Software at the Top of the AI Value Chain
The Search for AI’s Killer App Has Begun
iShares Software ETF Up 59% Year-to-Date
It would not be an exaggeration to call this year the first true year of generative AI. Since OpenAI’s ChatGPT was introduced to the public in December 2022, virtually every industry has begun exploring how to apply and invest in AI. Yet aside from ChatGPT—which has amassed an estimated 1.6 billion users—a true killer app has yet to emerge. Concerns remain that AI could ultimately follow the path of past themes such as the metaverse or Bitcoin, rising rapidly only to fade into a single-cycle trend.
Like any industry, AI has both a value chain and a timeline. Global consulting firm McKinsey divides the AI value chain into six stages, ranging from hardware computing to end-user services. At the very top sit applications and services. Barbarian Research, a U.S. equity research firm founded by former analysts and fund managers, has been closely tracking the shift of AI momentum from semiconductors and infrastructure toward software services, particularly since the third quarter. In other words, this is becoming the era of software.
Microsoft (MSFT), widely known as a major investor in OpenAI, began monetizing its AI-powered Copilot service within Microsoft Office 365 in November. The name combines “co,” meaning “together,” and “pilot,” reflecting its role as an AI assistant. Copilot generates documents in formats such as PowerPoint and Word and is priced at USD 30 per month for enterprise users. Wall Street estimates suggest that Copilot alone could add more than USD 10 billion in annual revenue by 2026.
Adobe (ADBE), best known for its PDF software, is focusing on AI-driven image creation and editing through a service called Firefly. Users can generate and modify images simply by entering text prompts. Adobe bundled Firefly with Photoshop and Illustrator under its Creative Cloud subscription and raised prices by 10%. Analysts expect Firefly to be a meaningful driver of earnings growth next year.
Another potential killer app is Unity (U), whose shares have risen 54% since November. Unity provides tools for creating and editing 2D and 3D content and is widely regarded as a game-changer for the gaming industry. Some observers also see it as a possible catalyst for making the metaverse more viable by addressing the chronic shortage of compelling content.
In the U.S., several ETFs focus on software investments. The SPDR S&P Software & Services ETF (XSW) diversifies across roughly 140 software companies, with individual weights typically between 0.5% and 1.0%, resulting in a near equal-weight structure. The Invesco AI and Next Gen Software ETF (IGPT) invests in about 100 companies, but its top 10 holdings account for 59% of the portfolio. Moreover, because it includes hardware companies such as Nvidia, Intel, and Qualcomm, it is not a pure-play software ETF.
The standout product is the iShares Expanded Tech–Software Sector ETF (IGV). While it holds 118 companies, nearly 58% of the portfolio is concentrated in the top 10 names. Major holdings include Salesforce (8.6%), Microsoft (8.2%), Adobe (8.1%), and Oracle (7.2%). IGV’s 59% year-to-date return far outpaces XSW (40%) and IGPT (27%).
Some analysts estimate that if just 20% of Microsoft’s 360 million Office 365 users adopt the paid Copilot service, Microsoft’s total revenue could rise by 10% next year. Notably, 77% of users report that once they start using Copilot, they find it difficult to stop. Adobe’s Firefly, meanwhile, generated approximately 3 billion images between its March launch and October. Because Firefly is trained largely on millions of images from Adobe Stock, the company’s proprietary library, it faces fewer challenges related to training costs and copyright issues.
Expectations are growing that 2024 will be the year of AI software. A wide range of services is likely to reshape everyday behavior, while equity markets may begin to form new investment themes around software-driven AI adoption. The monetization efforts that began in the third quarter should start to show up clearly in earnings reports released from late January next year.
Even amid persistent concerns about economic slowdown, history suggests that investing in clear and durable growth themes remains one of the most reliable ways to outperform the broader market. AI software may be entering precisely that phase.