Nvidia delivers record numbers again, its stock rises sharply. But there’s a catch lurking beneath the AI hype: many AI companies haven’t made any money at all. The risk of blisters is not over yet.
This means the chip giant remains the most important supplier in the AI boom. The data center division is “sold out until further notice,” confirmed CEO Jensen Huang. Demand in particular for the new Blackwell chips is “extremely high.”
The stock market was celebrating: Shares jumped more than six percent in after-hours trading. Nvidia targets even higher sales in the current quarter. Earnings per share rose to $1.30, above estimates of $1.26.
But as impressive as these numbers are, investors should be aware that Nvidia is just providing the material for an AI gold rush. It’s unclear whether the individual companies that want Nvidia’s chips will ultimately profit.
There is still money to be made with AI
Even though Nvidia is making billions in profits, many companies developing AI models are struggling very high costs. Prominent examples: Chat OperatorGPT OpenAI. Founder Sam Altman recently had to admit to investors that almost no profits were expected before the end of the decade. The reason: Training, further model development, and computing costs are enormous and will continue to rise for now.
A similar situation exists for other AI startups: They burn through capital while users expect free offerings and corporate customers are often hesitant to deeply integrate AI into their processes. The result: The majority of industries do not yet have a viable business model.
For investors, this means: Huge investments in AI have so far generated almost no revenue for software providers, but they are filling chipmakers’ coffers.
Why the billions of dollars flowing to big tech companies is at risk
Another thing that is increasingly worrying: Tech giants are increasingly pouring billions of dollars into start-up companies – which then buy computing capacity and chips from the same companies.
Example: Microsoft and Nvidia invested billions of dollars in Anthropic – which then purchased server capacity and hardware from Microsoft and Nvidia. A Growth cyclethat continues to drive Nvidia’s sales, but doesn’t say much about the sustainability of demand outside the tech bubble.
The more big tech companies finance each other, the more the AI sector becomes separated from the real economy. In other words, companies that will use AI productively.
Increased productivity will still occur for a long time
Big tech CEOs like Amazon boss Andy Jassy have announced that they want to replace thousands of jobs with AI. Just yesterday, tech billionaire Elon Musk said (again) that AI and humanoid robots will eliminate future poverty and make work a dream.
But in many companies, the reality is different:
- AI tools must be implemented painstakingly
- they require new processes and expensive infrastructure
- they often do not completely replace jobs
- the expected productivity only comes slowly
Important for investors: As long as AI doesn’t deliver measurable productivity gains, it’s unclear how long these billion-dollar investments will make economic sense.
Experts talk about “AI productivity paradox”: The technology already exists – but its benefits will only be seen at a later date. This has happened before: Even with the advent of PCs, the Internet, or smartphones, it took years for an economic breakthrough to become apparent.
What does Nvidia’s bull market mean for investors?
1. Nvidia will continue to make profits as long as the boom lasts
This group dominates the market, demand is high, and prices are affordable. In the short term everything looks like further growth.
2. AI software companies remain at risk
Many companies have yet to make money with AI, have unclear business models, and high costs. There is a risk of consolidation here, perhaps even a “mini AI crash”.
3. Circular transactions can distort the market
Investors should consider whether the earnings are truly consumption or industry based, or simply come from two-way deals. But this is complicated and, if there is any doubt, almost impossible to do.
4. The long-term potential is still very large
The AI revolution is coming. But this happened more slowly, more expensively and more chaotically than many expected.
5. Share widely
Rather than relying on superior AI, it is advisable to diversify broadly through ETF technology and look for solid companies that will actually use AI to increase returns later.
After Nvidia’s numbers, investors turned their attention to this
As soon as Nvidia’s numbers were digested, the market’s attention was already focused on the next push: there would be another push coming soon. US Federal Reserve cuts interest rates uncertain, which will not please the AI industry.
On December 9, the cloud giant Oracle quarterly figures before. Oracle is considered an important indicator of whether companies are truly investing in AI-powered cloud services or simply putting the brakes on them.
Lastly, they move US labor market data into focus, because the amount of work pending from Washington also has weight. This can provide information about how much leeway the Fed actually has to raise interest rates further and how resilient the US economy is.
After all: That Nvidia’s forecast should reassure investors. The company plans to earn at least $63 billion in the fourth quarter. That would be two billion more than the amount originally pushed for.
For investors, this means: Despite all the euphoria surrounding Nvidia, the next few days and weeks will remain dominated by interest rate fantasies, growth data, and attempts to classify AI euphoria.
How bubbles burst
If you want to know more about how dangerous an AI bubble is and what the warning signs are: Maximilian Gamperling from the team of experts at “Money + Mind” analyzes the situation in video. Just register for free and watch.
