Melius' Ben Reitzes: We're telling investors to 'sit tight' on these AI names

Key Concepts

• Toppy market: A market that is considered overvalued and likely to decline.
• Irrational exuberance: A term coined by Alan Greenspan to describe unsustainable investor enthusiasm that drives asset prices up.
• Inferencing: The process of using a trained AI model to make predictions or decisions based on new data.
• Compute: Refers to computational power, often measured in terms of processing capability (e.g., GPUs for AI).
• TAM (Total Addressable Market): The total revenue opportunity available for a product or service.
• Hyperscalers: Large cloud service providers (e.g., AWS, Azure, Google Cloud) that offer extensive computing resources.
• AI is eating software: A paradigm shift where AI capabilities are fundamentally changing and often replacing traditional software functions.
• Infrastructure software: Software that provides foundational services for computing, networking, and data management, often distinct from SaaS.
• On-prem data: Data stored and managed on-premises within an organization's own infrastructure, rather than in the cloud.
• Market multiple: A valuation metric (e.g., P/E ratio) used to compare a company's stock price to its earnings or other financial metrics.
• Single digit multiples: A low valuation multiple, often associated with mature or slow-growth companies.
• Quantum computing: An emerging computing paradigm that uses quantum-mechanical phenomena to solve problems intractable for classical computers.

Market Outlook and AI's Long Runway

Ben Reitzes, Head of Tech Research at Melius Research, addresses investor concerns about the market potentially being "toppy" due to recent gains in AI-related stocks. He advises investors to "sit tight," drawing a parallel to the 1996-1997 period when Alan Greenspan declared "irrational exuberance," yet the market experienced three more years of significant returns. Reitzes asserts that the current AI trend "has got a long way to go" because the underlying fundamentals remain strong. He cites Sora AI, a new social media application for generating video, as a prime example demonstrating the immense need for "inferencing" and "compute" resources, particularly for entities like OpenAI. This application's capabilities underscore the foundational demand driving the AI market.

Key Players and Investment Opportunities in AI

Melius Research's coverage extends beyond semiconductors to include AI in general, encompassing companies like Microsoft, Oracle, and Dell. In the semiconductor space, they recommend the "elite three": Nvidia, Broadcom, and AMD. The Total Addressable Market (TAM) for compute and networking is projected to reach $2 trillion by the end of this decade, with potential for even higher growth. This expansion is driven by the increasing demand for inferencing, autonomous vehicles, and robotics, all of which require substantial infrastructure build-out.

Beyond semiconductors, other significant opportunities are identified:

• Dell: Expected to play a crucial role as enterprises adopt AI and need to leverage their on-prem data. Melius Research recently raised Dell's target price to $200, noting its "cheap stock" valuation (currently around a 15x multiple compared to historical 30x) and Michael Dell's ongoing stock buybacks.
• CoreWeave: An "up and coming hyperscaler" recommended by Melius, allied with Jensen Huang (Nvidia CEO), is positioned to handle the "overflow" of compute demand.
• Oracle: Melius anticipates Oracle will likely raise guidance at its upcoming Analyst Day next week, amidst rumors and speculation about margin improvements. They recommend buying Oracle stock ahead of this event.

The Revaluation of Legacy Tech in the AI Era

A significant theme presented is that "AI is eating software," which represents a complete reversal of the previous paradigm, "software's eating the world." This shift is causing a "complete value shift that is revaluing hardware." Consequently, SaaS companies are increasingly behaving like hardware companies, and hardware companies are adopting characteristics of SaaS companies, while semiconductor companies are at the forefront of this transformation.

Reitzes notes that this is still the "early innings" for some legacy tech companies, which are experiencing a "new lease on life":

• IBM: Is performing well, particularly in quantum computing (expected to mature later this decade) and is reinventing itself around "infrastructure software" that is priced by the instance rather than being a SaaS model.
• Dell: As mentioned, is critical for enterprises needing to process AI workloads with their on-prem data.

This disruption could lead to hardware companies achieving a "market multiple or better," while some SaaS companies might see their valuations decline to "single digit multiples," akin to "old hardware."

Intel's Role and Future Prospects

Melius Research currently holds a "hold" rating on Intel, acknowledging it has been a "tough call" due to past focus on AMD gaining server share and "minutia" rather than strategic "deal making." Jensen Huang's recent statement, "we're lovers, not fighters," suggests a potential for collaboration between Nvidia and Intel. Pat Gelsinger, Intel's CEO, is praised as someone people "like" and "want to work with," with speculation even about potential collaboration with AMD, given Lisa Su's strong leadership.

While Intel might "take a rest for a little while" in the near term due to current numbers, Reitzes emphasizes that they "can't be counted out." This resilience is partly attributed to the backing of the US government (a move initiated by the Trump administration). However, Intel's primary challenge is that its AI strategy is not "as visible as the others that we recommend."

Conclusion/Main Takeaways

The AI market is far from "toppy" and has substantial growth potential, driven by fundamental demand for compute and inferencing, as exemplified by applications like Sora AI. Melius Research recommends key semiconductor players (Nvidia, Broadcom, AMD) and highlights opportunities in emerging hyperscalers (CoreWeave) and enterprise solutions (Dell, Oracle). A significant paradigm shift, "AI is eating software," is revaluing hardware and giving legacy tech companies like IBM and Dell a new lease on life. While Intel faces near-term challenges and a less visible AI strategy, its long-term prospects are supported by government backing and potential for collaboration, though its immediate numbers may require a period of consolidation. The overall sentiment is one of continued bullishness on the AI sector and its transformative impact across the tech landscape.