The world of artificial intelligence is accelerating at a dizzying pace, and at the center of this technological whirlwind are two companies making headlines for their outsized roles: Nvidia, the chipmaking titan, and IREN Limited, a rising star in AI infrastructure. As the AI boom continues to reshape the global economy, both firms are racing to meet surging demand for high-performance computing, each with their own distinctive strategies and challenges.
On August 25, 2025, Nvidia is set to report its fiscal second-quarter earnings—a milestone that marks two years since the generative AI revolution began to transform the company. According to CNBC, Nvidia’s revenue has more than tripled and profits have quadrupled since late 2022, when OpenAI’s ChatGPT ignited a new era of generative AI. In that time, Nvidia’s stock price has soared twelvefold, closing most recently at $177.99. Last month, the company became the first to hit a $4 trillion market cap, a testament to its central role in the AI gold rush.
But the story is about more than just numbers. As S&P Global’s Melissa Otto told CNBC, “The assumptions and performance of Nvidia really dictates what the market is going to start to price into the AI trade, and that whole AI trade has essentially been driving the market this past year.” Nvidia now makes up about 7.5% of the S&P 500, underscoring its influence over the broader tech sector.
Yet, even as growth remains robust, the pace has slowed. After five straight quarters of triple-digit expansion through 2023 and 2024, Nvidia’s revenue growth dipped to 69% in the fiscal first quarter of 2025. For the upcoming second quarter, analysts expect a 53% year-over-year jump to $45.9 billion, with data center revenue accounting for a staggering 88% of total sales. That’s a clear sign of AI’s dominance in Nvidia’s business model.
Much of this momentum is powered by Nvidia’s Blackwell architecture, the company’s flagship AI chip line. In May 2025, Nvidia reported that Blackwell products had reached $27 billion in sales, accounting for about 70% of data center revenue—up sharply from $11 billion the previous quarter. Blackwell Ultra GPUs, the next evolution, are expected to begin shipping in the second half of 2025. As Ryuta Makino of Gamco Investors noted to CNBC, “It solidifies that hyperscaler spending is still very strong with the Blackwell ramp.”
The Blackwell chips are coveted by the world’s largest internet companies and cloud providers, including Microsoft, Google, Amazon, and Meta. As more of these chips are installed, their superior computing power is expected to enable companies like OpenAI and Anthropic to build even more sophisticated AI models. OpenAI’s latest GPT-5, for instance, was trained on Nvidia’s previous-generation Hopper chips, not yet on Blackwell.
But with great success come new challenges. Nvidia’s rapid ascent has left it heavily reliant on so-called hyperscalers—giant tech firms that account for a large chunk of its sales. Last year, 34% of Nvidia’s total sales came from just three unnamed customers, according to CNBC. This concentration leaves Nvidia vulnerable to shifts in the macroeconomic environment and the unpredictable AI industry. OpenAI CEO Sam Altman recently cautioned that “investors as a whole are overexcited about AI,” even suggesting a potential bubble. Still, as OpenAI CFO Sarah Friar told CNBC, the company “constantly” lacks enough computing power, so demand remains sky-high for now.
Geopolitics have also entered the fray. Nvidia recently negotiated a deal with former President Donald Trump to regain access to the lucrative Chinese market, agreeing to pay 15% of its China chip revenue to the U.S. government in exchange for export licenses for its H20 AI chip. Trump revealed he’d initially asked for 20%, but Nvidia CEO Jensen Huang bargained him down. The H20 chip, worth an estimated $8 billion in potential second-quarter sales, was left out of Nvidia’s guidance due to ongoing export restrictions and pressure from the Chinese government for its cloud providers to use domestic chips from companies like Huawei.
While Nvidia’s dominance is clear, there’s a new player making waves in the AI infrastructure space: IREN Limited. On August 25, 2025, IREN secured 36-month leases for 2,400 Nvidia Blackwell B200/B300 GPUs and raised $550 million in convertible senior notes to fund its AI infrastructure expansion—all without diluting shareholders. According to AInvest, this non-dilutive capital strategy stands out in a sector where many competitors rely on equity financing, which can erode existing investors’ stakes.
IREN’s secret weapon? A unique blend of capital discipline and renewable energy. The company has locked in 2,910 megawatts of low-cost, 100% renewable energy at just $0.028 per kWh—30 to 50% below industry averages. Since energy costs typically make up 30–50% of AI cloud operations, this gives IREN a durable cost advantage. Its liquid-cooled data centers and direct-to-chip cooling technology further reduce operational expenses and boost margins.
AI cloud services are becoming IREN’s next major revenue engine. In Q3 2025, AI cloud revenue surged 33% year-over-year to $3.6 million, making up 10% of total earnings. Even more impressive, hardware profit margins in this segment reached 97–98%. By 2026, AI cloud is projected to become IREN’s dominant revenue stream, supported by a 4,300-GPU fleet and the completion of Horizon 1—a 50 MW liquid-cooled AI data center in Texas.
IREN’s growth is also fueled by strategic partnerships with Nvidia, ensuring access to the latest Blackwell chips. The Prince George campus can host up to 20,000 Blackwell GPUs, while the ambitious Sweetwater 2 project—a 2 GW data center hub in West Texas—is expected to be operational by 2028. This positions IREN to serve enterprise clients seeking hyperscale AI computing resources.
Financially, IREN is on solid ground. In Q3 2025, the company reported $148.1 million in revenue and adjusted EBITDA of $83.3 million, a 33% increase. Its Bitcoin mining business remains a stable revenue source, with a 50 EH/s hashrate and all-in cash costs of $41,000 per Bitcoin—among the lowest in the industry. Institutional ownership has climbed to 41.08%, with major holders like BIT Capital GmbH and BNP Paribas Financial Markets. Notably, IREN’s $550 million convertible notes offering was oversubscribed, signaling strong investor confidence.
IREN’s long-term vision is bold: to become a renewable-powered AI infrastructure leader. With projects like Sweetwater 2 and a strategic focus on capital-efficient scaling, the company is poised to capture a significant share of an AI infrastructure market projected to grow at a 37% CAGR through 2030. As AInvest put it, IREN’s “execution track record, institutional backing, and alignment with the next-generation computing paradigm make it a must-watch stock in the AI era.”
As the AI infrastructure race heats up, Nvidia and IREN exemplify different approaches to riding the wave. Nvidia’s technological leadership and global reach continue to set the pace, but IREN’s innovative capital strategies and green energy edge offer a compelling alternative. For investors and industry watchers alike, the coming years promise even more dramatic shifts as these companies—and the broader AI ecosystem—push the boundaries of what’s possible.
