Is the World’s Hottest Company Losing Its Mojo?

A year ago, Nvidia stood out as one of Wall Street’s most sought-after stocks. The semiconductor giant was already a leader in the video gaming graphics sector and had inadvertently made waves in the cryptocurrency mining space. It successfully transitioned to dominate the ultra-high-end AI processor market, soaring to become the world’s most valuable company briefly in June 2024.

As businesses globally announced significant investments in AI data centers and processing power, Nvidia emerged as a prime beneficiary. This success brought considerable weight to key market indexes and attracted a flood of investments from retail shareholders and super funds alike.

However, in January, the market experienced shockwaves due to the sudden rise of Chinese AI company DeepSeek, which, until then, had been relatively unknown. DeepSeek unveiled a conversational AI model it claimed could match the reasoning capability of leading American models, all at a fraction of the cost. This revelation caused Nvidia's market cap to plummet by $938 billion in a single day. While there was a slight recovery in February, the stock faced another significant decline afterward.

Earlier this month, Nvidia’s CEO, Jensen Huang, attempted to reassure investors of the company’s robust future during the GPU Technology Conference (GTC) keynote and in various investor interviews. He emphasized that while Nvidia continues to generate impressive profits—over $3,500 every second thanks to its booming AI and data center tech—the overall growth of these profits is becoming a concern.

Investors are increasingly questioning whether companies will sustain or ramp up their investment in AI hardware amid changing economic conditions. Additionally, the competition posed by DeepSeek adds trepidation, suggesting that affordable AI might soon be more accessible. On top of this, worries persist regarding potential cost increases due to U.S. tariffs affecting manufacturing and shipping.

At the GTC, Huang addressed these issues directly. He contended that AI would become the core operating system across all industries, and that despite economic uncertainties, companies would see increased AI investment as a primary growth channel. He asserted that firms would continue to purchase from Nvidia, as even the tech giants have struggled to produce chips rivaling Nvidia's latest Hopper line. He also noted that the new Blackwell generation had already secured more orders from cloud providers than Hopper did at a similar stage in its lifecycle.

Regarding DeepSeek, Huang acknowledged its R1 model but explained that its perceived advantages were misjudged. The processing power required for such advanced reasoning models would necessitate more computing resources than traditional models, leading to higher demand for Nvidia chips.

Another significant development is the launch of a Superpod composed of Blackwell Ultra units, yielding 288 CPUs, 576 GPUs, and a staggering 11.5 exaflops of computing power, complete with 300 TB of memory.

Huang highlighted that the industry would require “100 times more AI computing power than anticipated last year,” and the spending on data centers is rising faster than analyst forecasts have suggested.

On the topic of tariffs, Huang expressed optimism, stating that their impact would be negligible and short-lived, as Nvidia shifts its crucial manufacturing operations to the United States. Some analysts speculate that the tariffs could even offer Nvidia a competitive edge.

Richard Clode, a portfolio manager at Janus Henderson, pointed out that current U.S. semiconductor export restrictions essentially cap AI compute levels for China, thereby limiting future Chinese AI models while global standards are likely to increase.

Nevertheless, this year’s Nvidia announcements received a noticeably different reception compared to the previous two years, which were marked by intense excitement regarding AI applications. For instance, merely mentioning Dell at the 2024 GTC led to a surge in that company’s stock. In contrast, an announcement about supplying chips to General Motors failed to inspire notable movement in GM’s stock.

In the past, Huang painted a vision of immense economic and social progress stemming from AI advancements; however, despite substantial investments and announcements from leading names like OpenAI, Amazon, Google, and Microsoft, tangible benefits seem far from realization.

Nvidia’s chips continue to hold a pivotal role in the technology ecosystem. As noted by Wedbush analyst Dan Ives, they are the indispensable ‘new oil’ or ‘gold’ driving the foundation of AI technology.

Huang has reiterated the need for greater industry investment and a broader adoption of AI technology to unlock these anticipated benefits. Despite the drop of over 10% from its January peak and the loss of more than $1 trillion in market value, questions linger: Was the prior AI hype unfounded, and will Nvidia revert to a stable valuation based on its formidable products? Or is a resurgence in sight as competition and trade concerns ease?

Alec Young, the chief investment strategist at Mapsignals, remarked that the market appears skeptical about Nvidia's growth potential, reflecting doubts about whether projected growth will occur given the company’s current high valuation.

On a brighter note, demand for the Blackwell chip is on the rise, and no substantial competitors are on the horizon. A new Blackwell Ultra chip is set to launch later this year, matching the 20 petaflops of AI performance while offering approximately 50% more memory. Next year, Nvidia plans to introduce a new architecture named Rubin, aimed at delivering 3.3 times the performance of Blackwell.

These advancements are strategically designed for use in large-scale data centers, but Nvidia is also catering to smaller operations by releasing desktop PCs equipped with Blackwell chips, starting around $5,000.

In conclusion, according to analyst Dan Ives, Nvidia’s crucial role in the tech sector remains undeniable, with every dollar spent on Nvidia products generating an estimated $8 to $10 in multiplier effects across the tech ecosystem, benefiting areas like hyperscalers, software development, data center expansion, cybersecurity, and energy demand over the projected $2 trillion in AI capital expenditure expected in the next three years.

Nvidia, AI, Competition