Plutux

Semiconductors

Chips, from the fab floor to the income statement

Foundry capacity, HBM supply, packaging bottlenecks and export rules — traced through to the companies whose quarters they decide.

What Big Funds Own22 fundsWhat Berkshire, Bridgewater and other big funds reported holding last quarter.What Analysts Expect16 stocksRatings and price targets from the big institutions, each with its public source.

2026-08-26

Semtech turns “optics hype” into a Signal-Integrity-led AI backlog story—while Industrial IoT revenue stays the drag insight cover
Earnings
SMTC•7 min read

Semtech turns “optics hype” into a Signal-Integrity-led AI backlog story—while Industrial IoT revenue stays the drag

In Semtech’s Q2 FY2027 print, Signal Integrity surged while the company’s IoT connectivity pocket continued to lag the data-center growth impulse. The result is a quarter where margins and free cash flow strengthen enough to validate that AI connectivity demand is being pulled through the non-optical layer—an exposure that can move faster than optics when system-level signal quality becomes the constraint.

Q2 FY2027 net sales: $341.9MQ2 FY2027 Signal Integrity sales: $126.2M
Read article

2026-08-25

Marvell has a rare “price-vs-performance” setup in Q2 FY2027: its custom-silicon pitch meets Nvidia’s server cost shock insight cover
Earnings
MRVL•7 min read

Marvell has a rare “price-vs-performance” setup in Q2 FY2027: its custom-silicon pitch meets Nvidia’s server cost shock

Marvell’s Aug. 27 print is the first earnings-level test of whether demand for custom inference silicon can offset Nvidia’s reported 15%+ AI-server price hikes. Marvell has already guided Q2 FY2027 net revenue to $2.700B ±5%, with gross-margin and non-GAAP EPS ranges that will reveal whether “ASIC substitution” shows up in results—or stays a narrative.

Marvell reports Q2 FY2027: Aug. 27, 2026Q2 FY2027 net revenue guide (midpoint): $2.70B
Read article
OpenAI’s Jalapeño posts 1.5–1.9× more inference work per watt—turning “custom silicon” from an option into a pricing-power threat for NVIDIA insight cover
Industry News
7 min read

OpenAI’s Jalapeño posts 1.5–1.9× more inference work per watt—turning “custom silicon” from an option into a pricing-power threat for NVIDIA

OpenAI published its first Jalapeño inference-chip benchmark results on Aug. 25, claiming 1.5–1.9× more AI work per watt at peak throughput and 1.7–3.6× lower end-to-end latency versus stated comparison systems. Because inference is the spend that scales with live usage, the headline is that OpenAI can potentially shift its next compute build from “GPUs as the default” toward purpose-built accelerators—compressing NVIDIA’s margin cushion just ahead of its Aug. 26 earnings.

AI work per watt (peak): 1.5–1.9×End-to-end latency: 1.7–3.6× lower
Read article

2026-08-24

Amazon's 60% device price hikes confirm memory pricing power is moving into consumer hardware insight cover
Supply Chain
AMZN · MU•10 min read

Amazon's 60% device price hikes confirm memory pricing power is moving into consumer hardware

Amazon raised prices on multiple first-party devices—including the Echo Dot base model—by as much as 60%, explicitly citing “significant increases” in memory and storage component costs. The move is a retail-grade proof point that DRAM/NAND tightness can translate into end-demand pricing, sharpening 2H26 contract-price narratives for Micron, SK Hynix, and Samsung Electronics while increasing the near-term risk that Amazon’s device ecosystem growth depends on a tighter—or costlier—hardware attachment loop.

Amazon revenue (TTM): $775.68BMicron revenue (TTM): $90.27B
Read article
Broadcom's $60B+ AI-debt backstop has shifted credit risk from “equity optimism” to bond-market math insight cover
Industry News
AVGO•7 min read

Broadcom's $60B+ AI-debt backstop has shifted credit risk from “equity optimism” to bond-market math

A reported plan for Broadcom to raise more than $60B of AI-chip debt to support Anthropic and other buyers is now being read through the lens of credit: the company’s CDS spread widening implies lenders worry about who absorbs losses if AI demand timing or collateral values miss. For investors, the signal is less about whether Broadcom can finance chips and more about whether today’s AI “backlog math” understates tail risk across the buildout financing chain.

