This is absolutely ludicrous by Micron. Micron are refusing to pay their Taiwan workers higher bonuses despite generating more operating income last quarter ($44.64B) than the entire GDP of Iceland ($42.77B). Micron's Taiwan bonuses in September were just 4-5% of operating income according to the union. SK Hynix and Samsung's bonuses are 10% and 10.5% of operating income... Micron needs to close that gap.

JPM Memory Market Update: The order visibility is extending as we progress through the year, but we believe the market is anticipating AI funding risk and power bottleneck driven memory downcycle as early as in 1H27E. This puts memory in a tricky spot as investors debate between “too-cheap to ignore” vs. “it is cheap for a reason.” We believe EPS cuts are unlikely to happen in the next 6M as the memory S-D gap is too wide to narrow over a short period of time and memory earnings will have a buffer protected by LTAs and prepayments. On a medium-term basis, we continue to view risk-reward as attractive and remain OW across all our coverage companies and rank our next three months order of preference based on shareholder return policy upside: SKH (OW) > SEC (OW) > Kioxia (OW).

Absolutely insane... HFs have pulled back gross exposure by ~8% -- the most aggressive de-grossing in at least ten years -- according to Morgan Stanley’s Prime Brokerage data. In addition, TMT-focused HFs lost ~12% in October with losses concentrated in crowded Tech names -- according to Goldman Sachs Prime Brokerage. HFs are cutting both longs and shorts to protect their solid YTD gains (and bonus packages) accrued in H1 2026. That unwind reinforces selloffs in crowded stocks, which is why Amazon, Google and Intel saw violent sell-offs.

Intel's earnings call was horrific. The call sounded like we were listening to a nervous C-Suite meeting held at midnight rather than one where the aim was to ground (and even excite) investors in the face of adversity. Tone matters a lot when the message is bad.

Some quick thoughts on Intel, Sandisk...and AMD. You can make sense of Intel's earnings miss via Sandisk's updated revenue mix. Sandisk: Around 88% of Sandisk's Q/Q revenue gain came from Data Center. However Ex-DC revenue, which includes edge devices like PCs, grew by only 4% despite price increases. Follow that through to Intel & Lip-Bu Tan's comments: "Turning to our client business, in recent weeks, our customers began signaling a PC slowdown based on the memory prices and availability." While the market had already somewhat priced-in a weak PC market for Intel earlier in the year, I don't think anyone expected it to be *this* quick. I therefore naturally wonder whether this is a negative (but smaller) read on AMD, whose gaming business was already down 31% Y/Y in Q2 due to memory constraints. I ultimately think that's why AMD's stock is currently down 7% after-hours.

Blood in the water. $AMZN with good AWS growth of +39% Y/Y but a weak Q4 guide. $INTC with an ultra weak Q4 guide. Margins expected to compress to 38.0% vs. Est. 42.4%. What? LBT...please explain Sir. $SNDK with insane revenue growth. Shock. $RDDT with lower DAUs of 130.3M vs. est. 133.4M. U.S. DAUs of 50.7M are also down 2.5M Q/Q. Why do all these companies report on the same day?

