Quanta Q2 Revenue Surpasses NT$1 Trillion, AI Servers Account for 75%-80% of Server Revenue
Nashnova编辑部
Quanta Computer posted record Q2 revenue of NT$1.04 trillion (~US$31.6 billion), up 28.1% QoQ and more than doubling YoY, with AI servers delivering 75%–80% of all server revenue for a second straight quarter — the contract manufacturer's profit engine has fully shifted to AI.
How strong was this quarter?
Q2 consolidated revenue hit NT$1.04 trillion (~US$31.6 billion), up 28.1% QoQ and 105.6% YoY — revenue, gross profit, operating income, and EPS all set single-quarter records.
Operating income reached NT$34.2 billion, up 48.7% QoQ; EPS came in at NT$7.43.
Operating expenses rose to NT$17.8 billion, but the expense ratio fell to 1.72%. This means → revenue is growing faster than costs, and operating leverage is kicking in.
Gross margin held at 5% — why didn't high-ASP products drag it down?
Gross margin stayed at 5.02%, in line with prior guidance.
Management explained: Q1 saw concentrated shipments of high-end AI server racks, pushing up average selling prices but diluting margins. In plain terms = when the unit price is very high, the profit margin looks thinner as a percentage even if absolute profit grows.
In Q2 the product mix rebalanced and ASPs stabilized, easing margin pressure. This reflects a shift from a "ramp-and-price" phase to a "steady-state shipment" phase.
How did AI servers reach 80% of the server mix?
Both AI servers and general-purpose servers posted strong double-digit QoQ growth, yet AI servers accounted for 75%–80% of server revenue for a second consecutive quarter.
Quanta reiterated its full-year target: AI servers at 80% of total server revenue.
This means → Quanta's server business is no longer a "traditional + AI, two-legged" story — AI carries most of the weight, while general-purpose servers are the base, not the growth driver.
The Q3 consignment switch — what should investors watch?
Starting Q3, certain clients will shift specific high-value AI server components from buy-sell to a consignment model — components sit in Quanta's warehouse and count as revenue only when the client draws them.
Quanta says this eases working-capital pressure. In plain terms = Quanta no longer needs to front the cash for expensive parts sitting in inventory, freeing up cash flow.
But the revenue-recognition method changes accordingly, so Q3 server revenue may shrink on paper. This means → QoQ and YoY comparisons will hit a reporting-basis discontinuity — investors cannot compare Q3 directly with Q2 and should watch whether Quanta provides an adjusted basis.
Why did notebook shipments beat expectations?
Q2 notebook shipments reached 11.5 million units, up 15% QoQ but down 5% YoY — yet above the company's own forecast.
Two drivers: ① the pull-forward effect from Q1 new-product launches carried over; ② brand OEMs accelerated orders on fears that rising memory prices and AI server demand could squeeze notebook memory supply.
This reflects how the AI server boom is indirectly reshaping the notebook supply chain — two seemingly unrelated product lines are competing for the same pool of memory.
What needs proving next quarter?
Key question one: can AI server demand sustain the current doubling-pace growth, or does it start to plateau?
Key question two: the actual magnitude of the consignment switch's impact on Q3 revenue recognition — whether the market reads a paper revenue decline as "slowing growth" or "an accounting-basis change" depends on how transparently Quanta discloses the adjustment.
This means → Q3 headline numbers may look weaker than Q2, but what truly matters is the underlying AI server shipment trend after stripping out the basis change.
Content is for reference only, not financial advice.