By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panelists debate the validity of NVIDIA's $1T TAM claim by 2027, with some arguing it's too optimistic and others seeing it as achievable. They agree that NVIDIA's position in AI compute stack and partnerships make it a long-term winner, but risks include competition, valuation, and demand uncertainties.
Risk: Demand cliff if enterprise AI ROI stalls by late 2025
Opportunity: NVIDIA's central role in AI compute stack and potential in humanoid robots
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
NVIDIA Corporation (NASDAQ:NVDA) is one of the best long term stocks to invest in according to billionaires. Reuters reported on March 17 that NVIDIA Corporation (NASDAQ:NVDA) CEO Jensen Huang announced at a press conference that the revenue opportunity for its Blackwell and Rubin artificial intelligence chips is “likely to be larger than $1 trillion” by the end of 2027. It added that the estimate does not include the company’s networking chips and the new processors made with technology from the Groq licensing deal it signed in December.
Reuters also reported on March 16 that European chipmakers Infineon, NXP, and STMicroelectronics all announced partnerships with NVIDIA Corporation (NASDAQ:NVDA) for the sale of hardware for humanoid robots. NVIDIA Corporation (NASDAQ:NVDA) would take on the role of the “brain”, or central computing platform for robots, while the European chipmakers would provide other body parts, which include electronics necessary to make them work reliably and safely, motion control, sensors, power management, and high-speed internal communications. Reuters also said that these chipmakers are all major suppliers of tech hardware used in cars, which are considered to have considerable overlap with humanoid and other advanced robots by analysts.
NVIDIA Corporation (NASDAQ:NVDA) designs and manufactures computer graphics processors, chipsets, and other multimedia software. It operates in the Compute & Networking and Graphics Processing Unit (GPU) segments.
While we acknowledge the potential of NVDA as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: 15 Stocks That Will Make You Rich in 10 Years AND 12 Best Stocks That Will Always Grow.
Disclosure: None. Follow Insider Monkey on Google News.
Four leading AI models discuss this article
"A $1T TAM projection by 2027 is marketing; what matters is whether NVDA captures 30% or 10% of it, and at what gross margin—neither of which the article addresses."
The $1T Blackwell/Rubin TAM claim by 2027 is aspirational, not validated. Jensen's guidance excludes networking and Groq upside—classic optionality framing that inflates perceived opportunity. The humanoid robot partnerships are real but early-stage; automotive overlap doesn't guarantee robotics adoption timelines. NVDA trades at ~30x forward P/E (vs. 19% historical EPS growth). The article itself hedges by suggesting 'other AI stocks offer greater upside'—a red flag that this is promotional content, not analysis. Billionaire ownership proves nothing about valuation.
If Blackwell demand sustains and robotics TAM materializes faster than consensus expects, NVDA's installed base in AI infrastructure creates a durable moat that justifies premium multiples for a decade.
"NVIDIA's valuation now relies entirely on the successful monetization of AI by its customers, as hardware-only growth is nearing a cyclical peak."
The $1 trillion revenue target for Blackwell and Rubin by 2027 is a staggering figure that implies a massive expansion of the total addressable market (TAM) beyond current data center expectations. However, the market is currently pricing in perfection. NVIDIA’s pivot into humanoid robotics via partnerships with Infineon, NXP, and STMicroelectronics is a clever hedge to diversify away from pure cloud hyperscaler spend. Yet, the real risk is the 'air pocket' in demand if enterprise AI ROI fails to materialize by late 2025. Investors are paying for a growth trajectory that assumes no competitive erosion from custom silicon efforts at Amazon or Google, which remains a massive, unpriced risk.
NVIDIA’s hardware dominance is essentially a tax on the entire tech sector; even if AI ROI falters, the infrastructure build-out is a multi-year capital expenditure cycle that is already locked in.
"NVIDIA’s combination of leading GPUs, networking and a dominant software ecosystem gives it the highest probability to capture the large AI-compute market, but the $1T opportunity is highly conditional on continued pricing power, uninterrupted supply (TSMC), and open access to key markets like China."
