APH's AI Datacom Strength Grows: Can It Challenge TEL & MRVL?

Amphenol's Q2 2026 IT datacom revenue surged 89% year over year, including 63% organic growth, with virtually all of its 22% sequential increase attributed to AI products; its AI revenue run rate reached roughly $10.5-$11.0 billion. CommScope added more than $1.2 billion of Q2 sales and expanded APH's fiber-optic capabilities, while its IT datacom business nearly doubled year over year. The outlook is supported by rising AI demand for copper, optics and power interconnects, although APH faces strong competition from TE Connectivity and Marvell and trades at a 25.32x forward P/E versus 20.38x for the sector.

Analysis

The key investable distinction is not aggregate AI connectivity demand but where value accrues as architectures migrate from scale-out to denser scale-up fabrics. APH is positioned across passive copper, fiber and power, creating content-per-rack upside and reducing dependence on any single GPU/XPU winner; that breadth can support estimate durability over the next 2-4 quarters. The offset is that broad exposure also makes reported growth more sensitive to acquired revenue and integration execution than the market may be underwriting.

TEL is the cleaner relative beneficiary if hyperscalers standardize higher-voltage power distribution and copper-heavy short-reach links, while MRVL requires optical DSP and switching design-win conversion to monetize the same capex cycle. This creates a useful separation: APH/TEL are content and volume plays, whereas MRVL carries greater technology-transition and customer-concentration risk. Over 6-18 months, more optics entering scale-up systems is incremental for MRVL, but it could also cannibalize portions of passive-copper connectivity; the timing and architecture mix matter more than headline AI capex.

Consensus appears likely to extrapolate current AI-related growth without fully discounting the risk that hyperscaler capex shifts toward internally optimized architectures, reducing supplier share gains even if total spending remains elevated. APH's premium multiple leaves limited tolerance for a deceleration in organic datacom growth, weaker acquired-business margins, or any reduction in AI order visibility. The near-term catalyst is the next earnings cycle and booking commentary; falsification would be sequential datacom growth slowing materially, margin dilution from the acquired assets, or TEL's bookings converting into faster revenue growth.

Key Decisions for Investors

• Initiate a 3-6 month relative-value position: long TEL / short APH in equal dollar amounts only after confirming TEL trades at a meaningful valuation discount. The thesis is that TEL has greater catch-up potential if higher-voltage and data-network bookings convert; exit if APH sustains superior organic datacom growth by more than 20 percentage points for two consecutive quarters.
• Maintain APH as a core long only on post-earnings confirmation of organic growth and acquired-business margin progression; do not chase pre-results at an elevated multiple. Target a 2-4 quarter holding period, with the risk case defined by sequential IT-datacom growth falling below the broader AI infrastructure spending trend or integration-related margin pressure.
• Use MRVL as a higher-beta architecture-transition expression rather than a direct substitute for APH: accumulate only if management provides verifiable 1.6T, scale-up optics, or DSP design-win conversion metrics. A 6-12 month position has asymmetric upside if optics penetrates scale-up faster than expected, but should be reduced on customer-specific inventory digestion or delayed volume ramps.
• Set an earnings-season watch item for APH, TEL and MRVL: compare bookings, organic revenue, gross margin and AI-related order visibility rather than reported growth. A divergence in TEL bookings versus APH revenue would signal a likely 1-3 quarter relative-performance inflection.