--- title: "Jensen Huang Addresses Key Concerns: CoreWeave Debt, GPU Depreciation Period, and DeepSeek's High Gross Margin on Inference" type: "News" locale: "en" url: "https://longbridge.com/en/news/302644185.md" description: "Jensen Huang recently addressed three major focal points of AI investment: first, the reassessment of the GPU depreciation period from four years to seven or eight years, with high residual value serving as loan collateral; second, to mitigate debt risks for cloud providers like CoreWeave, NVIDIA introduced a modular solution to shorten deployment downtime and reduce financing costs; third, strong demand is the fundamental prerequisite for resolving current financing and supply issues, stating that \"with demand, all problems can be solved.\" Currently, the monetization price for GPU inference using DeepSeek models reaches as high as $38 per hour, far exceeding current market rental rates" datetime: "2026-10-22T04:08:44.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/302644185.md) - [en](https://longbridge.com/en/news/302644185.md) - [zh-HK](https://longbridge.com/zh-HK/news/302644185.md) generator: "portal-rs" --- # Jensen Huang Addresses Key Concerns: CoreWeave Debt, GPU Depreciation Period, and DeepSeek's High Gross Margin on Inference On Wednesday, following the conclusion of the GTC Berlin conference, NVIDIA CEO Jensen Huang and Silicon Valley tech investor Gavin Baker accepted an interview with Yahoo Finance. During the interview, Huang directly responded to three core issues currently troubling AI investors: the debt financing risks of new cloud computing companies, the actual service life of GPU assets, and the surprising gross margin data on inference disclosed by DeepSeek. Addressing debt financing risks and the service life of GPUs, Huang clearly stated that the service life of NVIDIA GPUs can reach seven to eight years, far exceeding the four years previously widely recognized by the market. This characteristic makes them eligible to serve as collateral for loans. He also pointed out that **NVIDIA has provided guarantees for the resale residual value of chips in certain financing projects to help new cloud computing companies like CoreWeave reduce their financing costs.** Furthermore, Gavin Baker cited the latest data disclosed by DeepSeek, noting that the monetization price for GPU inference using this open-source model reaches as high as $38 per hour, far exceeding current market rental prices. This may well be the reason why Blackwell chip rental prices have risen for 17 consecutive trading days. Huang responded to this by stating that **strong demand is the fundamental prerequisite for resolving all current financing and supply issues, asserting that "with demand, all problems can be solved."** These statements have direct implications for the market. New cloud computing companies (NeoClouds) like CoreWeave have recently sparked investor concern due to their highly leveraged financing structures, with their stock and bond prices under continuous pressure. Huang's response attempts to reconstruct the market's perception of debt risks in AI infrastructure from the perspective of asset quality, while DeepSeek's inference gross margin data provides new evidence for the sustainability of GPU demand. ## Reassessment of GPU Depreciation Period: From Four Years to Seven or Eight Years In the interview, Huang emphasized the durability of NVIDIA GPU assets, characterizing it as one of the core solutions to the current AI financing dilemma. He stated that the market had previously generally used four years as the depreciation cycle for GPUs, but this assumption is now outdated. "**People are discovering that what was once considered a four-year lifecycle could actually be eight years or even longer.**" He attributed this characteristic to NVIDIA's continuous upgrades to the CUDA software stack, which keeps hardware productive throughout its lifecycle. He also confirmed that the expected service life of the next-generation chip, Vera Rubin, is similarly seven to eight years. Huang further elaborated on the implications for capital markets: the high residual value of GPUs, their broad applicability (supporting all AI models and cloud service providers), and their software programmability collectively form a collateral base that lending institutions can underwrite. "Today, most GPUs have almost no residual value, **but people now recognize that NVIDIA's computing assets have quite high residual value, so they can be used to backstop the debt itself.**" He stated that, under this logic, NVIDIA's ability to bear the financing risk of AI infrastructure is actually superior to any other asset class. ## DSX Pod: Compressing Downtime, Accelerating Capital Returns Addressing the core financing pain point for new cloud computing companies like CoreWeave—where there is a gap of several months between purchasing equipment and generating revenue—Huang introduced NVIDIA's newly launched DSX Pod solution, a modular deployment format that prefabricates AI factories within containers. He explained that the construction cost of a gigawatt-scale data center is currently around $50 billion, and it typically takes four to six months from the time equipment leaves the factory to complete deployment and become operational. During this period, capital remains idle, but interest costs begin to accrue. **By moving some manufacturing processes upstream in the supply chain, the DSX Pod aims to compress this cycle to three to four months, or even shorter.