IPO now 2027, ‘RAMAgeddon’ reaches Apple. AI-RTZ #1130

Apple recognizes RAMageddon: The memory price hikes officially arrived this week, and it’s Apple that made them feel real. Media is dubbing it Apple’s ‘RAMageddon’— raising Macs, iPads, Apple TVs, Vision Pros and displays 15-25%+ to counter the memory shortage (the value MacBook Neo quietly went from $599 to $699). Even as it held off on base iPhones for now. The market took note. As outgoing CEO Tim Cook put it, it’s a “100-year flood” he hasn’t seen in four decades in the business. Apple is also adjusting their AI-ready chip release schedule for this year and next. My take: on a relative basis Apple still has the most resilient supply chain at scale — the best seat in the pouring rain versus peers without its wide, deep, global hardware/software supply chain. But it’s a relative win, not an absolute one: even Apple is passing the costs through, with little near-term relief for anyone. And because it’s Apple’s prices moving — not a Windows laptop or an Xbox — regulators and economists might actually notice.. More here and here.

US’s Memory Supply Option in China: There’s no easy respite from the global memory supply bottleneck and price hikes. There may be a somewhat difficult relief valve for the memory crunch — it just runs through China, and US companies like Apple haven’t been able to tap it. As the WSJ frames it, the crunch is “almost impossible to solve” quickly: the three memory titans (SK Hynix, Samsung, Micron) have swung wafer capacity to high-margin HBM for AI data centers, Micron’s new Idaho and New York fabs won’t open until mid-2027 and 2030, and national-security rules block American firms from working with China’s fast-rising CXMT (DRAM) and YMTC (NAND) — with Micron itself lobbying to keep that door shut. Apple’s Cook says “everything needs to be on the table,” and HP is reportedly exploring CXMT chips for Asia-bound products. My take: water, water everywhere, and not a drop to drink — the one near-term supply valve is capped by geopolitics, and it may take Xi’s White House visit this September and a broader tech deal to change that. Until then, the world just has to wait for the chip giants to make more, like regular potato chips. More here.

Microsoft sides against Frontier AI Models: Picking up the thread from ARD #98, Microsoft CEO Satya Nadella has now drawn the line in ink — “We can’t let AI giants eat the economy” — putting Microsoft firmly on the side of enterprise customers rather than the frontier model-builders. In a matter of weeks Microsoft rolled out a suite of low-cost models and Copilot Cowork (an agent that lets users pick cheaper models for long-running tasks), and is even weighing hosting China’s open-source DeepSeek — a striking move from one of OpenAI’s oldest backers. Nadella reframes AI as job reorganization over elimination, and says the industry still has to “earn the social permission.” My take: this is the Frenemies dynamic in full color — Microsoft wants to be far less dependent on a handful of frontier model companies (beyond Nvidia too), and it’s carving out an ‘us vs them’ position with the customers right ahead of the OpenAI and Anthropic mega-IPOs. A mechanism that’s worked in power politics for millennia. More here.

OpenAI & Anthropic vs SpaceX IPO Clock: OpenAI is considering pushing its mega-AI IPO back to 2027. And its new GPT 5.6, being rolled out in 3 versions, is being gated by the White House in terms of selective release to approved customers. Meanwhile, Anthropic also seemed to get some clearance to release Mythos (not Fable 5) to select, pre-approved US customers. Two weeks after the original restrictions. A bit better than the blanket Anthropic ‘Blip 2.0’ kill switch, where their Mythos/Fable 5 models were forced to be taken off the market altogether by the White House. For two weeks now. With Elon’s SpaceX/xAI IPO now weeks-plus old, and trading down of late, attention turns to the two mega-AI IPOs still ahead. Per The Information, SpaceX ran from confidential filing to listing in a record 74 days — about 37% faster than the typical big-tech IPO — helped by a more hands-off SEC under the current administration. Both Anthropic and OpenAI have confidentially filed, which on SpaceX’s clock could put their prospectuses public by mid-to-late July and a listing as early as mid-August — though August vacations likely push it toward the Fall. My take: it’s a race of narratives against the usual gauntlet of market risk — the longer it takes, the more time skeptics have to poke holes in the story. To match Elon’s execution, both would list before the Fall — but everything in this AI Tech Wave runs on different clock speeds than tech waves past. Indeed, there are indications OpenAI may slow down its mega-AI IPO to 2027. Anthropic may need to do the same, if the Blip 2.0 continues for a while longer. Both founders are being tested vs their pragmatism indeed. More here.

Meta’s WhatsApp acquihire in India: You have to hand it to Mark Zuckerberg for commitment to a strategy. Meta is investing $900 million into Indian fintech Cred (a ~20% stake) and tapping its founder Kunal Shah to run WhatsApp and its 3+ billion global users — replacing Will Cathcart — with a mandate to build out advertising and subscriptions. And to integrate AI agents into the platform, something that was potentially planned with the now China reversed $3 billion Manus AI Acquisition. It’s the same template as the $14B Scale AI acquihire of Alexandr Wang: buy in, and put the founder in charge of a core asset. It also deepens Meta’s big India footprint, from the $5.7B Jio stake to a just-announced first India AI data center. My take: Zuckerberg is singularly focused on sourcing world-class senior AI talent wherever it lives in the world — a differentiated page versus most of his big-tech peers this AI Tech Wave. More here.