For ETF investors, CES headlines are useful less as a short-term catalyst and more as a reminder of how a thematic AI fund is built. THNQ can look diversified at the ticker level, yet the driver set is still narrow: semiconductor platforms, data-center buildout, and edge-computing adoption. That means the fundโs long-run experience will depend on whether the AI stack keeps broadening beyond a few infrastructure winners.
The holdings cited here also highlight concentration and overlap risk. A top-10 or top-15 position can move the story, but it does not eliminate dependence on a small group of companies, customer cycles, and product road maps. Investors using THNQ as a building block should understand that thematic ETFs often concentrate economic exposure even when they hold many names. The practical question is not only โIs AI growing?โ but โWhich layer of AI is being financed, and how durable is that layerโs pricing power?โ
CES also shows why rebalancing matters in theme funds. As companies move from concept announcements to shipping systems, index weights can shift toward firms with stronger execution, while weaker stories fade. That can be helpful, but it also means a theme ETF may rotate into different parts of the AI ecosystem over time, changing the portfolioโs risk profile without changing the label. Long-term holders should review whether the fund still matches their preferred mix of infrastructure, hardware, and edge deployment.
In other words, THNQ is best understood as a mechanism for accessing a fast-moving industry transition, not as a substitute for broad equity diversification. The investor task is to match the ETFโs structure to the role it is supposed to play in the portfolio: satellite growth exposure, not core stability.
Nebius Group (NBIS)
Nebius Group, a top-10 holding in THNQ at a roughly 2.43% weighting, has differentiated itself as a first-mover in next-generation computing. The company announced it will be among the first globally to deploy NVIDIAโs Vera Rubin NVL72 systems across its U.S. and European data centers starting in the second half of this year. This offers superior performance for reasoning and agentic AI, along with cost benefits compared to prior generations.Ambarella (AMBA)
Ambarella, another THNQ holding, has intensified competition with larger rivals by simplifying edge AI development, a significant move toward democratizing the technology. Specifically, at CES, the company launched the Ambarella Developer Zone, a centralized platform providing โagentic blueprintsโ that allow developers to deploy complex AI rapidly without custom engineering. The hardware catalyst for this shift is the new CV7 SoC (system-on-a-chip). Built on Samsungโs 4nm process, the CV7 delivers a 2.5x jump in AI performance over their previous flagship CV5 while consuming 20% less power. This chip supports 8Kp60 video, transformer networks, and vision-language models (VLMs), targeting high-growth sectors โ drones, robotics, ADAS (advanced driver-assistance systems), and enterprise security โ represented within THNQ.Advanced Micro Devices (AMD)
AMD, a top-15 holding in THNQ, aggressively expanded its reach into the edge AI market to compete with NVIDIA and Qualcomm. AMD Original Postress-releases/2026-1-5-amd-and-its-partners-share-their-vision-for-ai-ev.html" target="_blank" rel="noopener" shape="rect">introduced the Ryzen AI Embedded P100 and X100 series. These chips target for โtight spaceโ applications like medical devices, automotive dashboards, and robots, allowing for split-second local decision-making. The launch of the Ryzen AI Halo developer platform further signals AMDโs intent to rival NVIDIAโs dominance in the local AI workstation market. Simultaneously, CEO Lisa Su unveiled the Helios rack-scale platform. AMD designed Helios to train trillion-parameter models and deliver up to three AI exaflops per rack.Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.