The chart below depicts an extraordinary surge in Nvidia’s quarterly revenues, soaring to around $25 billion—a more than five-fold increase since early 2024.
Technology behemoths like Google and Microsoft are observing these numbers with amazement. Analysts suggesting that Nvidia has peaked might need to reconsider their forecasts.
ADVERTISEMENT
CONTINUE READING BELOW
“Mainstream commentators are often blinded by the idea that it’s all about predicting the next word. They only see hype and a continuation of the status quo; at most, they consider the possibility of another major technological shift akin to the internet,” comments Leopold Aschenbrenner in his essay series titled The Decade Ahead.
They will soon awaken. What is about to unfold is both exhilarating and potentially concerning.
Artificial ‘superintelligence’
AI is advancing towards AGI (Artificial General Intelligence) much faster than anticipated, allowing machines to surpass human cognitive abilities. Beyond AGI lies ASI, or artificial superintelligence.
Nvidia is not the only player in the AI chip sector. Google has rolled out its Tensor Processing Units to bolster machine learning tasks.
Nvidia possesses over 70% of the AI chip market, with rivals Intel and Advanced Micro Devices eager to diminish this enormous market share.
There’s no denying Nvidia’s substantial edge—often referred to as a moat—but its attractive gross margin of 78% is bound to invite intense competition, which seems imminent.
The AI chip market is forecasted to reach $400 billion in annual revenue by the end of the decade, bringing in many new players with specialized chips designed for specific functionalities.
Read/listen:
WEF forecasts rapid growth in AI, big data, and cybersecurity jobs
Nvidia and its partners enhance SA unit trusts
Magic and Warhammer: The impact of wide-moat intellectual property
AI is becoming less expensive at an unprecedented rate compared to previous disruptive technologies.
“The cost of operating artificial intelligence models with similar performance has been halving every four months—a trend we expect to persist throughout this decade. In contrast, Moore’s Law in semiconductors has historically halved costs every 18-24 months, indicating that the AI revolution is advancing 4-6 times faster,” states ARK Invest.
Collaborations between Microsoft and OpenAI are rumored to involve a $100 billion data center revolving around an AI supercomputer named Stargate, scheduled for launch in 2028.
Major investments are being funneled into similar endeavors across the US, China, Dubai, and beyond. Microsoft and Google plan to spend more than $50 billion each in capital expenditures in 2024, with Meta not far behind. While not all this spending will go towards AI, the trend is unmistakable.
Source: Best Investing
According to Leopold Aschenbrenner, companies are more inclined to make significant AI investments if they believe the economic returns justify it.
OpenAI reported its revenue doubling to $2 billion over the six months ending February 2024. If this trend of doubling every six months continues, it could reach $10 billion by early 2025.
ADVERTISEMENT:
CONTINUE READING BELOW
Microsoft is already projected to generate an additional $5 billion in revenue from AI.
This indicates that the substantial investment in AI is poised to yield quick returns. As companies like Microsoft, Google, and Meta start to derive a considerable portion of their revenues from AI—expected within a few years—the demand for capital will inevitably rise.
If Microsoft can convince a significant share of its 350 million paid subscribers to pay an extra $100 a month for an AI add-on, revenue could increase dramatically.
“The implications are monumental. This would establish AI products as the primary revenue driver for the largest corporations in America and mark their most significant growth sector. Revenue growth projections for these companies would skyrocket,” continues Aschenbrenner.
“Stock markets would react accordingly; we may soon witness the arrival of the first $10 trillion company. At that point, big tech would likely be eager to invest heavily, with each firm throwing hundreds of billions [at minimum] into further AI expansion. We could also see our first corporate bond issuance surpassing a hundred billion dollars.”
The power constraint
The main bottleneck in the AI race is power supply. Power generation in the US has only risen by 5% over the last decade; however, a planned 100-gigawatt AI cluster could consume about 20% of the total US electricity supply by itself.
To meet the growing demands of AI, tech companies may need to venture into power generation or acquire aluminum smelters for their large power contracts.
While clean energy is desirable, a Trump-led administration might favor US supremacy in the AI race to prevent China or Middle Eastern countries from gaining the upper hand. This might involve leveraging extensive natural gas reserves while scaling back on green energy initiatives.
The stakes of falling behind in AI are alarming at the state level.
Superintelligence holds the power to be weaponized. Rogue states could potentially threaten annihilation or sabotage competing nations, undermining any nuclear deterrents. Countries that gain even a slight advantage in AGI will have a significant competitive edge.
Aschenbrenner argues that the US must secure its position in this race or risk losing AGI superiority to authoritarian regimes.
“These clusters can be established in the US, and we must coordinate our efforts to make this happen domestically. National security must take precedence over the allure of unregulated capital from the Middle East, complex regulations, or even commendable climate commitments.”
South Africa is currently lagging in this race. The considerable power demands of this developing sector cannot be satisfied in a country that has struggled to ensure a consistent electricity supply for most of the past decade, with the last year being an exception.
Stay informed with Moneyweb’s extensive finance and business news on WhatsApp here.