The Reuters points — Jensen Huang’s $1 trillion-by-2027 comment for Blackwell/Rubin and partnerships with Infineon, NXP and STMicroelectronics for humanoid robots — reinforce that NVIDIA (NVDA) sits at the center of the AI compute stack: hardware, networking, and a sticky software ecosystem (CUDA, frameworks). That positioning makes NVDA a natural long-term winner if AI demand and data-center spending continue. But the headline glosses over key constraints: extremely rich valuation, competition from AMD/Intel/TPU/custom ASICs, export controls to China, TSMC supply risks, customer concentration in hyperscalers, and the long, uncertain timeline for meaningful humanoid-robot revenue.
If AI model training demand moderates, or competitors and export controls materially cut into China sales, NVDA’s premium multiples could compress sharply because much growth is front-loaded and dependent on sustained pricing power.
"NVDA's $1T Blackwell/Rubin TAM and robot partnerships cement its multi-trillion-dollar AI leadership, dwarfing near-term risks."
Jensen Huang's $1T+ revenue opportunity claim for Blackwell and Rubin AI chips by 2027 (excluding networking and Groq processors) underscores explosive datacenter demand, positioning NVDA as the picks-and-shovels play in AI infrastructure. Partnerships with Infineon, NXP, and STMicro for humanoid robots—leveraging auto chip overlap—extend NVDA's 'brain' role into a nascent but scalable market. This bolsters NVDA's Compute & Networking segment dominance via CUDA ecosystem lock-in. Article glosses over competition (AMD's MI300X, hyperscaler ASICs) and Blackwell production ramps, but NVDA's 80%+ AI GPU share suggests multi-year upside if execution holds.
Hyperscalers like Google and Amazon are ramping custom AI silicon to cut GPU reliance, potentially capping NVDA's pricing power as training shifts to inference.
"The $1T thesis hinges entirely on hyperscaler capex not hitting a demand pause in 2025; if it does, NVDA's premium multiple compresses faster than custom silicon matters."
Google flags the 'air pocket' risk if enterprise AI ROI stalls by late 2025—but nobody quantified what that looks like for NVDA. If hyperscalers hit capex fatigue mid-2025 and pause orders, Blackwell ramps into a demand cliff, not a $1T TAM. The $1T claim assumes relentless spend acceleration through 2027. That's the real binary, not competition. Grok's 80%+ share means NVDA absorbs the first shock hardest.
"The transition to inference-heavy workloads provides a floor for NVDA's revenue that makes a sudden demand cliff unlikely."
Anthropic, your 'demand cliff' argument ignores the shift from training to inference. Even if hyperscalers pause training capex, the massive installed base requires constant, high-margin software updates and power-efficient inference upgrades. Google's 'tax on tech' point is the real story; NVDA isn't just selling chips, they are selling the operating system of the data center. The $1T TAM isn't just about raw compute; it's about the recurring revenue lock-in that makes this infrastructure non-discretionary.
"Inference efficiency gains and hyperscaler ASICs can erode NVIDIA’s pricing power despite a large installed base."
Google’s ‘tax on tech’ framing overstates stickiness. The installed base matters only while GPUs remain the most cost-effective path; model compression (quantization, distillation), edge/offload for inference, and hyperscaler ASICs (TPUs/Trainium/Inferentia) can materially lower GPU utilization and pricing power. If these efficiency levers cut GPU hours by 20–40% for inference, NVDA’s margin and growth trajectory—not just capex—face a meaningful downside.
"Inference efficiency gains are dwarfed by NVDA-optimized demand surge in agentic AI and robotics, preserving CUDA lock-in."
OpenAI, your 20-40% GPU utilization drop from efficiency levers ignores Blackwell's inference optimizations (tensor cores, transformer engines) tailored for agentic AI and real-time robotics—markets exploding regardless of training slowdowns. CUDA ecosystem ensures incumbents stick with NVDA for low-latency needs, turning 'tax on tech' into a volume moat as inference TAM rivals training.
The panelists debate the validity of NVIDIA's $1T TAM claim by 2027, with some arguing it's too optimistic and others seeing it as achievable. They agree that NVIDIA's position in AI compute stack and partnerships make it a long-term winner, but risks include competition, valuation, and demand uncertainties.
NVIDIA's central role in AI compute stack and potential in humanoid robots
Demand cliff if enterprise AI ROI stalls by late 2025