The great sage of modern baseball, Yogi Berra, said: “It’s hard to make predictions, especially about the future.” His observation applies in all layers of life, especially in the world of high technology.
Use the power of the smart score of Tipranks, a data -driven tool to help you discover best performing shares and informed investment decisions.
Check your share choices and compare them with the recommendations of the Top of Wall Street analysts with your smart portfolio
In November 2022, OpenAi chatgpt released and showed us how generative AI can crush our notions about what machine learning can do. Last month, the Chinese startup Deepseek released its latest AI model, R1, and shook the world of the Shakers.
R1 reportedly surpassed the best AI models on the American scene – but what makes this development really important is how. The R1 model is Open Source and Deepseek has claimed that the training costs of the AI - the combined costs of the background technology and hardware that makes it possible – came to just $ 6 million. That is a small fraction of the costs related to some of the American AI systems affected in the US, such as Chatgpt, Gemini or Grok.
But is this a sign of a seismic shift in AI development, or only a one-off deviation? BairD analyst Tristan Gerra leans to the latter, which suggests that this can be more an exception than the start of a trend.
“Deepseek is high in novelty claims, but time will learn whether it will create a disturbance in current AI models and AI architectures -we strongly suspect that this does not. The claimed costs seem unrealistic, built on existing AI research and GPU architectures, so not remotely comparable to the amount of AI costs that have already been deployed without which deep key would not exist. Moreover, we see lower/lower costs Implementations such as feeding the further demand for AI systems, where price elasticity floats a proportional increase in GPU volumes, ”Gerra said.
With that in mind, Gerra is doubling on Nvidia (Nasdaq: NVDA) and Micron (Nasdaq: MU) as top AI. Both companies are market leaders, both enjoying the consensus assessments of the ‘strong buy’ analysts and both can benefit from AI’s ruthless growth. Let’s look at it closer.
Nvidia
First on our list, Nvidia, is not only the world’s largest semiconductor company – it is one of the largest listed companies in Wall Street. The shares of the chip maker lost a historic $ 590 billion in the aftermath of the Deepseek newspaper heads, the largest Dal Day Day ever recorded.
However, the meteoric rise of Nvidia is well documented and has even been taken into account the recent lock-related loss, the share has won 441% over the past three years. The profits are hand in hand with very strong sales, because the NVIDIA GPU chips have proved most popular with data center operators and AI application developers.
The release of R1 was a direct challenge for that dominance. According to Deepseek, their newest AI model does not use the latest versions of the top-end processor chips or relies so much on high-speed computing. As a result, it only makes a fraction of the power requirements that we see in established AIS. If R1 fulfills this hype, it would be the first of a new AI model, and Nvidia, which has based its success on dominating the expensive, high-end, AI-compatible processor chips, will be forced to fit quickly. That apparently for the sharpened investors.
Nvidia seems to be triggering this development on his part. The company published a statement in which Deepseek is praised for its impressive progress in AI technology, especially when demonstrating the test time scales of the model. The company also noted that Deepseek’s R1 is still based on earlier AI models -and they used more traditional AI training methods.
Assuming that Nvidia is correct, the market for high-end AI-compatible chips does not go away and Nvidia holds the leading position in that market, despite delays in the new Blackwell series. And last summer, Nvidia unveiled the successor to the Blackwell series, Rubin, a new GPU architecture planned for release in 2026.
In the last series of financial results, with regard to fiscal 3Q25, NVIDIA reported $ 35.08 billion in income, an increase of 93% on an annual basis and beating the forecast of $ 1.95 billion. The bottom line was Nvidia’s non-Gaap EPS of 81 cents 6 cents per share better than expected. The company ended the tax third quarter with $ 38.5 billion in cash and other liquid assets, giving the company deep pockets to endure a storm.
For Tristan Gerra, the key here is that Nvidia remains a solid player with a strong product line. In his description of Nvidia, the 5-star analyst says: “We strengthen our 2h gross margin estimation for Nvidia, given our expectations of a lower mix of GB200 versus B200/B300 this year based on problems at system level. Lowering our estimate of the Grace Unit this year by 1 m to 1.5 million units to take into account implementation of the challenges; This mix shift can reduce the top of our AI income estimate for the year. No changes to our EPS estimates: after an expected B200 channel push we expect a strong B300 question, while the Rubin platform should catalyze further performance/revenue growth. NVDA remains a best idea. “
As a ‘best idea’, NVIDIA shares receive an outperform (ie buy) rating from Gerra, whose $ 195 price target on the share is an increase of 49% in the current level in the following year. (To view Gerra’s track record, click here)
In general, Nvidia shares have a strong buy -rating of the consensus of the street analyst, based on 39 recent reviews, including 37 to buy and hold 3. The shares are currently being traded for $ 131.14 and have an average target price of $ 178.86, which suggests an upward potential of one year of 50%. (To see NVDA -Sharing forecast))
Micron technology
The following on our list is Micron, a semiconductor chip company based from Boise, Idaho. The company is best known for its memory chips, essential components of the AI tree. Micron’s high-bandwidth memory (HBM) chips are especially suitable for AI use.
In February last year, Micron announced that it collaborated with Nvidia and his HBM3E chips for the larger company to use in its AI-compatible semiconductors. More recently, Micron’s HBM4 was chosen for use in the upcoming Rubin platform of Nvidia. Micron claims that his HBM chips can match or surpass the performance of his rivals, while consuming 30% less power, giving them a competitive advantage in a difficult area.
Just like Nvidia above, Micron took a hard hit with the release of Deepseek’s R1. Like most of his colleagues, Micron has benefited enormously from the rise in AI in recent years. But by acting and offering the powerful memory chips that are required by AI data centers and high-speed computing, the company is vulnerable to the depth of Deepeek of a cheaper, less energy-hungry AI competition. The chips consume less strength, but retain their performance: the most important qualities needed to compete with the Deep Pech Model.
Just like Nvidia above, Micron comes to this fight with a record of successes. The income and income of the company have both been trending in the past quarters. In the last reported period, tax 1q25, Micron’s $ 8.71 billion in turnover increased by 84% on an annual basis and met street expectations, while the non-gaap win per share of $ 1.79 defeated the prediction of 2 cents. At the end of the tax first quarter, Micron had $ 6.7 billion in cash and kasequivalents.
If we check in again with BairD analyst Gerra, we cheerfully find him on Micron – with reference to the company’s exposure to Rubin from Nvidia as an important point. Gerra says about Micron: “We are Bullish on the slope of Rubin, who will be very positive for HBM content by extension. The larger mold size and the I/O count of Rubin means an increase of 30%+ in HBM capacity that is needed to support the flat volume. This dynamic is to support a constantly tight HBM offer in 2026 and, in our opinion, will benefit the price/margin. Micron is well anchored as an important HBM supplier from 2h25 while Samsung continues to struggle. Lower Grace Volumes is positive for DDR5. Micron remains a best idea. “
Looking ahead, the Analyst MU shares gives an outperform (ie buy) rating, with a price target of $ 130 that points to an advantage of 42% for the coming 12 months.
The strong buy -consensus assessment on Micron is based on 23 recent reviews that break down to 21 purchases and 2 possession. The shares are priced at $ 91.68 and their average price target of $ 136.1 is a bit bullisher than the BairD view, which suggests an advantage of a year of 48%. (To see Mu -sharing forecast))
Go to the best shares of Tipranks to find good ideas for shares that act with attractive ratings, a tool that unites all Equity insights of Tipranks.
Disclaimer: The opinions in this article are exclusively those of the recommended analysts. The content is intended to only be used for informative purposes. It is very important to do your own analysis before you make an investment.