FY2025 revenue: $63.9BNet debt (FY2025): $48.9B
Read article
Samsung Electronics record payout stops moving the stock because memory cash returns are now “priced, not proven” insight cover
Industry News
005930.KS•8 min read

Samsung Electronics record payout stops moving the stock because memory cash returns are now “priced, not proven”

Samsung’s 2026 shareholder-return program is large enough to matter on cash, but investors are treating it as table stakes while they underwrite the next leg of AI/HBM demand. The market reaction fits the new memory playbook: capital returns are no longer the catalyst—unit pricing, HBM share, and capex discipline are.

Samsung 2026 shareholder return plan: KRW 90T–110TSamsung regular dividend framework: KRW 9.8T
Read article
Qualcomm's handset leverage meets Xiaomi’s Xring-and-TSMC bet — and the data-center plan is the real profit hinge insight cover
Industry News
QCOM · TSM•8 min read

Qualcomm's handset leverage meets Xiaomi’s Xring-and-TSMC bet — and the data-center plan is the real profit hinge

Verified filings show Qualcomm is actively repositioning beyond mobile into data-center compute through acquisitions and its nonreportable data-center business. Meanwhile, publicly available event coverage for Xiaomi’s TSMC-made Xring references an EV recall/product-safety library and does not substantiate the claim that a new smartphone SoC is already shipping and structurally replacing Qualcomm or MediaTek volumes. The investable takeaway: even without immediate share loss proof, Qualcomm’s current cash generation and segment mix determine how sensitive the “handset-to-data-center” narrative is if smartphone SoC competition accelerates.

Qualcomm FY2025 revenue: $44.28BQualcomm FY2025 net income: $5.54B
Read article

2026-08-23

AI fabrics hit the “fabric wall”: switch-ASIC + SerDes/retimer capacity, not GPU supply, is tightening the next data-center upgrade cycle insight cover
Supply Chain
ANET•8 min read

AI fabrics hit the “fabric wall”: switch-ASIC + SerDes/retimer capacity, not GPU supply, is tightening the next data-center upgrade cycle

As AI clusters scale toward higher-port-count, 800G-class and 1.6T-era fabrics, the limiting factor is shifting from GPU compute availability to the networking fabric bill of materials—especially switch-ASIC throughput and the SerDes/retimer ecosystems that make optical and copper links behave at scale. Arista has already disclosed that component shortages and lead-time dynamics force it to use non-cancellable semiconductor purchase commitments that can raise working-capital needs and pressure gross margins, while its 1.6T-ready 7060XE7 line shows the engineering push toward the next fabric tier. The practical investor takeaway: watch the silicon and interface bottlenecks that sit between “GPU arrival” and “fully networked training,” because that’s where revenue timing and margins can swing first.

Arista revenue pace: $10.54BArista gross margin sensitivity: Gross margin pressure risk
Read article
AI server build-out is paying ODMs with single-digit margins — the market sees GPU demand, but not the margin math between silicon and racks insight cover
Supply Chain
12 min read

AI server build-out is paying ODMs with single-digit margins — the market sees GPU demand, but not the margin math between silicon and racks

AI server demand is booming, yet the ODM layer that converts GPU shipments into hyperscaler-usable racks is structurally capped by cost-of-sales pass-through, tariffs, expedite charges, and component sourcing concentration. Using Supermicro’s latest SEC disclosure as a load-bearing reference point for gross margin drivers, and cross-checking the revenue scale of major ODMs like Hon Hai Precision, Quanta Computer, and Wistron, the payoff is less “margin expansion” than “volume absorption,” which changes how investors should price order growth and onshoring headlines.

Supermicro gross margin (quarter ended Mar 31, 2: 9.9%Supermicro gross margin (nine months ended Mar 3: 8.2%
Read article
Aug. 22’s semi-led selloff sets up Nvidia’s Aug. 26 call: the tape is repricing “duration,” while Canada tariff retaliation tightens risk appetite insight cover
Markets / Event
7 min read

Aug. 22’s semi-led selloff sets up Nvidia’s Aug. 26 call: the tape is repricing “duration,” while Canada tariff retaliation tightens risk appetite

Aug. 22 was not just another chip wobble: the Nasdaq’s weekly slide aligned with a macro mix of higher bond-yield sensitivity and escalating cross-border trade risk from Canada’s planned “dollar-for-dollar” tariffs. With Nvidia scheduled to report fiscal Q2 results on Aug. 26, the near-term question for investors is whether guidance can overpower duration-driven multiple compression—and how semiconductor buyers plan budgets under a more punitive policy backdrop.

