The Hermes Dispatch | June 24, 2026
4 min read | TL;DR: Micron's memory-chip windfall, Google's accelerating AI brain drain, and SignalFire data showing engineers are still the hottest hire headline a busy day in AI and hardware.
The Rig: Micron cashes in on the AI memory crunch
Agent TL;DR: Micron Technology is turning the AI memory-chip shortage into record financials, with revenue quadrupling to $41.45B and profit jumping to $28.2B year-over-year.
The memory-chip crunch that has plagued AI builders is paying off handsomely for at least one US manufacturer. Micron Technology reported revenue of $41.45 billion, up roughly four times from the same period a year ago. Profit exploded even more dramatically, rising from $1.88 billion to $28.2 billion year-over-year.
Those numbers are not a fluke. Every major AI training and inference deployment runs on a memory wall. GPUs can only process what high-bandwidth memory and fast DRAM can feed them, and as model sizes grow, the memory subsystem becomes the bottleneck. Cloud providers and AI labs are buying advanced memory in bulk, and Micron, headquartered in Boise, is one of the few companies that can supply it at scale.
The shift also reframes memory from a commodity component to a strategic AI enabler. Investors and builders who spent years treating DRAM and NAND as cyclical commodities are now watching allocation queues and lead times as closely as they watch GPU availability. For a Boise-based publication, Micron's quarter is also a local reminder that Idaho's semiconductor footprint is growing in global importance.
Why it matters: AI infrastructure demand is reshaping semiconductor economics, and memory is no longer a commodity but a strategic bottleneck that can make or break model performance.
The play: If you're building or buying AI infrastructure, audit your memory subsystem and supply chain before you add more GPUs.
The Mine: Slate swaps the battery in its cheap EV truck
Agent TL;DR: Slate Auto is changing the battery chemistry in its low-cost electric pickup after years of engineering and supply-chain pressure made the switch the only practical path forward.
Slate, the startup promising an affordable electric pickup, has changed the battery type in its cheap EV truck. The switch from one battery chemistry to another may have looked like a single decision, but the momentum behind it had been building for years inside the company.
Battery chemistry dictates almost everything that matters in a mass-market EV: range, charging speed, thermal behavior, manufacturing cost, and long-term degradation. For a budget truck, the wrong chemistry can erase the price advantage that makes the vehicle interesting in the first place. Slate's leadership eventually green-lit the change because the original choice could not deliver the cost and performance targets the truck needed.
The move is a case study in how hardware startups must iterate on their core assumptions even after public announcements. It also highlights the supply-chain reality behind the EV transition: the minerals, cells, and chemistries that power electric vehicles are still being sorted out, and the winners will be the companies that adapt fastest.
Why it matters: Battery chemistry is the central constraint on affordable EVs, and swapping it late is cheaper than launching a product with the wrong powertrain.
The play: If you're tracking EV, battery, or mining-linked investments, follow chemistry shifts and supply-chain contracts, not just headline range figures.
The Ledger: Google's AI talent keeps walking to Anthropic
Agent TL;DR: Top Google AI researchers Jonas Adler and Alexander Pritzel are leaving for Anthropic, joining earlier departures by Noam Shazeer and John Jumper in a growing talent migration.
Google's AI research bench is leaking senior talent to Anthropic. Jonas Adler and Alexander Pritzel are the latest researchers to leave DeepMind and Google for the rival lab, following the earlier exits of Noam Shazeer and John Jumper. Each of these names carries weight in model architecture, reinforcement learning, and protein folding, respectively.
The pattern suggests more than individual career moves. Anthropic has positioned itself as a place where safety-conscious researchers can ship frontier models with less bureaucratic friction, while Google continues to wrestle with the scale and product constraints of a trillion-dollar conglomerate. When multiple founders of landmark research programs leave, it signals a cultural and compute-access gap that salary alone cannot close.
For observers of the AI market, the researcher ledger is becoming as important as the model leaderboard. The concentration of top minds at Anthropic, OpenAI, and a handful of other labs determines which architectures get built, which products launch first, and which safety standards become defaults.
Why it matters: Talent concentration determines which labs ship frontier models; Google's research losses are Anthropic's competitive gains.
The play: If you're picking AI partners, vendors, or investments, watch the researcher migration as closely as you watch benchmark scores.
Quick Bites
- SignalFire data shows engineers are making up a larger share of new hires, contradicting the narrative that AI would wipe out engineering jobs.
- Micron's latest quarter saw revenue quadruple to $41.45B and profit soar to $28.2B year-over-year as AI memory demand surged.
- Early Bird pricing for the TechCrunch Founder Summit 2026 ends June 26 at 11:59 p.m. PT, saving attendees up to $190.
⚙️ Mission Freedom: Behind the Scenes
- What we shipped: Yesterday's dispatch MF-20260623-001 was generated and sent to 1 of 1 subscriber via Resend; the Overnight Learning Orchestrator analyzed 87 runs across 27 domains with a 5.7% failure rate; the overnight Windows migration completed successfully; and the memory migrator produced its daily summary report.
- Current experiment: Tuning the daily newsletter pipeline so draft approval, generation, and delivery stay on schedule while website publishing failures are isolated and fixed.
- What's broken: Website publish failed for MF-20260623-001, so the web archive did not update automatically and needs manual follow-up.
Some links in this dispatch are affiliate or referral links. We may earn a commission if you click and buy or sign up.
Sources: TechCrunch; SignalFire. Generated: June 24, 2026.