Morgan Stanley: Generative AI is expected to achieve positive returns this year, with revenue likely to exceed one trillion dollars by 2028

In this wave of artificial intelligence frenzy, tech giants have reaped trillions of dollars in market value growth, but the massive investments in AI have also raised concerns among investors about their return on investment.

Morgan Stanley's latest report on Wednesday pointed out that 2025 will be a watershed year for the return on investment (ROI) of Generative AI (GAI) to turn positive, and by 2028, this technological revolution is expected to generate over $1 trillion in revenue.

Since the launch of ChatGPT in November 2022, the total market capitalization of the Mag7 tech giants has soared by $9.2 trillion, an increase of 118% (compared to a 47% rise in the S&P 500 during the same period). This surge largely reflects investors' optimism about a new cycle of AI technology.

However, skepticism has never ceased: NVIDIA's data center GPU business has generated $83 billion in revenue over the past 12 months, with companies spending $500,000 annually on a single AI server. As global data center power demand is expected to double to 180GW within four years, when will this massive expenditure turn into real revenue?

2025 May Welcome ROI Turning Point, 2028 AI Revenue Expected to Exceed $1 Trillion

Morgan Stanley stated that after considering the comprehensive benefits brought by AI in the enterprise software and consumer internet markets, the ROI of generative AI will turn positive in 2025.

Morgan Stanley's model estimates that GAI will generate $153 billion in revenue in 2025:

Of which, $59 billion will come from improved productivity in enterprise software, with the potential productivity impact of generative AI solutions on the U.S. labor market estimated at $3.3 trillion; $94 billion will be driven by increased conversion rates in e-commerce, higher advertising revenues, and income from wearable devices, with generative AI search queries and social generative AI on Meta, YouTube, TikTok, Pinterest, Snapchat, and Reddit generating $56 billion.

On the expenditure side, capital expenditures in semiconductors and non-chip data centers are expected to reach $302 billion,

$128 billion will be allocated to investments in hardware, networking, cooling, memory, etc., and $174 billion will be for semiconductor expenditures, including GPU/ASIC sales. The annual cost of operating these infrastructures is expected to be $102 billion, which includes $77 billion in depreciation costs, $23 billion in electricity costs, and $2 billion in labor costs.

Overall, Morgan Stanley stated:

A gross margin of 34% marks the industry officially crossing the breakeven line, with generative AI expected to generate $51 billion in returns in 2025.

AI Revenue Expected to Exceed $1 Trillion by 2028

Looking ahead to 2028, Morgan Stanley stated that the total revenue from generative AI is expected to approach $1.1 trillion:

Among them, enterprise software contributes $401 billion, and the potential productivity impact of generative AI solutions on the U.S. labor market is $6.4 trillion; consumer internet-related spending contributes $683 billion, with generative AI's efficiency improvements amounting to $469 billion and generative AI advertising revenue at $197 billion.

Morgan Stanley predicts that capital expenditures for generative AI will reach $556 billion by 2025, with semiconductor spending at $280 billion and non-chip data center-related spending at $276 billion. Overall, by 2028, the monetization of generative AI is expected to generate $722 billion in returns before deducting sales and marketing, R&D, and other expenses, with a return rate expected to rise to about 67%.

In addition, several tech giants are expected to benefit from the AI wave:

Microsoft's Azure AI services and Copilot toolkit have formed an annual revenue scale of $13 billion, expected to swell to $78 billion by 2028; Amazon, with retail algorithm upgrades, can leverage a 4.5% GMV growth for every 10% increase in conversion rate; Meta's AI content engine could contribute an additional $101 billion to its revenue by 2028