In January, Chinese startup Deepseek introduced an artificial intelligence (AI) chatbot. It quickly became the most downloaded free mobile application in the US. Models on a specific benchmark.
Openai spent more than $100 million in training on the GPT-4 model, and has access to more advanced chips than Deepseek, which has rocked Wall Street. Investors rushed to the conclusion that US companies were spending on AI infrastructure and Nvidia (NASDAQ: NVDA) Stocks fell 17% on news, and market value disappeared at $600 billion. This is the biggest day’s loss in US history.
Nvidia’s stock is still down around 9% since Deepseek rattled the market, and its stock is currently at its all-time high. However, Nvidia shareholders got some good news from Amazon (NASDAQ: AMZN) and alphabet (NASDAQ: Google). Please see details.
Many analysts praise Deepseek for his engineering breakthrough. The company reduced costs by reducing data processing with innovative training technology. However, several analysts are also wondering whether DeepSeek is completely approaching the infrastructure and associated costs.
For example, Dan Ives of Wedbush Securities wrote, “To say Deepseek was built for $6 million without Nvidia Next Generation Hardware is probably fictional.” Similarly, research firm Semianalysis reports that Deepseek not only has the next-generation Nvidia GPU, but also paid about $1.6 billion during training the model.
Importantly, cost-efficiency could actually increase the demand for NVIDIA Graphic Processing Units (GPUs) by enabling faster spread of artificial intelligence (AI) through the economy. Put another way, if new training methods reduce model development costs, more software companies will build AI products and increase the aggregate demand for NVIDIA GPUs.
Image source: Getty Images.
Last week, Amazon and Google’s parent alphabet reported fourth quarter financial results. The companies said capital expenditures will increase significantly in 2025 as supply is currently limited if AI infrastructure is concerned.
Amazon CEO Andy Jassy predicts that costs associated with developing AI models will continue to fall. However, he also said, “I think it would make it much easier for businesses to allow inject all their applications with inference and generation AI.” CFO Brian Olsavsky told analysts that demand for AI infrastructure could potentially exceed $100 billion in 2025, from $83 billion in 2025.
The story continues
Alphabet CEO Sundar Pichai provided similar insights. He says that more percentage capital costs are shifting to inference as training becomes cheaper, meaning DeepSeek’s breakthrough will not slow AI spending. In fact, CFO Ana Ashkenazi told analysts that capital expenditures will increase by $75 billion in 2025 from $52 billion in 2025, primarily due to investments in data centers, servers and networking.
Capital expenditure guidance from Amazon and Google’s pro-Alphabet supports the narrative that demand for NVIDIA GPUs can actually increase as training costs decrease as companies move resources into inference. Andy Jassy compared the situation where cloud computing sales actually increased over time, despite the fact that cloud services became cheaper.
“What happens is that companies spend much less per unit of infrastructure, which is very, very useful for their business. When the cost of units is low, they spend the total on technology. Masu.
Wall Street estimates Nvidia’s adjusted revenue will increase by 52% each year until 2026, which closes in 2026. That consensus makes the current rating look 50 times cheaper with adjusted revenue. Certainly, analysts have set very high standards, so if Nvidia misses those expectations, inventory could drop sharply.
Here’s the bottom line: Nvidia’s shareholders are worried that Deepseek’s drama will give you a deep breath. Many analysts believe that more efficient AI training technology will lead to greater demand for NVIDIA GPUs. And evidence from the largest hyperscale cloud companies supports that conclusion.
Have you ever felt like you missed a boat when buying the most successful stocks? If you do that, you’ll want to hear this.
In rare cases, a team of analysts issue “double-down” stock recommendations for companies they think they are trying to pop. If you’re worried that you’ve already missed the chance to invest, now is the perfect time to buy before it’s too late. And the numbers speak for themselves:
nvidia:If you invest $1,000 when it doubled in 2009,$336,677! *
Apple: If you invest $1,000 when it doubled in 2008, you’ll have $43,109! *
Netflix: If you invest $1,000 when it doubled in 2004, then $546,804! *
Currently, we are issuing “double-down” alerts to three incredible companies, and we may not have a chance like this anytime soon.
learn more “
*Stock Advisor will return as of February 3, 2025
John Mackey, former CEO of Amazon subsidiary Whole Foods Market, is a member of Motley Fool’s board of directors. Suzanne Frey, an executive at Alphabet, is a member of the board of directors of Motley Fool. Trevor Jennewine holds roles for Amazon and Nvidia. Motley Fools are located and recommend in Alphabet, Amazon and Nvidia. Motley Fools have a disclosure policy.
Nvidia’s stock investors just got great news from Amazon, and Google-Parent Alphabet was originally published by The Motley Fool
Adnan is a passionate doctor from Pakistan with a keen interest in exploring the world of politics, sports, and international affairs. As an avid reader and lifelong learner, he is deeply committed to sharing insights, perspectives, and thought-provoking ideas. His journey combines a love for knowledge with an analytical approach to current events, aiming to inspire meaningful conversations and broaden understanding across a wide range of topics.