Nvidia call timing: Aug. 26, 2026
Read article
Credo’s AEC print is the first “copper-rack” signal for AI buildouts — and it matters because optics don’t fix the margin math insight cover
Earnings
CRDO•8 min read

Credo’s AEC print is the first “copper-rack” signal for AI buildouts — and it matters because optics don’t fix the margin math

Credo’s latest guidance implies the AI-connectivity bottleneck is shifting downstream to the rack-level interconnect layer, where its active copper AECs should protect gross margin. In FY2026, Credo linked revenue growth overwhelmingly to AEC ramp and credited margin expansion to scale rather than optics mix — a setup where a “quiet” AEC beat can outweigh the market’s focus on the optical leg.

FY2026 revenue: $1.34BFY2026 gross margin: 68.0%
Read article
NVIDIA's 15%+ server price hike raises the payoff for custom inference silicon—and reframes who captures AI profit into 2027 insight cover
Industry News
NVDA · MRVL•8 min read

NVIDIA's 15%+ server price hike raises the payoff for custom inference silicon—and reframes who captures AI profit into 2027

NVIDIA’s reported 15%+ AI-server price increases—linked to memory costs—will likely make “build vs. buy” math harder for OEMs and easier for hyperscalers considering custom inference silicon. The underappreciated consequence is that higher platform BOM costs can accelerate ASIC substitution, shifting a bigger share of compute economics toward winners already embedded in custom stacks like Marvell, Broadcom, and hyperscaler-designed accelerators tied to Microsoft.

NVIDIA revenue (TTM): $253.5BNVIDIA gross profit (TTM): $188.0B
Read article
Quantum’s next supply chain winner won’t be the “best qubit”—it’ll be the firm that owns the atoms-to-atmosphere stack insight cover
Supply Chain
10 min read

Quantum’s next supply chain winner won’t be the “best qubit”—it’ll be the firm that owns the atoms-to-atmosphere stack

DoD, Oracle, and hyperscalers are pulling quantum hardware decisions forward, but the bottleneck is no longer the algorithm—it’s cryogenics (or its absence), trapped-ion control electronics, and the foundry/packaging chain. IonQ’s planned SkyWater acquisition and its $5.7M DoD-backed networked-quantum design work point to a defensible economic play: closing the loop between device physics and manufacturability before the market standardizes.

SkyWater purchase consideration: $35.00Deal structure: Cash + stock
Read article

2026-08-22

NVIDIA price hikes are the market’s missing AI-spending signal insight cover
Industry News
NVDA•7 min read

NVIDIA price hikes are the market’s missing AI-spending signal

Reports say NVIDIA notified AI server customers that prices for systems containing its chips will rise by more than 15% in many cases, effective on early-next-year shipments. The timing—four days before its Aug. 26 results call—turns a “demand is still tight” message into something closer to accounting: if hyperscalers absorb higher system prices, Nvidia’s margin and backlog quality matter more than the headline revenue growth rate.

FY2025 revenue: $130.5BFY2025 net income: $72.9B
Read article
Synopsys’ guide is a stress test of the “design is the bottleneck” AI-chip thesis—and China licensing exposure is the real swing factor insight cover
Earnings
7 min read

Synopsys’ guide is a stress test of the “design is the bottleneck” AI-chip thesis—and China licensing exposure is the real swing factor

Synopsys’ latest guidance is designed to separate two demand drivers: whether the AI-chip bottleneck is dominated by design/verification work (EDA and silicon IP) rather than wafer starts. At the same time, its China/export-control language frames how much of any design-led upside can actually be recognized through its licensing model under shifting approvals.

Q2 FY2026 revenue mix: Time-based: 42%Q2 FY2026 upfront share: Upfront: 24%
Read article
Idaho lab probes Chinese lidar security risk—Hesai and RoboSense face a potential AV sensor “ban-or-divest” shockwave insight cover
Industry News
7 min read

Idaho lab probes Chinese lidar security risk—Hesai and RoboSense face a potential AV sensor “ban-or-divest” shockwave

A U.S. Department of Energy lab is investigating whether widely used Chinese lidar sensors could introduce cybersecurity risks into U.S. vehicles. The immediate threat isn’t just reputational—it’s the procurement layer, where OEMs may accelerate qualification gaps, firmware testing, and supplier swaps before any formal restriction lands.

