Is Nvidia the New Cisco?

By Daniel Morgan, Synovus Trust Senior Portfolio Manager

Almost everyone remembers the “Four Horsemen” — Cisco, Dell, Intel and Microsoft — and its leadership during the dot-com era. (For what it’s worth, the phrase was rooted in the nickname given the 1924 Notre Dame football backfield.) That tech group was followed by the “FAANG” stocks - Facebook, Apple, Amazon, Netflix and Google. At one time it represented 25% of the entire S&P 500 Index’s market cap. That leadership has now given way to the “Magnificent Seven.” The Magnificent Seven stocks — Alphabet, Amazon.com, Apple, Nvidia, Microsoft and Tesla — are up an average 87% year to date (YTD).

Will Nvidia (NVDA) fall into the same trap that Cisco did some 25 years later? Nvidia shares have nearly tripled in this year's artificial intelligence (AI) boom. It's not dissimilar from Cisco's trajectory at the turn of the century. Cisco shares still haven't managed to recapture the peaks it scaled back in 2000. Cisco shares hit an all-time high of $70 a share in August of 2000 and now trade south of that milestone at $54 a share!

Investors have piled into Nvidia because it makes the chips backing AI models, like OpenAI's ChatGPT and the Cloud Data Center. Nvidia is now the fifth-biggest U.S. stock, trading at 40 times its sales. Cisco was one of the largest tech stocks during the dot-com bubble, trading as high as 30 times revenues. And much as bulls see Nvidia as well-placed to benefit from today's most exciting emerging technology, Cisco's networking equipment was viewed as integral to the rise of the internet era. But the company's growth couldn't catch up to the hype.

Today, Cisco’s growth has slowed down dramatically since the internet’s boom days. Efforts to diversify into video conferencing, security and services has had some success. However, Cisco’s profits are currently being powered by the old router/switch product group that drove the stock up above 1000% in the 1990s.

The Secure Agile Networks (68% of product revenues) consist of Cisco’s core networking technologies of switching, enterprise routing, wireless and computer products. These technologies consist of hardware and software offerings, including software licenses and SaaS, which help Cisco’s customers build networks and automate, orchestrate, integrate and digitize data. Cisco is shifting its business to software and subscriptions across its core networking portfolio and expanding its software offerings. Cisco’s Secure Agile Networks unit posted $8.13 billion for the 4Q23 Revenue, a 33% Year-over-Year (YoY) increase with switching revenue growing strong double digits — with strength in campus and data center switching. Enterprise routing growth was driven by Cat 8K, SD-WAN and IoT routing. Cisco noted that wireless saw strong double-digit growth driven by Wi-Fi 6. This segment is most likely Cisco’s best bet for generative AI as it produces the backbone of any enterprise or Internet network.

Nvidia is benefiting from the trend of generative AI spending and the data center cloud space is taking priority over the transition to newer, more expensive platforms, as general CAPEX in cloud slows. How is this growth even possible, in the context of 2023 cloud budgets that appear up only slightly? Anecdotes, quantitative modeling and common sense all lead to the same conclusion — cloud customers are rapidly shifting budgets toward AI. This strength is coming at the expense of traditional servers and other legacy areas. There are obviously going to be areas where new servers are needed, but expect that customers are extending the lives of server CPUs, delaying transition to newer, more expensive Sapphire Rapids and Genoa platforms to make room for mission-critical investments in AI. For example, CAPEX spend rates in the data center space lead to all the top IaaS cloud units to post slower growth rates in the 2Q24: AWS (+12%), Azure (+26%) and GPC (+28%). NVDA seems to be bucking this cap-ex general slowdown as the increased usage of AI chips in the data center cloud space is providing a catalyst.

Can Nvidia stock continue to ride this AI mania and duplicate the share performance of Cisco System’s huge run pre-2000? As Cisco shares were driven by building switches and routers that ran the backbone of an enterprise and internet network, Nvidia is being driven by the increase in usage of generative AI in the cloud data center. Data center has been impacted by the ramp of NVDA’s A100/H100 products and is expected to significantly inflect higher in the 2H24, driven by strong demand pull from Generative AI and large language models.

Moreover, given the strong demand for its data center products, NVDA has extended visibility for the next several quarters and has secured more supply to support significant revenue growth in the 2H24. Part of management’s confidence in the 2H24 is being driven by the introduction of NVDA’s new generative AI inferencing engines (L4, L40, H100NVL platforms). Demand for its high-performance networking products were also strong driven by the very high attach rate of InfiniBand solutions to NVDA’s accelerated computer platforms. The company is also ramping multiple new products across its data center portfolio — H100 (GPU), Grace/Grace Hopper (CPU/GPU), BlueField 3 (DPU) and Spectrum 4 (Networking). Expect the DGX H100 is set for volume ramp in late 2023 but could see some challenges to Chinese demand, given the recently announced bans for both the H100 and prior-gen A100. This new Grace Hopper chip will be targeted toward developing and supporting large language models. NVDA’s data center segment now generates more revenues than both Intel Corporation and Advanced Micro Devices, Inc.’s (ADM) combined data center units! NVDA’s data center business is expected to nearly double its revenues in FY2024 to $30.05 billion from $15 billion in FY2023 with AI being the major catalyst.

NVDA’s management does not seem overly concerned about competition given its strong performance lead. Nvidia today accounts for more than 70% of AI semiconductors sales — with Amazon, Google, IBM and Meta all producing their own AI chips. NVDA continues to dominate performance benchmarks and the company’s eco-system advantage would be hard to match. Management recently highlighted that NVDA GPUs already co-exist with custom silicon solutions such as TPUs.

While AMD’s upcoming MI300X looks promising for AI inferencing workloads due to large amounts of integrated Memory (192GB), NVDA management pointed that the company’s Grace Hopper chip supports nearly 600GB of Memory. AMD’s MI300X, a GPU-only version, is designed to go toe to toe with Nvidia’s H100 Hopper AI accelerator and optimized for large language models and generative AI. Further, AMD is working directly with Microsoft to develop additional AI chips under the code-name Athena. Broadcom’s Jericho3-AI will not be ready to ship until 2024. Jericho3-AI is designed to connect supercomputers and features a high-performance fabric for AI environments. Marvell Technology, Inc. is expected to be major beneficiary of the aggressive spending on generative AI by its cloud customers. It’s also on track to double its AI-based revenue to $400 million in FY24, driven by strong demand pull for Marvell’s 800G PAM4 DSP chipsets and the 400ZR DCI solutions.

So how does Nvidia’s valuation today compare to Cisco’s during its prime in August of 2000 when the stock hit an all-time high of $70 a share? In August of that year, Cisco’s shares traded at a $480 billion market cap, on 30x sales, with a forward P/E of 16x Earnings. On the other hand, NVDA’s shares today trade at a market cap of $1.032 billion, on 40x sales, with a forward price-to-earnings of 47x earnings. Therefore, Nvidia’s valuation today does seem much higher when compared to Cisco’s during its prime.

The global AI chip market was valued at $11.2 billion in 2021 and is projected to reach $263.6 billion by 2031, growing at a compound annual growth rate of 37.1%, according to a research report produced by Allied Market Research. That means one could make the case that the market opportunity for Nvidia in generative AI chips today may be larger than the prospects for Cisco were in the enterprise and internet network were back in the 1990s. Of course, as the proverb goes, only time will tell.