Listen to the blog post below:

Semiconductors as AI’s Backbone

The AI market has evolved significantly in the past two years, shifting from a heavy reliance on mega-cap and semiconductor dominance to a more diverse set of beneficiaries. The launch of ChatGPT was a landmark moment for AI and a wake-up call for the market. Semiconductors,¹ particularly Nvidia, took center stage with its graphics processing units (GPUs), driving a 53% gain in the space over the following 18 months—outperforming the S&P 500 Index by more than 20% for the period. Key players like Micron and SK Hynix played important roles with high-bandwidth memory (HBM) chips, while Taiwan Semiconductor booked its advanced chip manufacturing pipeline well into the future. This semiconductor hardware enabled AI’s early growth, solidifying semiconductors as the foundation of the AI revolution.

Figure 1: What’s Driving AI Equity Performance in a Post-ChatGPT World?

Other firms offering AI solutions—spanning software, robotics, autonomous vehicles and other AI applications—failed to catch the initial wave that benefited semiconductor stocks. Hyperscalers such as Amazon, Microsoft, Alphabet and Meta are leading the AI data center buildout, buying Nvidia’s chips and other semiconductor equipment to create what Nvidia CEO Jensen Huang calls “AI factories.” This infrastructure development is ongoing, with hyperscalers investing more than $200 billion in capex in 2024 and planning an additional $300 billion in 2025. As these AI factories come online, the foundation for AI will be set, and value creation could extend further across the entire AI ecosystem.

Earnings in Focus: Networking Shines While Software Faces Mixed Results

Earnings over the last week² have highlighted that we haven’t made the shift just yet. While compute and memory semiconductors like GPUs and HBMs have historically been the focus, networking firms are now stepping into the spotlight. Companies like Credo Technology (+47.9% following its earnings announcement) and Marvell (+23.2%) posted stellar results due to their capacity to deliver high-speed connectivity solutions for data centers. Broadcom (+24.4%), the designer of Google’s tensor processing unit, also had strong earnings, driven by its networking capabilities. Notably, Elon Musk’s Colossus supercomputer, which connects 100,000 Nvidia GPUs, relies on low-latency, high-bandwidth networking infrastructure to support training at an unprecedented scale. With the latest Nvidia Blackwell chip expected to suffer from supply constraints, investing in the most advanced networking equipment to scale up to a greater total number of GPUs is becoming essential in training more intelligent large language models (LLMs).

Figure 2: Early December Earnings Recap, Select AI Firms

In contrast, software firms have faced a more mixed reception. While MongoDB—a company offering vector database solutions for LLMs—posted a 68.9% earnings surprise, its stock fell 16.9%, signaling investor concerns about valuation amidst a challenging macroeconomic backdrop. GitLab’s earnings beat barely moved its stock (+1.7%), further reflecting a market that is focused on earnings quality and forward guidance. The valuation reset in the software sector—driven by inflation concerns and questions about interest rates—has led to a recent heightened scrutiny of software firms.

Signs of a Broadening Market Ahead?

The past six months have offered possible early signs of a narrative shift. While the broader semiconductor group’s performance has recorded a 2% loss, and the “AI factory” buildout continues, other players are entering the spotlight. Investors are beginning to recognize that AI’s impact extends beyond Nvidia and a few other key players. Software firms are evolving their strategies to monetize AI capabilities. Meta’s growing advertising revenues, Salesforce’s release of “AgentForce” AI service agents and OpenAI’s $200/month premium tier for advanced AI are all notable examples. Google’s “Willow” project has drawn attention to developments in the space of quantum computing, and smaller players like IonQ have shown early promise. Innovative companies leveraging AI for drug discovery, cybersecurity, robotics and autonomous systems have also captured investor attention, further broadening the scope of beneficiaries. As these technologies and applications mature, the convergence of AI’s influence across industries will become more pronounced.

Figure 3: Innovation Companies Take the Lead over the Past Six Months

Conclusion

The AI ecosystem is approaching an inflection point. What began as a hardware-driven story is now transforming into a broader narrative, encompassing software integration, next-generation infrastructure and beyond. Investors who understand how these elements fit together will be better positioned to capture the AI investment theme. The future of AI is vast, and as adoption scales, value will be created across the entire ecosystem—from hardware providers to end users.

At WisdomTree, the rubber meets the road with the WisdomTree Artificial Intelligence & Innovation Fund (WTAI), which is designed to track the total return performance, before fees and expenses, of the WisdomTree Artificial Intelligence & Innovation Index. This index is seeking to generate investment returns by allocating across the different aforementioned groups: 1) Semiconductors, 2) Software, 3) Innovation and 4) Other Hardware.

Figure 4: Standardized Returns

Within figure 5, we see the “2024 experience,” at least so far, with the year nearly on the books. Many are familiar with a strong S&P 500 Index benchmark, delivering nearly 30% cumulatively over this period. Even though the AI topic has generated a lot of attention, WTAI has delivered a return closer to 10%.

We can say that this is a relatively short period and that when we think of AI, we are thinking of a 10–15-year megatrend—not any single year—but we also understand the reality that people may look year-by-year, and they may compare strategies to the big benchmarks.

However, we are encouraged by the period that looks like it starts in early August 2024 in this figure and continues on through the end of the period.

Figure 5: The Year-to-Date 2024 Performance Picture

If we zoom in on the period from August 5, 2024, to December 13, 2024, we do, in fact, see an acceleration of WTAI’s performance. The strategy did deliver outperformance over the S&P 500 Index benchmark over this period.

We are, of course, encouraged, but we know that this is still a relatively short period, and many investors would want to see this on a more sustained basis. We believe there is a chance it can continue. We look at this this way:

• The chances of this acceleration look more positive the more we see broad-based companies performing well, offering strengthening earnings guidance and being featured in different important breakthroughs for the space. It was refreshing, for example, to hear a bit about quantum computing and move away from every AI advance featuring Nvidia.
• The big benchmarks, like the S&P 500 Index, are well positioned if the biggest companies continue to drive with the best performance. That was the story for the better part of the last two years. We know that “trees don’t grow to the sky,” but we also know that we are not aware of the exact limit as to how high they can climb.

Figure 6: Zooming In

We remain very excited about the prospects for AI and think that 2025 may be an exciting year.

Check out our PATH suite of tools yourself to explore more around WTAI’s performance and read the latest WisdomTree U.S. Thematic Monthly Update for more AI insights.