Bull! Why is the AI Bull Market is Back?

The market previously sold off because investors feared that AI spending was rising faster than the profits it generated. The latest earnings season has started to answer that concern: demand remains strong across almost every layer of the AI supply chain, while AI investment is increasingly translating into revenue.

1. Cloud: AI is finally being monetised

Amazon, Microsoft and Alphabet reported accelerating cloud growth:

* AWS revenue increased 37%, its fastest growth in 18 quarters.

* Azure revenue rose 43%, while Microsoft 365 Copilot surpassed 30 million paid seats.

* Google Cloud revenue surged 82%, supported by AI infrastructure and enterprise AI solutions.

The market’s main concern was the enormous increase in AI capital expenditure. Higher capex pressures free cash flow and raises questions about how long it will take companies to earn an adequate return on their investments.

However, investors became more comfortable when cloud revenue, contracted backlog and AI-related sales accelerated alongside spending. In other words, capex is no longer rising without evidence of monetisation.

2. Semiconductors: Demand remains strong, but margins matter

AMD reported record revenue of US$11.5 billion, up 50%, while Data Centre revenue more than doubled to US$6.7 billion. Arm also delivered record quarterly revenue, supported by greater adoption of Arm-based processors in AI data centres.

The market was less satisfied with AMD’s gross-margin guidance of approximately 56%. Investors worried that early Helios system costs, product-mix changes and competition could limit how much of the revenue growth converts into profit.

The message remains positive for the industry: AI-compute demand continues to expand. The debate has shifted from whether customers still need AI chips to how profitably suppliers can meet that demand.

3. Memory and storage: Customers are securing supply years ahead

SK Hynix, Sandisk, Western Digital and Seagate all reported strong AI-related demand:

* SK Hynix delivered record results and signed multiple long-term supply agreements.

* Sandisk’s data-centre revenue increased 437%, with multi-year agreements expected to cover around two-thirds of capacity by FY2028.

* Western Digital said AI customers are negotiating storage agreements extending as far as 2031.

* Seagate’s quarterly revenue rose approximately 49%, with data-centre revenue up 57%.

The market’s concern is that the traditional memory cycle could eventually return. Higher prices encourage new capacity, potentially creating oversupply from late 2027 onward—particularly if smartphone and PC demand remains weak.

Investors are therefore questioning whether today’s exceptional pricing and margins can be sustained. Nevertheless, long-term contracts and customer deposits provide far better demand visibility than in previous memory cycles.

4. AI infrastructure: Spending is spreading beyond GPUs

Arista Networks, Lumentum and Nebius confirmed that AI investment is flowing into networking, optical connectivity and cloud capacity:

* Arista revenue rose 38%, supported by demand for high-speed AI networking.

* Lumentum revenue more than doubled as AI data centres adopted faster optical connections.

* Nebius revenue jumped 454% as customers demanded more AI-compute capacity.

The market welcomed the stronger revenue, improving margins and higher guidance. Remaining concerns include capacity bottlenecks, customer concentration and the substantial financing required to expand infrastructure quickly.

These results demonstrate that AI spending is broadening beyond GPUs into the equipment required to connect, power and operate complete AI clusters.

5. Software: AI is moving from experimentation to real usage

Palantir, ServiceNow and Cloudflare provided evidence that companies are paying for AI applications:

* Palantir revenue surged 93%, while US commercial revenue increased 149%.

* ServiceNow subscription revenue grew 24.5%, and the company raised its full-year outlook.

* Cloudflare revenue rose 36%, supported by AI workloads and increased usage of its Workers platform.

The market liked the accelerating adoption and stronger guidance. However, investors remain cautious about premium valuations, ongoing investment requirements and whether high growth can eventually produce sufficiently strong margins and cash flow.

Software earnings are especially important because they show that AI is creating business value—not merely generating demand for infrastructure.

Why did investors return?

The latest earnings collectively show that the entire AI value chain remains in demand:

Cloud companies are monetising AI → chip demand remains strong → memory and storage are being contracted years ahead → networking and optical infrastructure are expanding → software companies are converting AI into recurring revenue.

At the same time, July CPI rose only 0.1% month-on-month, while annual headline and core inflation moderated. Softer inflation reduced fears of another rate increase and lowered pressure on bond yields.

The combination of strong AI earnings and lower interest-rate risk encouraged investors to return to technology stocks, pushing the S&P 500 to another record high.

⸻

How long will the Bull Run continue?

Four Major Tests 2026

The earnings season confirms that AI demand remains strong. However, with the S&P 500 already at a record, the bull market now faces four major tests before the 3 November midterm elections.

Test 1: Will the Federal Reserve remain patient?

The 19 August FOMC minutes will reveal the degree of disagreement among Fed officials at the July meeting. Broad support for keeping rates unchanged would help stabilise rate expectations and Treasury yields.

The next decisive signals will come from:

* 4 September: August employment report

* 11 September: August CPI report

* 15–16 September: FOMC meeting

The employment and CPI reports will be the final major labour and inflation releases before the September decision.

Weakening employment alongside controlled inflation would support another pause. Conversely, stronger growth or renewed inflation could revive expectations that the Fed will resume tightening.

Test 2: Can inflation cool without causing a recession?

Two major economic reports will be released on 26 August:

* July core PCE inflation

* Revised second-quarter US GDP

Current estimates point to core PCE increasing approximately 0.2% month-on-month and around 3.3% year-on-year.

The most supportive combination for equities would be resilient GDP growth alongside moderating inflation—a soft landing rather than recession or stagflation.

A hotter core PCE reading would be negative because it could lift Treasury yields and pressure highly valued growth stocks. A sharp downward GDP revision would create the opposite concern: inflation may be cooling because the economy is weakening too quickly.

Test 3: Can Nvidia and Big Tech justify the AI spending boom?

Nvidia reports earnings on 26 August. Investors will focus on:

* Blackwell demand and deliveries

* Early visibility for Rubin

* Data-centre revenue growth

* Gross margins

* Supply constraints

* Cloud customers’ capital expenditure

* Whether AI investment continues translating into revenue

Nvidia is especially important because its results provide a direct read-through for the entire AI ecosystem—from cloud providers and memory suppliers to networking, optical and power-equipment companies.

The next confirmation will come during the third-quarter earnings season beginning in October. Microsoft, Amazon, Alphabet, Meta and other technology companies must demonstrate that higher AI capex continues generating faster cloud revenue, larger contracted backlogs and eventually stronger cash flow.

Test 4: Can external risks remain contained?

Ahead of the midterm elections, investors will continue monitoring:

* Oil prices and the Strait of Hormuz

* Tariff developments

* US fiscal policy and government borrowing

* Long-term Treasury yields

Another energy-price surge could lift inflation and complicate the Fed’s decision. Meanwhile, a breakout in the 10-year Treasury yield could trigger a valuation correction—particularly among technology stocks already trading on demanding earnings expectations.

Conclusion: What does the bull market need next?

The AI bull market has returned because earnings have validated demand throughout the value chain:

Cloud monetisation → semiconductor demand → memory and storage contracts → networking and optical expansion → enterprise software adoption.

However, strong AI demand alone may no longer be sufficient to drive the next leg higher. With the S&P 500 already at a record, further gains require three conditions:

1. Inflation continues cooling.

2. The US economy avoids recession.

3. AI-related earnings continue justifying the capex boom.

Corporate earnings and AI capital expenditure remain strong. If core PCE does not accelerate and Nvidia confirms continued AI demand with the Fed stays on hold, the bull market still has room to continue into the midterm elections.