U.S. semiconductor giant NVIDIA (NVDA Quick QuoteNVDA - Free Report) plunged roughly 17% (approximately $593 billion) on Monday after Chinese Artificial Intelligence (AI) company DeepSeek showcased models that cost a fraction of what U.S. tech giants are spending on model development.

Apart from NVIDIA, AI chip and component providers, Broadcom (AVGO Quick QuoteAVGO - Free Report) , AMD and Amphenol (APH Quick QuoteAPH - Free Report) fell 17.4%, 6.37% and 12.57%, respectively. Hyperscalers, including Microsoft (MSFT Quick QuoteMSFT - Free Report) , Alphabet and Oracle, dropped 2.14%, 4.2% and 13.79%, respectively.

Nevertheless, we believe APH, AVGO and NVDA will continue their momentum amid the disruptions caused by the DeepSeek’s low-cost AI models. In fact, the dip offers great buying opportunity in our view.

DeepSeek Models Performing Similar to U.S. Tech Giants

DeepSeek’s models are gaining attention due to efficient performance at low costs and energy use. Its latest open-source model that was released last week, DeepSeek-R1-Zero and DeepSeek-R1, showed similar performance to that of Microsoft-backed OpenAI’s reasoning models and Meta Platforms’ (META Quick QuoteMETA - Free Report) Llama models on leading benchmarks such as the AIME 2024. 

Earlier in December 2024, DeepSeek showcased a different model (V3) that apparently costs $5.6 million to train and develop on NVIDIA’s low-capacity H800 chips. DeepSeek uses less than 2,050 GPUs, while Meta Platforms reportedly uses 16000 chips of NVIDIA H100s to train its Llama 3 model.

Meanwhile, DeepSeek-V3 powered AI chatbot surged to the top of Apple’s App Store downloads. The V3 model is comparable to Meta Platforms’ Llama 3.1 and OpenAI’s 4o.

Are U.S. Tech Giants Spending Too Much on AI?

DeepSeek’s low-cost models are definitely raising questions about the spending trend of U.S. tech giants, which are expected to spend billions on model development and AI infrastructure. 

OpenAI is estimated to have spent close to $3 billion on training models and data in 2024, per data from DeepLearning.ai. According to Reuters, U.S. tech giants are estimated to spend roughly $250 billion on AI infrastructure, with META alone planning to spend between $60 billion and $65 billion in capital expenditures on AI in 2025.

Moreover, OpenAI, SoftBank and Oracle, along with the Trump administration, announced the Stargate joint venture, which intends to invest $500 billion over the next four years building new AI infrastructure.

Top Semiconductor Picks

We expect U.S. tech giants to continue their spending momentum as models become more complex and deployment time reduces drastically. Amphenol, Broadcom and NVIDIA are three semiconductor stocks that are well-positioned to benefit from the increasing deployment of AI and Generative AI (GenAI) technology.

Amphenol’s diversified business model lowers the volatility of individual end markets and geographies. This Zacks Rank #1 (Strong Buy) stock rides on a strong portfolio of solutions, including high-technology interconnect products. You can see the complete list of today’s Zacks #1 Rank stocks here.

Expanding spending on both current and next-generation defense technologies bodes well for APH’s top-line growth. Apart from Defense, APH’s prospects ride on strong demand for its solutions across Commercial Air, Industrial and Mobile devices. For the first quarter of 2025, Amphenol expects a mid-to-high single-digit range increase in Commercial Air sales.

APH shares have jumped 34.1% in the trailing 12-month period. The Zacks Consensus Estimate for the company’s 2025 earnings has increased 3.7% to $2.25 per share over the past 30 days and indicates 19.05% growth over the 2024’s reported figure.

Meanwhile, Broadcom is benefiting from strong demand for its networking products and custom AI accelerators. Strong demand for Broadcom’s application-specific integrated chips, designed to support AI and machine learning, aids top-line growth. The acquisition of VMware has benefited Infrastructure software solutions.

Broadcom’s expanding AI portfolio, along with a rich partner base, indicates solid top-line growth potential. This Zacks Rank #2 (Buy) stock expects the momentum in AI connectivity to remain strong as more hyperscalers deploy Jericho3-AI in their fabrics.

AVGO shares have appreciated 68.1% in the trailing 12-month period. The Zacks Consensus Estimate for the company’s fiscal 2025 earnings has increased a couple of cents to $6.31 per share over the past 30 days and indicates 29.57% growth over fiscal 2024’s reported figure.

We expect NVIDIA to continue benefiting from accelerating usage of its GPUs, including Hopper and Ampere architectures. In fact, we believe DeepSeek’s success in using NVIDIA’s low-capability chips to develop advanced models validates the latter’s technological superiority. 

Moreover, a surge in hyperscale demand and higher sell-ins to partner across the Gaming and ProViz end markets following the normalization of channel inventory are acting as tailwinds for NVIDIA, which currently has a Zacks Rank #2.

NVDA shares have appreciated 93.1% in the trailing 12-month period. The Zacks Consensus Estimate for the company’s fiscal 2026 earnings has increased five cents to $4.21 per share over the past 30 days and indicates 43.15% growth over fiscal 2025’s estimated figure of $2.94 per share.