Andy ZHU

China supply chain intelligence — deep analysis for investors, tools for buyers

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/Ep76Ztd8NJ

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/Ep76Ztd8NJ

[Andy ZHU @AndyZHUfvdu]

Sinoma Science hit a daily limit on "electronic cloth" hype — but wind-turbine blades are 34% of its revenue at just 12% margin. Fiberglass products are 29.5% at 33.6% margin. The AI-server sliver is diluted across a diversified base. Price hikes are genuinely improving earnings quality, but the market is pricing a pure-play that doesn't exist. The gap between the story and the revenue mix is the trade.

[Andy ZHU @AndyZHUfvdu]

@EvilCountryFuck Thanks — section 6 turning into a ready-to-send Ask is the core win, and a traceability/origin field for components is a clear next gap. Logged.

[Andy ZHU @AndyZHUfvdu]

Missing MOQ / lead time / tiers on the public page isn’t “fine in English.”
Treat blanks as a signal before you pay for a sample.
https://t.co/b6m5Op86vu

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/Ep76ZtdGDh

[Andy ZHU @AndyZHUfvdu]

China Jushi raised electronic-cloth prices 15-20% in September. An AI server uses 3-5x more fiberglass cloth than a regular server. Goldman Sachs forecasts the AI-server PCB market hitting $84B by 2028. But the industry leader admits its low-dielectric specialty cloth has generated zero orders.

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/Qh0rPYedu0

[Andy ZHU @AndyZHUfvdu]

China produced 773,000 industrial robots in 2025 (+28%) — the world's largest market. Edge computing for robots needs low-latency, low-power inference. Hygon has the server-CPU DNA and is now pushing downmarket. But at 180x trailing earnings, the market is pricing perfection. A 5-point margin decline while expanding into a cheaper segment is the warning sign. Scale without margin discipline is just expensive growth.

[Andy ZHU @AndyZHUfvdu]

WeChat can unlock a lower MOQ, but don’t negotiate blind. Read the on-page MOQ and every price tier first — the ≥1/sample price and the first bulk break are often very different, and that gap is your baseline. Then ask in chat only for what the page doesn’t already show: whether they’ll match a lower tier, what’s included (print/pack), and lead time at that qty. If the chat number beats the page, get it on the order before you sample.

[Andy ZHU @AndyZHUfvdu]

Good catch. One thing that flips sample math: a lot of 1688 pages show a ≥1 price that looks fine, then the next bulk tier is where the unit price actually drops — sometimes by half. Comparing at your quantity only works if you read every tier, not the first number on the page. What qty are you testing these against?

[Andy ZHU @AndyZHUfvdu]

On 1688, ignore the headline MOQ and read the tier ladder first. Anime merch listings often show a high MOQ next to a cheap unit price, while the first real price break is lower — and the ≥1/sample price is usually 2–4× that break. Before you pay for a sample, check the same row: the qty where unit price drops, whether print/pack is included, and the lead time on that tier. Quality you only learn from the sample. If you paste one 1688 offer link, I’ll read the MOQ and each tier out loud so you can decide sample-or-skip.

[Andy ZHU @AndyZHUfvdu]

@EvilCountryFuck Thanks — that’s exactly the kind of two sentences that help. Per-part lead time with evidence + date-code are the right make-or-break fields for electronic components; I’ll keep that as the bar for what “missing” means in that category.

[Andy ZHU @AndyZHUfvdu]

Looking for sellers who open 1688 / China supplier tabs.
Founder ask: open Demo (no signup) + 2 sentences on what’s missing.
https://t.co/b6s8atBViF

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/Qh0rPYedu0

[Andy ZHU @AndyZHUfvdu]

Hygon's H1 revenue grew 67% and profit 50%, but gross margin fell 4.9 percentage points in a year. Now it's launching a "physical world" chip for robots, entering a lower-price tier just as margins compress.

Ambition or dilution?
A) Ambition
B) Dilution

Vote below.

[Andy ZHU @AndyZHUfvdu]

@EvilCountryFuck Thanks — here’s the no-signup Demo (six sections → tick Ask → Keep):
https://t.co/QskrpZfrs4

If you open it: 2 sentences on what helped / what’s missing for your category.

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/hgOkvnKxyv

[Andy ZHU @AndyZHUfvdu]

Packaging is a volume business with high fixed costs. When utilization ticks up, profit explodes non-linearly — that's leverage, not growth. JCET's Q1 revenue was actually DOWN 1.76%. The H1 5% growth conceals a weak first half. The 79% profit jump looks heroic until you realize it's margin catching up to a flat top line. Watch Q3 revenue to see if the leverage holds.

[Andy ZHU @AndyZHUfvdu]

Full analysis: https://t.co/hgOkvnKxyv

[Andy ZHU @AndyZHUfvdu]

JCET's H1 revenue grew just 5%, but profit jumped 79%. In a capital-heavy business, that didn't come from orders. It came from margin expansion.

What is the 79% really telling you?
A) Capacity utilization is turning up
B) Margin catching up to a flat top line

Vote below.