Initial notes on $MRVL's rumoured $LITE acquisition: 1. Why?! - $LITE's CEO has said they're sold out into 2029. So do management think the InP laser shortage is shorter than what their orders suggest? Need more clarity on why Lumentum would green light this. - With lasers sold out into 2029, I think owning the fab might be the only way $MRVL can guarantee supply for FY28 growth. So maybe Marvell and Lumentum have different reads on the InP laser shortage? 2. This has to be mostly stock. - $MRVL has ~$3.9B cash vs. $5.0B debt. Combined run-rate EBITDA is circa $7-7.5B (my numbers), so I think staying investment grade caps new debt at around $15-20B. - At a ~30% premium that leaves $85-90B+ paid in stock, with $LITE holders owning ~30% of the combined company. - I feel that’s a lot of Marvell to hand over. 3. I think the strategy is much better than the reaction. - $MRVL are essentially buying the part of the optical link that survives the move to CPO which is a pretty smart hedge on their own DSP franchise. - Marvell announced Inphi around 6 years ago (~$10B in cash + stock), and that deal built the core of MRVL's optics business. - This is basically the same move just one layer down, from DSP to laser, stacked on Celestial AI. - Probably a good signal that Marvell are running the same optics playbook again? 4. Dilution. - $MRVL trades on about 52x run rate EPS. - $LITE is about 60x on its FQ1 guide at tonight's AH price and about 70x with a 30% premium. - That's ~15% dilutive to Marvell's run-rate EPS (my rough napkin math number) assuming a 30% premium, ~$17.5B of new debt at ~5%, and the rest in stock at $228. (Again, very rough numbers). - It only turns accretive if Lumentum keeps outgrowing Marvell at today's guided rates (+24% q/q vs +15%) for another five or six quarters. - I think that's achievable while the laser shortage lasts, but I feel the whole deal ends up being a bet on how long that is. 5. Fewer lasers on the open market. - If LITE's output gets prioritised for $NVDA and Marvell's own modules, everyone else buying LITE EMLs and CW lasers gets tighter allocation. - I think that's bullish for laser pricing at $COHR and for other laser sources like $SIVE (lol). - Feels like the module makers that depend on LITE lasers are the quiet losers. 6. Channel conflict. - Marvell is the leading DSP supplier to 800G/1.6T module makers, and LITE already sells its own modules through Cloud Light. - I think Innolight and Eoptolink start shifting DSP share to $AVGO and $CRDO? - I feel hyperscalers will push harder to second-source lasers from $COHR.

The main takeaway here is that one week OpenAI has the best model, then it's Anthropic, then it's Google, in a never-ending cycle of incramental model improvements. Feel very lucky to be living through this. PS: feels like this is an early signal that $GOOGL will announce increased CapEx in Q3 earnings later today.

Foreign inflows into SK Hynix is a key watchpoint for investor sentiment now. So far this week, US investors have favoured Samsung (+$167M net bought) compares to SK Hynix (-$31M net sold). If foreign inflows into SK Hynix pick up after their stellar earnings, I think that'll be a big step in the right direction.

Don, there is a missing term in your equation. The crunch spread is not: Token Price - GPU rental price. Rather it is: (Token Price * Token Volume) - GPU rental price. More simply: Revenue per GPU - GPU rental price. Continuous throughput improvements for the same exact token mean that volume growth can more than offset token price declines for the exact same model/token i.e. Kimi K3 token throughput per GPU has recently been improving at a 37% annualized rate driven by inference engine optimizations. GPU prices are vertical because: 1) Mix is shifting from higher margin frontier tokens to lower margin open-weight tokens. $10 in Claude revenue requires $1 of inference compute. $10 of of open-weight revenue requires $3 to $7 of inference compute given competition between inference providers leads to lower margins. Open-weight tokens taking share is positive for AI infrastructure demand given lower margins at the model layer. The reverse is also true in that frontier tokens taking share is negative for AI infrastructure demand. Open-weight models have been taking share since early summer. 2) Throughput is steadily improving for each individual token type such that GPU compute becomes more valuable even as true like for like token prices decline.

Recommended reading for the weekend: 1. @jpmorgan — Credit markets look for AI’s weak links 2. @BernsteinPWM — Climbing the Wall of Worry | 3Q Macro & Markets Insights 3. @gdb — Towards securing the internet 4. @AswathDamodaran — From Eyeballs to ARR: Buzzy Pricing Metrics and the AI Trade! 5. @UBS — Higher oil and AI concerns drive risk-off move 6. @neil_shah — NVIDIA GTC Berlin 2026 – Containers, Safety and Sovereignty Rough week in the markets. Hope this weekend serves as some welcome respite!