After DeepSeek launched the "King Bomb" model before the Spring Festival, this round of shockwave has clearly continued to "resound" on Wall Street...

Tracking the performance of the "Magnificent Seven" in the US stocks.$Roundhill Magnificent Seven ETF (MAGS.US)$In the past five trading days, it has fallen by 2.4%, among which $Alphabet-A (GOOGL.US)$ due to poor earnings reports, it plunged by 9.2%.$Amazon (AMZN.US)$It has also dropped by 3.6%.

This is just the latest headwind faced by these tech giants in the USA. At the end of last month, DeepSeek triggered a massive sell-off among American tech stock investors, as fears of cheaper open-source large language models created panic, and the tech industry is still trying to recover from the sell-off.

However, company executives and industry experts generally believe that the development of DeepSeek should ultimately be seen as an upward catalyst for the entire tech industry, rather than a headwind.

A former OpenAI employee stated that the market should view the development of DeepSeek as a "victory"—because they have the potential to accelerate AI innovation and application.

Zack Kass, former marketing director of OpenAI, explained, "DeepSeek's R1 model is a breakthrough… We should all hope for scientific progress, and seeing key resource costs decrease is something we should also hope for."

Executives from major tech companies expressed similar views. Alphabet CEO Sundar Pichai praised DeepSeek's work as "outstanding" during the company's earnings call earlier this week, while Meta CEO Mark Zuckerberg remarked that DeepSeek has done "many novel things," and the company is "still digesting".

Therefore, in the eyes of many Wall Street professionals, DeepSeek's rise is, in fact, a critical turning point in the current AI revolution—compared to the soaring "Big Seven" over the past few years, it may be time for investors looking to profit from the next growth phase in AI to shift their focus toward massive tech giants and$NVIDIA (NVDA.US)$、 $Advanced Micro Devices (AMD.US)$ And beyond chip manufacturers...

At this crucial moment, many Wall Street professionals are coincidentally focusing on the same 'track': network communication stocks—betting on companies that provide foundational communication service facilities for AI applications.

Is the network gradually becoming a bottleneck for demand?

The reason behind this is actually not hard to understand: as the demand for computation increases significantly for AI models like DeepSeek's R1, the demand for high-speed network solutions will also increase day by day.

Chun Him Cheung, an Asia Emerging Markets Forex Strategist at Bank of America Global Research, wrote in a report to clients this week that the reduction in computing costs should lead to 'broader AI applications,' and demand should be 'strong.'

T. Rowe Price Technology Stock Strategy Portfolio Manager Tony Wang also stated that he believes that betting on this field is 'well-positioned.' 'The network is definitely the bottleneck for providing AI infrastructure, and we need to see more innovation at this level,' Wang remarked.

Analyst Ruben Roy from the USA-based independent investment banking and financial services company Stifel also believes that there is room for growth in this field, as he sees DeepSeek's R1 model as a driving force behind the global demand for robust and high-speed network infrastructure.

Roy believes that as the efficiency of AI workloads continues to improve, telecom system suppliers$Ciena (CIEN.US)$and Hardware manufacturers$Coherent (COHR.US)$and$Celestica (CLS.US)$will become the biggest beneficiaries.

The underlying channel for mobile data will become more challenging... with interface speeds increasing, the challenges of data migration may continue to become more complex, Roy explained.

He indicated that, as "data related to AI leaks out from the datacenter into enterprises and consumers using AI," Ciena has begun to see an increase in demand from hyperscale enterprises. Furthermore, Celestica, which manufactures switches and servers for hyperscale enterprises, will also benefit as the industry transitions into "custom computing — custom chips that hyperscale enterprises are designing and developing."

Morgan Stanley research analyst Meta Marshall is bullish on AI networking companies.$Arista Networks (ANET.US)$Arista Networks is a company focused on developing, marketing, and selling cloud networking solutions, covering switching and routing platforms as well as related network applications.

Marshall wrote in a recent earnings reports release that the company's stock price is now more attractive after the recent sell-off triggered by DeepSeek. She has a rating of shareholding on the stock ahead of the company's earnings report scheduled for February 18.

The three major telecom operators in China have launched DeepSeek.

It is worth noting that, on one hand, DeepSeek's free open-source strategy requires more bandwidth resources, while on the other hand, the development of AI will also feed back into the telecommunications and cloud computing service sectors. Last week,$CHINA TELECOM (00728.HK)$、$CHINA UNICOM (00762.HK)$、$CHINA MOBILE (00941.HK)$The three major telecommunications operators have collectively announced full access to DeepSeek, marking a new height in the integration of AI technology and the telecommunications industry.

As a result, CHINA TELECOM, China United Network Communications, and CHINA MOBILE all saw significant stock price increases in the A-share market on Monday.

It is reported that CHINA TELECOM has made DeepSeek fully available through its Tianyi Cloud, providing full-process services from deployment to inference and fine-tuning. Users can experience this through the Tianyi Cloud intelligent computing product system, which includes Xirang - Research Assistant, Tianyi AI Cloud Computer, Mole Community, Xirang Intelligent Computing Platform, and GPU Cloud Host/ Bare Metal. Additionally, Tianyi Cloud's AI Community - Mole Community has also launched an experience space for DeepSeek.

The China United Network Communications Cloud also realizes compatibility with domestic and mainstream computing power configurations for various specifications of the DeepSeek-R1 model based on the Starfall platform, accommodating both privatization and publicization scenarios, and can be pre-deployed in over 270 backbone cloud pools nationwide, fully integrating with products such as the China United Network Communications Cloud Desktop and Programming Assistant.

CHINA MOBILE's Mobile Cloud has fully launched DeepSeek, achieving complete version coverage, full size adaptation, and all functional usage. The 13 intelligent computing centers across the country operated by CHINA MOBILE have all implemented the aforementioned capabilities, allowing users to choose any intelligent computing resource pool for deployment, distillation, and intelligent agent orchestration. Additionally, Mobile Cloud seamlessly integrates DeepSeek into the Mobile Cloud Intelligent Agent Platform and customizes computing power solutions for the DeepSeek-R1 model, providing high cost-performance computing power and development environments for Internet companies and universities.

Many industry insiders have stated that the integration of DeepSeek has built a precise user behavior analysis system for operators, effectively mining high-value data assets based on multi-dimensional demand insight capabilities. This technological empowerment can not only optimize service response efficiency and quality but also achieve personalized product and service recommendations through user profiling, ultimately enhancing user stickiness and satisfaction.

Editor/new