Nvidia controls the AI universe now - but these tech giants will lead the next boom

By Daniel Newman

Simple, out-of-the box AI to run a business will be a moneymaker for software providers

SaaS companies will be increasingly interesting as part of the unfolding AI movement.

The tech world is littered with boom-and-bust cycles - from the early days of the personal computer and halcyon days of IBM, to the internet (hello, AOL), cloud (the rise of AWS) and the current generative-AI craze, as seen through the market-value prism of Nvidia.

A central core of the narrative is the convenient theory that for every winner, there is a presumed loser - even if only fleetingly. That viewpoint is often myopic, if not downright wrong, but almost always repeated. Yet over the decades, those presumably left in the dust have shown an ability to roar back by adapting to the shifting winds of tech or by reinventing themselves.

The current storyline revolving around artificial intelligence is arguably the most pronounced yet. The rise of the anointed few - Nvidia (NVDA), Microsoft (MSFT), OpenAI and Google (GOOGL) (GOOG) - has drawn attention to a perceived corresponding decline for Salesforce (CRM), SaaS (software as a service) and enterprise-software companies. Over the past few weeks, stocks of these companies have been eviscerated because of earnings results, outlooks that have raised questions around their long-term viability, and general questions about perceived weakness in the sector.

Is the SaaS bubble bursting? It's important to view the current market and the outsized impact of Nvidia on it. Remove Nvidia from the S&P 500 SPX and the U.S. benchmark index would lose around one-third of its gains so far this year. What is more, the AI market has been largely hardware (again, market-maker Nvidia), with contributions from the likes of Broadcom (AVGO), Dell Technologies (DELL) and Super Micro Computer (SMCI).

We are in the front-loading phase of the AI boom, where building out infrastructure and data centers to support accelerated computing is essential.

This stands to reason because we are in the front-loading phase of the AI boom, where building out infrastructure and data centers to support accelerated computing is essential. It is also why GPUs have become such a scarce resource and why we are witnessing big cloud hyperscalers like Meta Platforms (META), Amazon.com (AMZN), Microsoft, Oracle (ORCL), Google and, soon, Apple (AAPL) gobbling up GPUs and investing big to develop their own homegrown AI chips.

In essence, we are in an era of building the next-generation App Store supported by AI, and it won't be possible without infrastructure. At the same time, dollars are flowing from other parts of IT to AI infrastructure and projects. Our research at Futurum Group shows this in the number of OEMs shifting from CPU to AI servers, as well as system integrators moving their IT projects almost exclusively to AI.

The tide is turning for SaaS and software

SaaS companies will be increasingly interesting as part of the unfolding AI movement. This will occur much sooner than many in the market expect as enterprises will be aggressively seeking to show returns from AI investments.

Oracle, SAP (SAP), Salesforce and ServiceNow (NOW), to note a few, have created enormous digital moats with deeply entrenched customers. Each have more than 100,000 customers that depend on them to run their businesses. Abruptly switching to generative AI would be extremely difficult and expensive, taking years to reach just an elementary stage of achieving a user experience for a voice or text prompt.

There will be a move to out-of-the box AI, or simpler-to-implement AI, that will come from companies like those mentioned above, but also companies that will enable enterprises to build their own AI factories with simpler infrastructure and governance. This is what OEMs like Dell, Hewlett-Packard Enterprise (HPE) and Lenovo Group (HK:992) are touting, while software companies including IBM (IBM) and VMware are developing offerings that simplify the governance of enterprise AI. Opportunities will further extend into SaaS-powered data platforms like Snowflake (SNOW) and Databricks, which enable companies to develop data pipelines for AI with less friction on a use basis.

Consequently, in the next two to three years, the consumption layer of enterprise software - much like the consumption of AI infrastructure -will be led by service offerings that let companies quickly adopt AI. Outside of megascale cloud companies and the largest industry-centric companies, this consumption will take place in the cloud or in a SaaS-like model.

Despite unparalleled excitement for the AI infrastructure that Nvidia is pioneering, its revenue is concentrated to a small number of massive customers including Meta, Microsoft, Tesla (TSLA) and others - with at least two individual customers reportedly making up more than 10% of Nvidia's data-center GPU revenue. As an investor, I would better appreciate companies that have 100,000 users and have the ability to raise prices by adding valuable AI-powered features.

So far, the monetization of AI has been limited to a few software players. SaaS revenue is durable, their customers are sticky and AI consumption can be achieved at scale fairly quickly. Big SaaS players have the market and resources that will allow them to successfully pivot. Salesforce and others have yet to prove that subscribers will pay substantially more for their products because of gen-AI features. It is a gap they need to fill.

Daniel Newman is CEO and chief analyst at The Futurum Group, which provides or has provided research, analysis, advising or consulting to ServiceNow, Intel, Nvidia, Microsoft, Amazon.com, IBM, AMD and other technology companies. Neither he nor the firm have any positions in any of the companies cited. Follow him on Twitter @danielnewmanUV.

More: Nvidia's rivals are seen as being in perpetual catch-up mode

Also read: Is Nvidia doomed to be the next Cisco or Intel? Here's what investors need to know.

-Daniel Newman

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

06-29-24 0933ET

Copyright (c) 2024 Dow Jones & Company, Inc.