Nvidia and the issue of scaling

The rise of Nvidia, driven by AI’s “scaling law,” is now facing challenges.

This principle, which fueled the AI boom, suggests that feeding more data into larger models results in smarter systems. However, this idea is reaching its limits. Recent models from companies like OpenAI and Google show diminishing returns, despite massive investments in computing power.

AI leaders are rethinking their strategies.

OpenAI’s Ilya Sutskever, who once predicted AI would dominate global energy consumption, now says the “age of scaling” is giving way to the “age of wonder and discovery.” The challenge is no longer just building bigger models but enhancing their reasoning and adaptability after initial training.

For Nvidia, this shift is pivotal.

The company, which dominated AI hardware, now faces questions about the sustainability of scaling as a growth driver. CEO Jensen Huang argues that while traditional scaling may slow, the demand for advanced “inference” capabilities—how AI systems respond to queries—will require even more of its chips.

The broader AI industry, fueled by over $200 billion in annual investments from tech giants, is searching for breakthrough applications.

Nvidia’s future hinges on Big Tech’s ability to translate this spending into tangible returns, making the next phase of AI a high-stakes game for all.

________

One more thing, at Pfortner, we take communications privacy very seriously. We encrypt email, messaging and network communications to provide our clientele with uncompromised privacy.

If you need to protect sensitive communications, please see www.pfortner.co.za or send an email to info@pfortner.co.za, and we will get back to you.