Nvidia Projects Longer Lifespan for Blackwell GPUs

CEO Jensen Huang expects new GPU architectures to remain functional for up to eight years.

Updated on Oct. 21, 2026 in Semiconductors

Isometric editorial illustration of stacked geometric server rack hardware in deep teal and slate blue, representing industrial equipment longevity. — Nvidia projects its Blackwell GPU architecture will maintain a seven-to-eight-year useful lifespan, allowing data center operators to lower annual investment expenses. AI Illustration. Upload story photo >

Live Poll

Does an extended lifespan for high-end technology make large business investments a safer bet today?

Nvidia CEO Jensen Huang stated that the company's Blackwell and Vera Rubin GPUs possess a useful lifespan of seven to eight years. This projection significantly exceeds the four-year lifespan previously assumed in financial models for data center investments.

Why it matters

By extending the projected useful life of hardware, operators can spread infrastructure costs over more years to improve investment health. Longer-lived chips also serve as better collateral, helping firms secure more favorable credit terms for large-scale data center projects.

Nvidia estimates its Blackwell and Vera Rubin architectures will have a useful life of seven to eight years. This outpaces current industry depreciation schedules, which typically range from five to six years.

The players

Jensen Huang

He is the chief executive officer of Nvidia and has been a central figure in the expansion of AI hardware.

Nvidia

The company is a leading technology firm specializing in the development of graphics processing units and artificial intelligence hardware.

CoreWeave

This is a specialized cloud provider that builds infrastructure optimized for large-scale GPU computing workloads.

The details

Spreading hardware costs over a longer timeframe allows companies to lower annual expenses and adjust capital allocation strategies. The company has already seen evidence of this trend as CoreWeave extended its usage contracts for the older A100 GPUs through 2029.

2020: Nvidia launches the A100 GPU.
2025: Blackwell GPUs enter volume production.
H2 2026: The Vera Rubin architecture begins its production ramp.
October 21, 2026: CEO Jensen Huang provides the new lifespan projections.
2029: CoreWeave A100 contracts expire.

The Tech Race

By moving beyond the standard four-year depreciation model, Nvidia is attempting to redefine the economic sustainability of AI infrastructure. This shift positions the company to maintain dominance as rival hardware providers continue to adhere to shorter, more traditional lifecycle expectations.

While this change primarily affects large-scale cloud operators, it signals that the hardware running your favorite AI tools may become more cost-effective to maintain over time. This efficiency could eventually stabilize the pricing of AI-driven cloud services as infrastructure providers amortize their investments over longer periods.

The takeaway

Nvidia's aggressive revenue outlook is predicated on the idea that its newer architectures will stay relevant in the data center for nearly a decade. Investors should monitor whether these longer lifespans actually materialize as the industry continues its rapid pace of innovation.