OpenAI's Jalapeño chip lands on Nvidia's earnings day
Of all the days to announce your own silicon, OpenAI picked the one where Nvidia reports earnings. The AI trade's two biggest questions — who makes the chips, and who pays for them — collide in a single session.
OpenAI's Jalapeño Chip Turns Up the Heat on Nvidia
OpenAI unveiled Jalapeño, its first custom inference chip, built with Broadcom and aimed at the workloads it currently rents from Nvidia. Even if the silicon underdelivers, a credible in-house alternative gives OpenAI leverage on price and supply that it has never had — and every other frontier lab now has a template to copy.
Nvidia's $92 Billion Quarter Lands Tonight — And Wall Street Is Braced for a $280 Billion Swing
Nvidia reports fiscal Q2 after the close with roughly $92 billion in sales on the line, and options traders are pricing a 5.4% move — about $280 billion in market value. No single print does more to set the price of AI infrastructure, and the whole index is along for the ride.
Nvidia's Earnings Week Arrives With Wall Street Watching One Chip: Vera Rubin
The number that matters may not be the quarter at all but Vera Rubin, Nvidia's next-generation platform, which shares slid roughly 2.6% into the print waiting on. Guidance on the roadmap tells you whether the 2027 capex cycle is real; the trailing quarter is already spent.
Everyone Owns the AI Trade. That's Exactly What Has Investors Nervous
With the AI trade now consensus positioning across funds, the risk has shifted from being wrong to being crowded. When everyone is on the same side, a merely fine earnings report can still trigger a violent unwind — which is exactly the setup into tonight.
Apple's M6 and M5 Ultra Push AI Off the Cloud and Onto Your Desk
Apple's M6 and M5 Ultra, in new Mac mini and Mac Studio desktops, push serious model inference onto local hardware. For anyone building with AI, a capable on-desk machine changes the math on privacy, latency, and the cloud bill for development work.
Apple's New Macs Get Faster Chips — and Higher Price Tags
The Mac mini now starts at $899 with 256GB and 16GB of memory — a real jump from Apple's old entry price. The cost of the memory and bandwidth that local AI demands is showing up on the sticker, and it will show up in your hardware budget too.
Claude Now Remembers You Across Chats and Cowork — Unless You Say No
Anthropic has merged memory across Claude and Claude Cowork, so context from a chat carries into agentic work, with opt-out available. Persistent cross-surface memory is what turns an assistant into something that accumulates value — and it makes data governance a question worth answering before rollout.