** Gavin Baker agreed with this assessment, referencing the $1 trillion scale of AI capital expenditure. He pointed out that if the deployment cycle is six to nine months, it means hundreds of billions of dollars in capital remain non-productive for extended periods, creating systemic pressure on financing structures. Huang stated, "If we can shorten the time from system shipment to deployment and operation from months to weeks or even a month, it will be a total game-changer for all AI factories." ## DeepSeek Inference Gross Margin Data Draws Attention During the interview, Gavin Baker cited the latest operational data disclosed by DeepSeek, providing a new quantitative perspective on the strength of current GPU demand. According to him, DeepSeek publicly stated that the monetization price for GPU inference using its open-source model reaches $38 per hour, while the mainstream rental price for GPUs in the current market is far below this level. He believes this data may explain why Blackwell chip rental prices have risen for 17 consecutive trading days. Huang responded by stating that strong demand on the demand side is the fundamental prerequisite for resolving all current financing and supply issues. He listed the robust demand from organizations such as OpenAI, Anthropic, Meta, xAI, Google Gemini, DeepSeek, and Moonshot AI, stating: **"Without demand, no problem can be solved; with demand, all problems can be solved."** He also characterized the current timeframe as a key turning point where AI officially shifts from "invention mode" to "practical mode," believing that the continued rise in token demand has solid fundamental support. ## China's AI Development and the AI Employment Narrative Additionally, Huang stated that China is the only country in the world capable of achieving large-scale output simultaneously in three dimensions: chip manufacturing, open-source model development, and the training of AI researchers. "No country can train AI researchers on the same large scale as China." He also noted that the United States' advantage lies in possessing the world's top-tier frontier laboratories, chips, and computing infrastructure, and stated that NVIDIA will continue to compete actively globally, including in emerging markets such as Africa, Central Asia, Brazil, and Egypt. On the issue of AI and employment, Huang directly characterized the current mainstream narrative as "completely wrong." Citing employment growth data from various industries including energy, chips, infrastructure, construction, and healthcare, he argued that AI is actually creating more jobs at various levels rather than eliminating them. "The narrative about employment is almost entirely reversed; it is completely wrong." He also pointed out that builders of AI data centers should communicate in advance and proactively integrate when entering communities to improve public perception, noting that the industry as a whole has made improvements in this direction. ### Related Stocks - [CRWV.US](https://longbridge.com/en/quote/CRWV.US.md) - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) - [ANV.US](https://longbridge.com/en/quote/ANV.US.md) - [CORD.US](https://longbridge.com/en/quote/CORD.US.md) - [CRWG.US](https://longbridge.com/en/quote/CRWG.US.md) - [CRWU.US](https://longbridge.com/en/quote/CRWU.US.md) - [CRWX.US](https://longbridge.com/en/quote/CRWX.US.md) - [CWVX.US](https://longbridge.com/en/quote/CWVX.US.md) - [CWY.US](https://longbridge.com/en/quote/CWY.US.md) - [DIPS.US](https://longbridge.com/en/quote/DIPS.US.md) - [NVC.US](https://longbridge.com/en/quote/NVC.US.md) - [NVD.US](https://longbridge.com/en/quote/NVD.US.md) - [NVDB.US](https://longbridge.com/en/quote/NVDB.US.md) - [NVDD.US](https://longbridge.com/en/quote/NVDD.US.md) - [NVDG.US](https://longbridge.com/en/quote/NVDG.US.md) - [NVDL.US](https://longbridge.com/en/quote/NVDL.US.md) - [NVDO.US](https://longbridge.com/en/quote/NVDO.US.md) - [NVDQ.US](https://longbridge.com/en/quote/NVDQ.US.md) - [NVDS.US](https://longbridge.com/en/quote/NVDS.US.md) - [NVDU.US](https://longbridge.com/en/quote/NVDU.US.md) - [NVDW.US](https://longbridge.com/en/quote/NVDW.US.md) - [NVDX.US](https://longbridge.com/en/quote/NVDX.US.md) - [NVDY.US](https://longbridge.com/en/quote/NVDY.US.md) - [NVIB.US](https://longbridge.com/en/quote/NVIB.US.md) - [NVII.US](https://longbridge.com/en/quote/NVII.US.md) - [NVIT.US](https://longbridge.com/en/quote/NVIT.US.md) - [NVW.US](https://longbridge.com/en/quote/NVW.US.md) - [NVYY.US](https://longbridge.com/en/quote/NVYY.US.md) - [NYYY.US](https://longbridge.com/en/quote/NYYY.US.md) - [07388.HK](https://longbridge.com/en/quote/07388.HK.md) - [07788.HK](https://longbridge.com/en/quote/07788.HK.md) - [09388.HK](https://longbridge.com/en/quote/09388.HK.md) - [DPSK.NA](https://longbridge.com/en/quote/DPSK.NA.md) ## Related News & Research - [Nvidia backs Indian startup helping companies run AI models with fewer GPUs](https://longbridge.com/en/news/303217827.md) - [HPE puts storage to work on AI’s expensive memory problem](https://longbridge.com/en/news/303529539.md) - [NVIDIA Corporation $NVDA Shares Purchased by LifeGoal Investments LLC](https://longbridge.com/en/news/303853182.md) - [Blockmate Agrees to Buy NVIDIA B200 Compute Node for About USD 635,000](https://longbridge.com/en/news/303220950.md) - [CodeBoxx Unveils CrewKit Forge, an AI Software Factory in One Cubic Meter That Companies Own and Run On Premises](https://longbridge.com/en/news/303983793.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**