Morgan Stanley Research: With $16 Billion In R&D Spending This Year Alone, NVIDIA “Can Maintain A 4–5‑year Development Cycle By Running Three Design Teams Sequentially — Each With An 18–24‑Month Architectural Cadence”

Custom AI-focused ASICs are increasingly being billed as NVIDIA's Achilles heel, capable of bringing down the entire demand-related edifice that NVIDIA has painstakingly built by leveraging its strength in the GPU arena. Yet, Morgan Stanley Research now thinks that ASICs pose no threat to NVIDIA's broader success story.

To wit, Morgan Stanley Research has now penned an exhaustive note on the evolving GPU vs. ASIC paradigm, arguing that NVIDIA faces no material threat to its dominant position in the market.

The research arm of the Wall Street titan concedes at the outset that "momentum in the AI sector has clearly shifted toward custom silicon," and so has the market, with NVIDIA currently commanding a ~$3 trillion market capitalization on the back of $32 billion in quarterly revenues, while Broadcom (AVGO) now supports a $1.1 trillion market cap on just $3.2 billion in quarterly revenues.

Morgan Stanley Research then notes that the development budget of a given ASIC is generally less than $1 billion, and often much lower. In contrast, NVIDIA spends an insane amount of money to deliver an immaculate product:

"... NVIDIA will invest approximately $16 billion in R&D this year alone. With that funding, NVIDIA can maintain a 4–5‑year development cycle by running three design teams sequentially—each with an 18–24‑month architectural cadence—delivering innovation over a five‑year span. In addition, they invest billions in interconnect technologies to boost rack‑scale and cluster‑scale performance..."

Of course, custom ASIC's such as Google's Tensor Processing Unit (TPU) are much more customizable. However, Morgan Stanley argues that the "largest AI training and inference clusters are not currently highly customized," and NVIDIA continues to optimize its GPUs for transformer models. Going forward, Morgan Stanley thinks that the greatest utility for highly customized ASICs will lie in "legacy workloads."

Yet, some might argue that the cost factor is the biggest draw for customized ASICs, which can cost as little as $3,000 while NVIDIA's H100 is priced at $20,000.

Here though, Morgan Stanley argues that there are hidden costs that should be taken into account. For instance, the cluster cost for ASICs is often materially higher as they employ the much more expensive optical connect technology, while NVIDIA's 72‑GPU NVLINK domain leverages the much cheaper copper-based technology.

Moreover, with its unmatched buying power, NVIDIA can easily secure preferential rates for high-bandwidth memory (HBM) chips. The Wall Street titan goes on to note:

"The same applies to CoWoS; because many ASICs use smaller dies with larger stacks, the CoWoS cost can be higher than that of NVIDIA. Of course, NVIDIA’s wafer costs might be higher due to reticle‑limited dies, but overall, Nvidia delivers exceptional value."

Next, Morgan Stanley argues that 'software developer hours' should also be incorporated in the Total Cost of Ownership (TCO) for ASICs, where NVIDIA's CUDA (Compute Unified Device Architecture) SDK is at a distinct advantage and provides for a much more efficient software development experience.

Morgan Stanley then goes on to assert that "NVIDIA and AMD will outperform ASIC competitors [this year] — especially in the second half of the year."

"The scale of AMD’s investments across the ecosystem tends to far exceed that of ASIC vendors. This year, AMD has completed two acquisitions of AI software assets. One of these—the acquisition of ZT Systems—involved acquiring a major server ODM, divesting the ODM business while retaining key engineering talent related to rack‑ and cluster‑scale computing. Once in possession of such assets, AMD can deploy them across multiple cloud environments, which drives third‑party support and accelerates ecosystem development. Can ASIC designers within the cloud replicate this?"

Next, Morgan Stanley Research notes that in 2024, commercial silicon controlled 90 percent of the market, led by NVIDIA's $98 billion in chip-based revenue (on a provisional basis since the GPU manufacturer has yet to announce its earnings for the last quarter) and trailed by AMD at $5 billion. In contrast, custom ASICs controlled only 10 percent of the market in 2024, led by Broadcom with $8 billion in revenue, and trailed by Alchip and Marvel whose combined revenue amounted to only ~$2 billion.

In what is a critical prediction, Morgan Stanley then declares:

"We expect the 90% share for commercial products to increase slightly this year."

To bolster this prediction, the Wall Street behemoth notes that Broadcom is overly reliant on Google's TPU to drive sales. Yet, in 2025, NVIDIA "will grow 50 – 100% more than TPU."

Similarly, Marvel is increasingly reliant on Amazon's Trainium ASIC. While Amazon is all set to double its ASIC purchases to ~$4 billion this year, it is also expected to "more than double" its purchases from NVIDIA.

Therefore, Morgan Stanley thinks that NVIDIA's "revenue momentum in the second half [of 2025] will be significantly stronger than the builds from ASIC or AMD."

Beyond 2026, ASIC growth can explode, with Broadcom's own SAM analysis predicting that "Google, Meta, and Bytedance could each achieve between $60 billion and $90 billion in revenue in the 2027 fiscal year with clusters numbering in the millions."

Nonetheless, Morgan Stanley thinks that "the total addressable market (TAM) for AI ASICs would grow from $12 billion in 2024 to $30 billion in 2027—a figure considerably lower than what is being discounted in the market now."

Towards the end, the investment bank notes:

"NVIDIA's biggest short‑term risk is U.S. export controls, which are equally problematic for AVGO. In the long term, the greatest risk is not competition but a slowdown in investment— which we forecast to occur around mid‑2026."

About the author: Writing is my one incontrovertible passion. Over the past six years, he has authored over 2,200 distinct articles on financial and tech-related topics, spanning nearly 1 million words. And he has been a member of Wcctech mobile team since 2025. As an alumnus of the University of Toronto, Rotman Commerce Program, I bring nuance, in-depth knowledge, and a unique perspective to every topic that I cover. When I'm not writing, I'm traveling the world, exploring hidden confectionaries and restaurants as an aspiring food connoisseur.

Follow Wccftech on Google to get more of our news coverage in your feeds.