Nvidia (NVDA) CEO Jensen Huang Doubles Down on 70% Revenue Growth Target Through 2027

Key Takeaways

• CEO Jensen Huang expressed strong confidence in achieving 70% year-over-year revenue expansion by 2027, noting that unconstrained market demand exceeds 100%
• Cybersecurity has been pinpointed as AI’s next significant expansion opportunity, with active collaborations involving CrowdStrike, Cisco, and Palantir
• System pricing has escalated significantly: from $18,000 for Hopper GPU systems to an expected $40,000 for the forthcoming Vera Rubin platform
• Nvidia’s client roster encompasses OpenAI, Google, Meta, Anthropic, xAI, and virtually all leading AI model creators
• Physical AI applications, encompassing autonomous transportation and robotics, represent another crucial expansion avenue highlighted by Huang

During his Thursday presentation at the Goldman Sachs Communacopia and Technology Conference, Nvidia CEO Jensen Huang presented a bullish outlook for his company’s trajectory while singling out cybersecurity as artificial intelligence’s next major expansion territory.

Shares of NVDA were changing hands near $218 on Thursday, giving the chipmaker a market capitalization of $5.27 trillion alongside a price-to-earnings multiple of 27.6.

NVIDIA Corporation, NVDA

The CEO reiterated his projection calling for 70% year-over-year revenue expansion in 2027. According to Huang, actual unconstrained market demand is running north of 100%, though supply chain capacity continues to act as the primary constraint.

Trailing twelve-month revenue reached $303 billion, accompanied by gross profit margins approaching 75%.

Huang characterized Nvidia’s evolution as extending far beyond GPU sales into a comprehensive AI factory platform. He dubbed the company “the world’s first and only growth value stock.”

System Prices Continue Upward Trajectory

The pricing structure for systems has experienced consistent increases with each new generation. Hopper GPU-based systems commanded approximately $18,000, while Blackwell systems reach roughly $25,000, and the forthcoming Vera Rubin platform is expected to arrive at approximately $40,000.

Certain integrated system configurations now command prices reaching $8.5 million. Grace, Blackwell, and NVLink system deployments expanded 27% on a month-over-month basis in 72-rack setups.

According to Huang, Nvidia systems offer durability, are suitable for rental arrangements, and are now being utilized as loan collateral, representing what he characterized as an innovative form of asset-backed computing capital.

His projections suggest AI infrastructure expenditures could climb to a range of $3 trillion to $4 trillion by decade’s end, propelled by generative computing demands and the diminishing performance improvements from Moore’s Law.

Cybersecurity Positioned as Primary AI Expansion Vector

Huang identified cybersecurity as the most immediate emerging growth sector for AI applications, positioning it ahead of physical AI implementations such as robotics.

The company has established cybersecurity collaborations with CrowdStrike, Cisco, and Palantir, incorporating its Nemotron models into these strategic partnerships.

Huang emphasized that the perpetual red-team versus blue-team dynamics inherent to cybersecurity generate substantial and continuous demand for AI computational resources.

Regarding physical AI applications, Huang projected meaningful advancement in autonomous vehicle technology within the next two to three years. He suggested manipulation systems designed for mid-market manufacturing operations could materialize in approximately two years.

He also drew attention to a recently unveiled development called Alpamayo, which he indicated can substantially decrease training data requirements for autonomous vehicles by emphasizing reasoning capabilities.

Nvidia’s customer portfolio currently includes OpenAI, Google Gemini, Meta, Anthropic, xAI’s Grok, major cloud service providers, enterprise clients such as Eli Lilly, Merck, and Jane Street, plus neocloud partners including CoreWeave and Lambda.

Huang asserted that Nvidia stands as the sole provider powering every significant AI model, encompassing both open-source and proprietary variants.

Addressing supply limitations, Huang noted that land availability, electrical power access, and data center real estate represent more significant bottlenecks than component procurement. The company revealed plans for a 2 gigawatt data center commitment in Australia scheduled for 2027, signifying an $80 billion infrastructure investment.