AI Revenue Trends in 2025 and 2026: What the Numbers Tell Investors

The artificial intelligence sector has moved from a period of promise into one of measurable financial output. Revenue figures from the leading companies in hardware, software, and cloud services now provide a clearer picture of where value is being created, how fast it is growing, and where the risks remain. For investors and financial analysts, understanding these numbers in context is essential before drawing conclusions about the sector's long-term trajectory.

---

Revenue Growth Across the Major Players

The scale of revenue growth across AI-focused companies in 2025 and 2026 is notable by any measure. However, the growth stories differ significantly depending on whether a company is selling hardware, software subscriptions, or cloud computing capacity.

NVIDIA sits at the infrastructure layer of the AI economy. Its graphics processing units (GPUs) are the primary hardware used to train and run large AI models. In Q4 of its fiscal year 2026, NVIDIA reported data center revenue of $62.3 billion, a 75% increase compared to the same period the previous year, according to NVIDIA's investor relations page. Data center revenue refers to the income generated from selling chips and related hardware to cloud providers, enterprises, and research institutions building AI systems. This figure represents the clearest single data point showing the scale of capital flowing into AI infrastructure globally.

Microsoft has taken a different path, embedding AI capabilities into its existing software products and cloud platform Azure. As of April 2026, Microsoft's AI business had reached an annualised revenue run rate of $37 billion, up 123% from the prior year, according to reporting by GeekWire. A run rate is a projection of annual revenue based on the most recent period's performance. Microsoft's position is significant because it reflects demand not just from developers but from enterprise customers integrating AI into everyday business workflows.

OpenAI, the company behind ChatGPT and the GPT series of models, reported annual recurring revenue (ARR) of $3.7 billion by May 2026, representing a 340% year-over-year increase, according to WhoEarns. ARR measures the predictable, subscription-based revenue a company expects to receive over a year. While OpenAI's absolute revenue figure is smaller than Microsoft's or NVIDIA's, the growth rate reflects rapid adoption of its API (application programming interface, the technical gateway businesses use to build products on top of OpenAI's models) and its consumer subscription products.

Alphabet, the parent company of Google, reported that its Google Cloud division generated $13.6 billion in revenue during Q2 2025, a 32% year-over-year increase, according to Android Central's coverage of the earnings report. Google Cloud competes directly with Microsoft Azure and Amazon Web Services, and its AI-specific offerings, including access to Google's own Gemini models, have contributed to this growth.

Alibaba's Cloud Intelligence Group reported revenue of 41.6 billion yuan, equivalent to approximately $6.1 billion, for the January to March 2026 quarter. This represented a 38% year-over-year increase, according to the Associated Press. Alibaba's cloud business serves as the primary AI infrastructure provider across much of Asia, and its growth rate suggests that demand for AI compute is not confined to North American markets.

---

Sector-Specific Patterns

Looking at these companies by sector reveals distinct dynamics that matter for investors assessing risk and opportunity differently.

In AI hardware, NVIDIA's dominance is reflected in the concentration of its data center revenue. The company benefits from the fact that training large AI models requires enormous quantities of specialised chips, and NVIDIA's H100 and Blackwell series GPUs have become the standard tool for this work. The 75% year-over-year growth rate in data center revenue shows that demand has not yet plateaued, though the sustainability of that rate over multiple years is a separate question.

In AI software and services, OpenAI's 340% growth rate is the most striking figure in the dataset. However, growth rates at this level are more common in early-stage companies that are expanding from a smaller base. The more meaningful question for analysts is whether OpenAI can convert that revenue growth into sustainable profit margins, which it has not yet demonstrated.

In cloud services, both Alphabet and Alibaba are growing at rates in the 32% to 38% range. These are large, established businesses, so growth at that pace is meaningful. Cloud platforms benefit from AI demand because every AI workload, whether training a model or running inference (the process of generating outputs from a trained model), requires cloud compute capacity. This positions cloud providers as indirect but reliable beneficiaries of broader AI adoption.

---

Capital Expenditure and Infrastructure Investment

Sustaining these revenue figures requires substantial ongoing investment. The major cloud providers and AI companies have committed to significant capital expenditure programmes to expand data centre capacity, acquire chips, and build out networking infrastructure.

Microsoft, Alphabet, and others have each publicly signalled plans to increase infrastructure spending materially through 2025 and 2026. These investments are necessary to meet demand but also create a structural cost base that affects near-term profitability. For investors, the key question is whether the revenue growth being generated will eventually exceed the cost of building and maintaining the infrastructure required to deliver it.

NVIDIA is a beneficiary of this spending cycle, as each new data centre built by a cloud provider typically involves significant GPU purchases. This creates a feedback loop where infrastructure investment by cloud companies flows directly into NVIDIA's revenue.

---

Profitability Challenges and Sustainability Concerns

Not all revenue growth translates into profit, and this distinction is important for any financial analysis of the AI sector.

OpenAI and Anthropic, despite their rapid revenue expansion, remain unprofitable, according to reporting by The Atlantic. Both companies are reinvesting heavily in model development, safety research, and infrastructure. The cost of training frontier AI models (the most capable and resource-intensive models at the edge of current technology) runs into hundreds of millions of dollars per training run, and these costs do not diminish proportionally as revenue grows.

Some analysts have raised concerns about market saturation and whether current growth rates are sustainable over a multi-year horizon, as noted by The Atlantic. The argument is that a significant portion of current AI spending reflects an initial wave of enterprise experimentation and infrastructure build-out, and that spending growth may moderate once companies have established their core AI capabilities.

This does not mean the sector is in decline, but it does suggest that investors should distinguish between companies that have demonstrated a path to profitability and those that are still operating at a loss while scaling revenue. NVIDIA and Microsoft, for example, are both highly profitable. OpenAI and Anthropic are not, at least not yet.

---

Implications for Investors and Analysts

The revenue data across these companies points to several observations worth considering.

First, the hardware layer, represented primarily by NVIDIA, has captured a disproportionate share of the financial value generated by AI so far. When every major technology company is competing to build AI capacity, the company selling the tools to build it tends to benefit reliably.

Second, cloud platforms are growing steadily and are likely to remain durable revenue generators because AI workloads require compute infrastructure regardless of which AI models or applications become dominant. Alphabet and Alibaba's figures support this view.

Third, pure-play AI software companies like OpenAI are growing fast but carry more financial risk in the near term because they have not yet demonstrated consistent profitability. The 340% growth rate is attention-grabbing, but the underlying economics of model development remain challenging.

Fourth, geographic diversification matters. Alibaba's growth figures indicate that AI infrastructure demand in Asia is substantial and growing at comparable rates to Western markets. Investors focused exclusively on US-listed companies may be missing part of the picture.

---

Conclusion

The revenue data from NVIDIA, Microsoft, OpenAI, Alphabet, and Alibaba collectively confirm that the AI sector is generating real and growing financial returns. NVIDIA's $62.3 billion in data centre revenue, Microsoft's $37 billion AI run rate, and OpenAI's 340% ARR growth are not projections but reported figures from recent earnings periods and financial disclosures.

At the same time, the picture is not uniformly positive. Profitability remains elusive for several of the fastest-growing companies, and the sustainability of current growth rates is a legitimate open question. For investors and analysts, the most useful approach is to assess each company on the basis of its specific revenue model, cost structure, and competitive position rather than treating the AI sector as a single homogeneous opportunity. The numbers are large and the growth is real, but the distribution of that value across the sector is uneven.