King Yuan Electronics: Google’s 3x CPU Demand and the XPU Rebound Are Not in Consensus Models
Core Thesis
King Yuan Electronics (KYEC) sits at the intersection of two accelerating vectors that consensus models have not fully incorporated: a 3x expansion in Google’s Arm-based CPU testing demand by end-2026, and a sharp sequential XPU ramp starting in 3Q26. The combination pushes revenue growth into a higher trajectory while customer-consigned equipment lowers capital intensity — a setup that should drive positive estimate revisions and free cash flow improvement over the next four quarters.
What the Market Is Underestimating
The Street sees KYEC as a steady OSAT with a cyclical smartphone exposure. Management’s latest disclosures contradict that view. Google, through GUC, is demanding a tripling of dedicated CPU testing capacity by December 2026. This is not a tentative allocation request — it is a committed ramp with 2027 revenue implications. Separately, the largest customer’s XPU business — encompassing both B-series and the new R-series — is expected to generate a sharp quarter-on-quarter revenue lift in 3Q26, not a gradual build.
The market is pricing these as 2027 stories. The evidence suggests they start impacting the P&L within the next two quarters.
Evidence Chain
1. Google CPU Capacity: 3x in Six Months Is a Signal, Not Ambition
The customer, routed through design partner GUC, has explicitly requested KYEC to triple its Google CPU testing capacity between now and end-2026. That request implies a steep volume ramp in production wafers already in the pipeline — consistent with Google’s server CPU deployment cycle for its Arm-based Axion processors.
A 3x capacity call within a six-month window reflects urgent, committed demand, not optional planning. When an OSAT receives this type of directive, the typical lag between capacity installation and revenue recognition is short: test floors fill within a quarter. This means material revenue contribution from Google CPU starts in 1H27, but the build-out itself drives activity and utilization from 2H26.
Investment implication: Consensus revenue estimates for FY27 are likely 5-8% too low on this vector alone. The capacity tripling is not yet in Street models, which still treat Google CPU as an early-stage contributor.
2. XPU Testing: The Q/Q Step-Change in 3Q26
Two XPU generations from KYEC’s largest customer are set to overlap in 3Q26. The B-series XPU, already contributing through 1H26, maintains momentum. The R-series XPU, a new architecture, begins its volume ramp in the same quarter. Management flagged potential strong sequential growth — language that implies a double-digit percentage Q/Q move.
XPU testing is high-value, long-duration work. A simultaneous B-series sustain and R-series ramp creates a step-function increase in tester utilization, not a marginal gain. The revenue impact compounds because XPU units carry higher test intensity and ASPs versus smartphone SoCs.
Investment implication: 3Q26 revenue consensus likely embeds a linear build, not the non-linear jump implied by parallel generation ramps. Positive surprise risk is asymmetric here, given low Street expectations for near-term XPU contribution.
3. Capacity Expansion Without Dilution
KYEC stated it has no plans to raise 2026 capex above current budgeted levels. The major XPU customer will consign additional testers directly, shifting the capital burden off KYEC’s balance sheet. This is structurally important: growth is being funded by the customer, not by shareholder capital.
Investment implication: This breaks the historical pattern where OSAT growth required proportional capex increases that compressed FCF. With revenue growth accelerating and capex flat, FCF conversion should improve meaningfully in 2H26. This creates conditions for a re-rating as the market recognizes KYEC’s improved capital-light profile.
Key Risks
Smartphone Segment Drag Is Real
MediaTek revenue is expected to be flattish in 2H26, with longer new-model testing times providing some offset but no growth. Worse, OmniVision and other smartphone customers remain weak. This segment represents a non-trivial share of revenue and could partially offset the CPU/XPU acceleration.
Mitigant: The magnitude of Google CPU and XPU growth likely dwarfs the smartphone softness. A 3x capacity expansion in one account plus a multi-generation XPU ramp creates absolute dollar gains that flat smartphone revenue cannot neutralize. But timing mismatches could create noisy quarters.
Customer Concentration
Two customers — Google (via GUC) and the unnamed largest XPU customer — drive the entire growth thesis. Any delay in deployment, architecture shift, or allocation change would disproportionately impact KYEC.
Mitigant: Both customers are expanding their silicon programs aggressively and testing is a non-discretionary part of the value chain. KYEC’s position as a qualified, scaled test partner provides some insulation. But concentration risk warrants a lower multiple ceiling than more diversified peers.
Trade Implications
This memo identifies a window between operational reality and consensus expectations. With no additional capex needed and customer consignment lowering capital intensity, the setup favors positive operating leverage in 2H26 and 2027. If Google CPU tripling materializes and XPU ramps in 3Q26 as management signaled, successive estimate revisions should follow. The FCF improvement story, combined with growth acceleration, supports a valuation expansion argument that is not yet reflected in the current multiple.