Day: February 18, 2026

$AVGO and OpenAI announced Jalapeño, purpose-built silicon for LLM inference as the two companies deepen their hardware partnership.

The chip targets the inference bottleneck that has become critical as enterprises scale AI deployments.

One of my favorite moves in 2026 was $AMD acquisition of MEXT.

MEXT makes predictive memory software that uses AI to move "cold" data into cheaper flash and prefetch it back into DRAM before an app needs it effectively making low-cost flash behave more like expensive memory

$CBRS Q1 EARNINGS

• Revenue $193M vs Est. $181M (+92% YoY)
• EBITDA $13M vs. Est. ($5M)

FY26 Outlook
• Revenue $860M vs Est. $830M (+69% YoY)
• Gross Margin 40% vs. Est. 44%

The company also announced a $20B+ OpenAI deal for 750MW of capacity and new $AMZN AWS partnership

"Customers will care more about memory efficiency and full systems economics rather than just GPU performance."

The $AMD & MEXT deal is a bet on inference economics. It is not an HBM killer.

$QCOM Investor Day tomorrow. Expecting a strong slate of datacenter wins to back our long term thesis that it has a massive multi-billion revenue expansion op by winning a small fraction of a massive DC TAM. 👀

$META CEO Mark Zuckerberg has tasked the company with developing a prediction market platform.

Meta joins a growing roster of tech firms exploring prediction markets as a new user engagement and monetization vector.

"There's nobody more important than $ASML in the AI economy for the equipment layer."

It isn't a memory name, but a durable memory cycle pushes more capex straight to the equipment layer ahead of tomorrow's $MU print.

Select Date

THE BIGGEST WINNER FROM $NVDA + $META PARTNERSHIP

Podcast

sentiment

powered by

DAILY RIP

powered by

INSIGHTS

powered by

Rate your experience