In This Article:
Baron Funds, an investment management company, released its “Baron Technology Fund” first quarter 2025 investor letter. A copy of the letter can be downloaded here. Market trends are often driven by sentiment in addition to fundamental elements, and the first quarter of 2025 was a clear example of this. January saw a strong performance for the Fund, driven by investor excitement around long-term growth trends in AI. Optimism was fueled by expectations of the new Trump administration's policies to accelerate economic growth. However, by mid-February, fears of tariffs, a potential trade war, and geopolitical shifts reversed market gains, leading to significant volatility. In the first quarter, the fund fell 14.80% (Institutional Shares), underperforming an 11.64% decline for the MSCI ACWI Information Technology Index (the Benchmark) and a 4.27% decline for the S&P 500 index. In addition, please check the fund’s top five holdings to know its best picks in 2025.
In its first-quarter 2025 investor letter, Baron Technology Fund highlighted stocks such as NVIDIA Corporation (NASDAQ: NVDA). NVIDIA Corporation (NASDAQ:NVDA) offers graphics and compute, and networking solutions. The one-month return of NVIDIA Corporation (NASDAQ:NVDA) was 5.37%, and its shares gained 29.04% of their value over the last 52 weeks. On May 9, 2025, NVIDIA Corporation (NASDAQ:NVDA) stock closed at $116.65 per share with a market capitalization of $2.846 trillion.
Baron Technology Fund stated the following regarding NVIDIA Corporation (NASDAQ:NVDA) in its Q1 2025 investor letter:
"NVIDIA Corporation (NASDAQ:NVDA) is a semiconductor and systems company specializing in compute and networking systems for accelerated computing. Its unmatched leadership in AI infrastructure, spanning GPUs, systems, software and networking solutions, continues to drive robust performance. However, NVIDIA’s stock came under pressure during the quarter, as media and investor narratives shifted toward skepticism, ranging from concerns over slower AI adoption to DeepSeek-related fears that future AI training and inference workloads may become more compute-efficient, reducing demand of accelerated computing systems. As discussed above, we believe these concerns are premature. Training cluster buildouts are progressing in line with expectations, while inference will progressively and steadily scale with usage as enterprises integrate AI into real-world workflows and consumers continue to adopt AI applications, such as ChatGPT, Grok, and Perplexity, to name just a few. Moreover, as we shift from standard Gen 1 (“gut based”) AI models to reasoning Gen 2 (long thinking) models, the query response can demand about 100 times more inference compute to provide a better answer. In contrast to these skeptical narratives, NVIDIA delivered a strong January 2025 quarter, which exceeded Street expectations, driven by data center compute revenues growing 93% year over year to $35.6 billion, with $11 billion of revenue from NVIDIA’s new Blackwell architecture, the fastest product ramp in the company’s history. On the February earnings call and at the GTC conference in March, CEO Jensen Huang reiterated a number of NVIDA growth drivers, including: (1) accelerated (GPU-based) computing architectures replacing legacy (CPU-based) computing architectures; (2) multiple generative AI scaling laws, including pre-training (more data, more compute, smarter models), post-training using reinforcement learning from human and AI feedback, and inference with test-time, long-reasoning compute; (3) agentic AI (autonomous, non-human workers); and (4) physical AI (robots, EVs, etc.)."