Hesai Group revenue (TTM): $3.33BHesai Group gross profit (TTM): $1.36B
Read article

2026-08-21

YMTC’s $4.9B Shanghai IPO turns into a NAND capacity lever — and it puts SanDisk’s “floor pricing” thesis on trial insight cover
IPO
8 min read

YMTC’s $4.9B Shanghai IPO turns into a NAND capacity lever — and it puts SanDisk’s “floor pricing” thesis on trial

YMTC’s parent is progressing toward a Shanghai STAR Market IPO, targeting roughly $4.9B, to fund a NAND buildout that—by company-reported planning—could reach ~300,000 wafers/month across three lines by 2026. That matters for the pricing-battle thesis because SanDisk’s reported backlog strategy is built around sustaining floor-like economics through the cycle; any step-change in supply intensity raises the probability that those floors get tested in negotiations, even if demand remains strong.

IPO funding size (reported objective): ~$4.9BPlanned NAND investment: $24B
Read article

2026-08-20

Marvell just got a Google route into TPU-attached custom AI silicon—tightening the “ASIC duopoly” pressure on NVIDIA’s inference economics insight cover
Industry News
MRVL · GOOGL•7 min read

Marvell just got a Google route into TPU-attached custom AI silicon—tightening the “ASIC duopoly” pressure on NVIDIA’s inference economics

Marvell’s newly disclosed agreement with Google ties Marvell-designed custom semiconductor products to the TPU ecosystem, with Google receiving a warrant equivalent to about $12.2B of Marvell shares and vesting linked to Custom Products revenue. The investment implication is not “Marvell wins a chip”; it is that hyperscalers can diversify beyond Broadcom-led custom designs while still optimizing inference compute near TPU—raising the probability of faster cost-per-inference improvements that directly pressure NVIDIA’s accelerator pricing power.

Warrant size (Marvell shares): 58,970,907Exercise price per share: $206.58
Read article
SK Hynix's $29B buyback is a cycle bet: management is paying in the stock selloff to pull the bottom forward insight cover
Industry News
000660.KS•8 min read

SK Hynix's $29B buyback is a cycle bet: management is paying in the stock selloff to pull the bottom forward

SK Hynix approved a KRW 40 trillion (about $28.6B) share repurchase and cancellation on Aug. 19, with buying starting Aug. 20 and running about three months. In a memory market still trading through a selloff, that timing uses buybacks as a “bottom-signal,” while the parallel strength of its U.S. ADR creates an unusual capital-structure setup for investors and forces peers to answer the same payout question.

Program size: KRW 40T (~$28.6B)Repurchase window: ~3 months
Read article
Wolfspeed’s outlook break shows the SiC trade is splitting: AI power could stay tight, but EV/industrial SiC demand is no longer a one-way bet insight cover
Earnings
7 min read

Wolfspeed’s outlook break shows the SiC trade is splitting: AI power could stay tight, but EV/industrial SiC demand is no longer a one-way bet

Wolfspeed’s Aug. 19 guidance reset (and the stock’s -7.5% pre-market move into Aug. 20) is the first “hard” datapoint that AI datacenter power silicon and EV/industrial SiC are moving out of lockstep. The message for investors: the market may reward the SiC supply chain differently by end-market—tight 48V-to-rack power needs versus a cooling EV/industrial ramp—creating winners in adjacent power silicon and control platforms while pressuring wafer/device oversupply.

Q1 FY2027 revenue range: $140M–$160MGross margin trajectory: Non-GAAP negative
Read article

What to expect

Evidence-first notes with a visible point of view.

This section collects sharp takes on earnings, shareholder meetings, and market structure. Each new piece should make the thesis, the facts, and the implications obvious within the first few screens.

Expect direct analysis, not generic commentary.

Expect the data to be explicit and the argument to be easy to follow.

About UsPrivacy PolicyTerms of ServiceDisclaimerStock DirectoryETFsIndices13F HoldingsAnalyst TargetsFree ToolsTechnology MapJoin our Slack

Plutux is not an investment adviser. Market data and AI-generated analysis are for information and education only, not investment advice. Disclaimer

© Plutux Technology Limited 2026