Is an AI Mega-Tech Correction Underway?

By Daniel Morgan, Synovus Trust Senior Portfolio Manager

The biggest question that clients appear to be asking:  Is this a correction or the start of something bigger? First off, earnings estimates for the second half of 2024 are seeing revisions lower with 3Q coming down 1.4% since July 1 and 4Q decreasing by 0.2%. The bigger concern, however, is the waning “AI enthusiasm.” The circular capital expenditures (CapEx) cycle for the largest, most liquid companies does not appear to be ending just yet, but investors are now questioning what the return on investment (ROI) will be in the future. “AI disappointment” remains the biggest risk to the market, and near the tail end of July this played out. Wall Street investors continue to raise concerns that many of the large Technology companies may be spending too much CapEx on artificial intelligence (AI) and the ROI will not be realized. With the top Tech players (Amazon, Google, Microsoft and Meta) all projected to spend collectively up to $185 billion in CapEx in 2025, a large portion of those dollars will be dedicated toward the AI data center.

After Alphabet’s (GOOGL) 2Q24 financial results, shares sold off as investors digested the news. For the 2Q24, earnings per share of $1.89 (up 31%) on revenue of $84.7 billion (14%) beat analysts’ expectations of earnings per share of $1.85 on revenue of $84.3 billion. Advertising revenue topped $64.6 billion (11%) versus analysts' $64.5 billion expectations. That represented a bit of a pullback in growth from 1Q24 (+13%). YouTube ad revenue, however, fell short, with the segment bringing in $8.66 billion (13%) versus $8.95 billion expectations. That represented a drop in growth from 1Q24 (+21%). This miss could potentially be due to heightened competition from other players like TikTok. The CFO commented on a recent conference call that YouTube “continued to close the monetization gap with Shorts,” the company’s TikTok rival.

Google Cloud Platform (GCP) revenue was at $10.35 billion (28%); that's better than analyst expectations of $10.1 billion. That matched GCP growth posted in the 1Q24 (+28%). I believe many investors were hoping for an acceleration in growth for GCP that could be directly related to all the Gen-AI CapEx spending Alphabet is doing. Alphabet reported spending $2.2 billion building AI models across its DeepMind and Google Research organizations. CapEx was $13 billion in 2Q24. Alphabet’s CapEx outlook for 2024 guidance is consistent at an estimated $50-52 billion. Investors left “scratching their heads” after Alphabet’s conference call, wondering when will AI begin to generate revenue for Google’s cloud business and its core ad segment. At this point, that appears to still “be up in the air.”

On the other hand, IBM (IBM) stock rallied following its strong 2Q24 earnings report, beating on both the top and bottom lines. IBM's disclosure of $2 billion in booked revenue from generative AI, doubling from the first quarter. IBM's approach to AI differs from its competitors, focusing on the hybrid space. IBM has a different strategy toward AI and cloud than a lot of these hyper-scalers. IBM, at this point, is not as focused on the data center like Amazon’s AWS, Microsoft’s Azure and Alphabet’s GCP. Simply put, IBM isn't playing in the same sandbox as other these other prominent AI names. For FY2024, IBM is expected to spend approximately $1.8 billion on CapEx and is on target to generate $12.1 billion FCF.

IBM’s AI initiative surrounds its Watson X product. During the recent “Think 2023” event management highlighted IBM's continued AI focus, with the new Watson X suite set to enable enterprises with AI apps, foundational models and AI governance tools (all run on OpenShift). IBM's platform and incumbency could give a competitive edge, but incremental monetization remains the key to success. From 2017-23, AI adoption has doubled to reach 30-40% of enterprises, and IBM aims to capitalize on this opportunity through multiple new product launches, including its new WatsonX suite of offerings, a GPU-as-a-service offering on IBM Cloud and new AI-focused teams within IBM Consulting. All of these offerings expand on IBM's AI work from the last decade. They are also focused on bringing AI capabilities to the enterprise in order to improve key applications such as IT operations, customer care, cyber and digital labor, among others. IBM's key strength in AI is its platform – or ability to provide software, services and hardware – which creates not only a compelling value proposition for customers but also a platform multiplier effect, where consulting becomes the "tip of the spear" that drives AI software purchases.

As the “Street” searches for tangible evidence of a growth spike from Gen-AI spending, Microsoft is expected not to disappoint. Projecting Microsoft's CapEx increasing from $30 billion in FY2024 to $52 billion in FY2024, and then accelerating to $78 billion into 2029 driven by generative AI specific investments, the questions now surround ROI. Can MSFT capture enough incremental increase in profit growth from its investments?

For FY23, Microsoft generated about $250 million of GenAI revenue (0.1% of total), which is expected to grow to $5.5 billion in FY24 (2.2% of total), and $19.2 billion in FY25 (6.7% of total), settling at 50% compound annual growth rate (CAGR). Broadly divide the GenAI industry into four buckets: 1. Developer tools (i.e., coding assistants); 2. ChatGPT/Co-Pilots; 3. Media generation (image, audio and video de-noising models); and 4. Platforms (tooling, models, data, studios and marketplaces). Microsoft will be a particularly strong participant in ChatGPT/Co-Pilots and Platforms. MSFT appears to be in the pole position with other hyper-scalers closing in.

Currently, platforms generate almost twice as much revenue as ChatGPT and Co-Pilots for the industry. For Microsoft, platforms will remain 3x over-the-top subscriptions as Azure will remain dominant hosting hub for both OpenAI and other models, including open source. During 3Q24, Azure growth accelerated to 31% Year-over-Year (YoY), benefiting from a 7% contribution from AI services (vs. 6% in 2Q24). AI commentary remained robust, with demand outpacing supply, while Azure core continued to gain share, also benefiting from an acceleration in workload migration from on-premises. Azure guidance of 30-31% growth in 4Q24 also beat expectations of 28% YoY growth, suggesting a combination of sustainable demand strength and share expansion. Of the three primary vendors in the cloud infrastructure market (AWS, Azure and GCP), MSFT is separating itself from GCP's (growth of 28.4%) and AWS (growth of 17.3%), with growth north of 30% YoY.

Microsoft’s other AI-bet is Copilot. Office 365 Commercial results likely will prove to be disappointing to investors. However, management remains bullish on the longer-term M365 Copilot opportunity. Copilot has been a dominant lead sale for many organizations and inspires enterprises to adopt other Microsoft AI services, which are mostly sold in unique bundles — productivity apps, security, communications — all on Azure with Copilot embedded. The first generally available generative AI product was GitHub Copilot, and while its overall revenue impact is marginal, it serves as an important on-ramp for other Copilot services. Copilot deployment remains gradual but sticky. Commentary on paid seat growth attributed to Copilot would be valuable. Copilot will likely shore up Microsoft’s multi-product relationship within Enterprise.