JPMorgan analysts Claudia Hueston and Pedro Martins just dropped a major AI ripple effect report, and the verdict is clear: DeepSeek's latest LLM is shaking up the tech hierarchy.

With the promise of slashing training and inference costs, some companies stand to gain big, while others might be in for a rough ride.

Winners: The AI Powerhouses

The bullish camp — tagged under JPMorgan's DeepSeek Positive Impact — is a who's who of AI titans and cloud leaders. Alphabet Inc GOOGL GOOG, Amazon.com Inc AMZN and Nvidia Corp NVDA top the list, unsurprisingly, given their AI ambitions.

But the wave also lifts Broadcom Inc AVGO, Marvell Technology Inc MRVL and Micron Technology Inc MU — all crucial to AI infrastructure. Even software players like Snowflake Inc SNOW, Databricks rival Confluent Inc CFLT and GitLab Inc GTLB look poised to benefit from AI-powered cost efficiencies.

The key theme? AI is moving from experimental to execution mode, and companies that enable inferencing and data management stand to gain.

Read Also: Sundar Pichai Downplays DeepSeek Threat — Google’s Gemini Still Leads In Cost, Performance, And Latency

Losers: The Hardware Setback

On the flip side, JPMorgan's DeepSeek Negative Impact list puts companies like Intel Corp INTC and Oracle Corp ORCL in the danger zone. With cheaper and more efficient AI models, legacy hardware players could see slower demand for expensive, high-power training chips.

Even industrial giants like Caterpillar Inc CAT and Cummins Inc CMI might take a hit if AI-driven automation undercuts traditional heavy machinery investments.

The Takeaway: AI's Next Economic Shift

While hyperscalers are still throwing money at AI hardware, DeepSeek's disruption could eventually tilt the value proposition toward application-layer innovation.

Companies that enable seamless AI adoption—think Salesforce Inc CRM, Shopify Inc SHOP and CrowdStrike Holdings Inc CRWD—could see an acceleration in demand, while those overly reliant on expensive, legacy infrastructure may need a strategy refresh.

Read Next:

• AMD CEO Lisa Su On DeepSeek’s AI Disruption: ‘Training And Inference With Less Infrastructure Is A Good Thing’ — But GPUs, Not ASICs, Will Still Power The Future Of AI Chips

Image: Shutterstock