Nvidia's $81.62B Quarter Confirms Semiconductor Era Over Software

Nvidia reported quarterly earnings Wednesday evening that gave concrete weight to Jim Cramer's repeated argument: semiconductors are the market's new center of gravity, and software has lost its pole position. The chip maker posted adjusted earnings of $1.87 per share on revenue of $81.62 billion, topping Wall Street expectations. On CNBC's "Mad Money," Cramer called the moment a turning point. "Semis are now in charge. Software is taking a back seat."

That statement is not a throwaway line. The generative AI boom has permanently reordered the hierarchy of technology sectors. Investors who treat the shift as a cyclical rotation risk mispricing the next decade. The numbers, the capital flows, and the enterprise adoption patterns all point in the same direction.

Nvidia's Earnings and the Sector Reordering

Nvidia's $81.62 billion quarter is the headline, the scale of that revenue tells the deeper story. Enterprises and cloud providers are spending heavily on AI compute infrastructure. The same companies that once allocated IT budgets to subscription-based software are now writing big checks for Nvidia GPUs, AMD accelerators, and networking silicon from Broadcom and Intel.

Cramer said the generative AI era has reshaped the old hierarchy. "It's a new era," he said. "Software's facing new competition from the much cheaper products you can develop yourself from AI, and...it's growing more slowly than the physical side of tech: semiconductors, hardware, the tools that allow for the artificial intelligence revolution."

Before AI, software-as-a-service (SaaS) was Wall Street's prized business model. Vendors like Salesforce and Adobe generated recurring revenue, high margins, and predictable growth. Investors paid premium revenue multiples for that stability. That world is cracking.

SMH Up 72%, IGV Down 12% – The Divergence is Structural

The performance data backs Cramer's claim. Year to date, the iShares Semiconductor ETF (SMH) has climbed roughly 72%. Over the same stretch, the iShares Expanded Tech-Software Sector ETF (IGV) has fallen about 12%. A gap of 84 percentage points is not a short-term trade. It reflects a fundamental change in where capital is deployed and where value accrues.

The beneficiaries of the shift are the companies supplying the infrastructure:

• Nvidia – the dominant AI chip supplier, now the most valuable company globally.
• AMD – gaining traction with its MI300 series accelerators.
• Arm – the architecture behind energy-efficient inference chips.
• Intel – chasing foundry and GPU ambitions, still a wild card.
• Broadcom – networking and custom ASIC design for hyperscale data centers.

Cramer's Charitable Trust owns shares of Arm, Broadcom, Nvidia, and Salesforce, so his view is informed by direct exposure. The portfolio's positioning itself reflects the thesis he articulated on air.

Why Generative AI Undermines the SaaS Model

The mechanism is straightforward. Traditional SaaS vendors sold convenience and integration. A company paid a monthly fee per user for Salesforce to manage customer relationships or Adobe for creative tools. Generative AI changes the calculus. A firm can now fine-tune a foundation model on its own data, run inference on Nvidia hardware, and build a comparable internal application. The upfront capital cost is higher, the variable cost per user is dramatically lower.

Cramer added that legacy software companies are not disappearing. Businesses will still use platforms from Salesforce and Adobe. Yet AI is making customers rethink how much they are willing to spend. The pricing power that software vendors once enjoyed is eroding.

Enterprise fear as a pricing pressure

"They've sown fear into the very fabric of the enterprise," Cramer said, referring to the AI model providers and hardware suppliers. That fear is rational. Any CFO can now look at a $10 million annual software subscription and ask whether an internal AI tool built with open-source models could do 80% of the work for a fraction of the cost. The answer is often yes.

Valuation Context: Alpha Scores Reflect Semiconductor Momentum

The Alpha Score from AlphaScala provides a quant view of the quality behind the narrative. For NVIDIA, the Alpha Score stands at 66/100, labeled Moderate. The score reflects strong fundamental momentum – high revenue growth, expanding margins, and dominant market share – tempered by elevated expectations already priced into the stock. For AMD, the Alpha Score is 59/100, also Moderate. AMD's position in the AI chip market is gaining ground, though it remains the clear number two. Investors can review the full profiles at the NVDA stock page and AMD stock page.

For context, the software-heavy indexes do not enjoy similar quant support. Many legacy SaaS names face compression on both multiples and growth rates, a dynamic explored in AlphaScala's earlier piece Why Capital Is Abandoning IGV for Semiconductor Growth.

What Confirms or Weakens the Semiconductor Thesis

Investors watching this theme need to track specific signals, not just macro sentiment. The following clusters separate confirmation from risk.

Confirming signals:

• Nvidia's next earnings – continued acceleration in data center revenue, especially as the H200 and Blackwell architectures ramp. A slowdown in quarter-over-quarter growth would be the first crack.
• Software earnings calls – mentions of longer sales cycles, higher churn, or reduced seat counts from Salesforce, Adobe, or Workday would validate the pricing power erosion thesis.
• Enterprise AI adoption data – rising percentages of companies building custom AI applications instead of buying off-the-shelf software, as surveyed by firms like Gartner or IDC.

Weakening signals:

• A cyclical downturn in chip orders – hyperscalers pausing or reducing data center expansion plans would hit semiconductor revenue growth directly.
• Software vendors successfully monetizing AI – if Salesforce or Adobe demonstrate that AI features drive higher average revenue per user without cannibalizing subscriptions, the thesis of a structural shift would be challenged.
• Regulatory intervention – export controls on AI chips to key markets, or new compliance costs that slow infrastructure buildout, would introduce headwinds.

Cramer was unequivocal about the direction. "The world has changed," he said. "We are not going back to the way things were. Not now. Not ever."

For now, the data aligns with his call. The SMH–IGV divergence is not a short-term trade. It reflects a reordering of which companies control the critical layer of technology. Investors who fail to adjust their framework risk owning the wrong assets in the next cycle. The broader stock market analysis and the NVIDIA profile provide additional context for tracking this theme.