Nvidia, one of the faces of the AI boom, has recently announced the pricing details for its highly anticipated next-generation graphics processor designed specifically for artificial intelligence (AI) tasks.
Speaking with CNBC, Nvidia CEO Jensen Huang disclosed that the new processor, dubbed Blackwell, will be priced between $30,000 and $40,000 per unit.
Nvidia CEO Talks About Latest Blackwell Chip Pricetag
Speaking with CNBC's Jim Cramer, Huang emphasized the groundbreaking nature of the Blackwell chip, revealing that Nvidia invested approximately $10 billion in research and development to bring this innovative technology to fruition. "We [even] had to invent some new technology to make it possible."
The pricing strategy aligns Blackwell with its predecessor, the H100, indicating a continuity in Nvidia's pricing model for its flagship AI chips. Analysts had estimated the cost of the H100 to be in a similar range, ranging from $25,000 to $40,000 per chip, making Blackwell's pricing a natural progression within the industry.
AI-Backed Chip Demand
Nvidia consistently releases new generations of AI chips approximately every two years, with each iteration promising improved performance and energy efficiency.
Blackwell, boasting a larger physical size and the integration of two chips, is poised to deliver enhanced capabilities compared to its predecessors. This advancement comes at a time when the demand for AI processing power is skyrocketing, fueled by the proliferation of AI applications across various industries.
The significance of Nvidia's AI chips in driving exponential growth in the company's sales cannot be overstated. Since the onset of the AI boom in late 2022, Nvidia has witnessed a tripling of quarterly sales, with major players like Meta investing heavily in Nvidia's H100 GPUs for training their AI models.
Recently, Jensen Huang has made strong remarks regarding the superiority of Nvidia's GPUs over competitors, suggesting that even if rival chips were free, they would not match Nvidia's performance.
The Blackwell series is expected to bolster further Nvidia's position as a leading provider of AI hardware solutions, catering to the evolving needs of AI-driven enterprises worldwide.
What Makes 'Blackwell' Standout
Moreover, the unveiling of Blackwell underscores Nvidia's commitment to innovation in AI hardware, as evidenced by the chip's remarkable performance metrics.
According to Nvidia, Blackwell can deliver a staggering 30 times improvement in performance for generative AI services, such as OpenAI's GPT-4, while consuming 25 times less energy compared to the previous generation of Hopper GPUs. This leap in efficiency enhances the computational capabilities of AI systems and addresses concerns regarding energy consumption and sustainability.
In addition to its technical prowess, Blackwell's pricing strategy and availability raise questions about accessibility and environmental impact. With an anticipated price tag reaching eye-watering levels, Blackwell may only be accessible to a select few organizations, potentially exacerbating existing disparities in AI adoption.
However, the escalating electricity demand from data centers supporting AI training raises concerns about carbon emissions and environmental sustainability, highlighting the need for responsible innovation in the AI industry.
According to the International Energy Agency, current data center energy usage is around 460 terawatt hours in 2022. It could climb up to 1,050 TWh in 2026-equivalent to the energy demands of a country like Germany.
Join the Conversation