The headlines scream crisis, but the underlying fundamentals tell a different story. While pundits obsess over OpenAI CEO Sam Altman's character and speculate about NVIDIA's competition from Google's TPUs (tensor processing units), the real narrative is being buried: artificial intelligence compute is undergoing genuine liftoff as demand for agentic AI (systems that can autonomously complete multi-step tasks) explodes higher .

Why Media Drama Shouldn't Drive Your AI Investment Thesis?

In April 2026, The New Yorker published a 16,000-word feature examining Altman's character and integrity, prompting widespread criticism of OpenAI's acquisition of TBPN and speculation about leadership turmoil. Meanwhile, skeptics are questioning whether NVIDIA can maintain dominance as Broadcom and Google develop competing chip technologies, and whether the company's next-generation Rubin chips will see the volume growth expected for 2026 .

But here's what matters: personality-driven narratives have a poor track record predicting company success. In early 2013, legendary investor Julian Robertson abandoned his Apple position after reading Walter Isaacson's biography of Steve Jobs, concluding that a "mean person" couldn't build a great company long-term. Apple subsequently entered a generational run-up. The lesson applies directly to OpenAI today. Whether Altman is ruthless in business is largely irrelevant to whether OpenAI's model and product teams can execute .

The TBPN acquisition itself is not a strategic pivot or warning sign. OpenAI is simply providing resources to an organization that aligns with its pro-technology worldview. This is thin analysis masquerading as insight .

What Do the Real Numbers Tell Us About AI Demand?

If you look past the noise, the data reveals explosive growth across the entire AI industry. Anthropic, OpenAI's primary competitor, disclosed that its run-rate revenue surpassed $30 billion in early 2026, up from $9 billion at the end of 2025, as demand for Claude continues to accelerate . This represents a more than doubling of revenue in less than two months.

Token demand, a key metric for measuring actual AI usage, is growing even faster. OpenRouter, a platform that aggregates AI token consumption across multiple providers, shows aggregate AI token demand is up 15 times year-over-year . This isn't speculation; it's a direct measure of how much computational work the industry is performing.

Google executives are signaling that this growth will continue. Google CEO Sundar Pichai stated that AI models one year from now will be dramatically better, noting that memory supply is currently acting as a constraint on performance. According to industry sources, Google plans to release its next two major models on a rapid six-month cadence, suggesting the company expects sustained demand .

How to Evaluate AI Company Strength Beyond the Headlines

• Revenue Growth Trajectory: Look at actual run-rate revenue and year-over-year growth rather than leadership drama. Anthropic's jump from $9 billion to $30 billion in two months demonstrates real market demand, not speculation.
• Token Consumption Metrics: Monitor aggregate token demand across platforms like OpenRouter as a proxy for actual AI usage. A 15x year-over-year increase indicates the industry is in genuine growth phase, not a hype cycle.
• Compute Capacity and Roadmaps: Assess whether companies have sufficient hardware infrastructure to meet demand. OpenAI's advantage lies in having dramatically more compute capacity than rivals if demand materializes over the next few years.

The competitive landscape is not zero-sum. The fact that Anthropic and Google are thriving simultaneously suggests OpenAI is doing well too. Anthropic has benefited from focusing on enterprise coding and using large amounts of coding data in its training runs. OpenAI is pivoting more resources to enterprise and following suit, indicating the market is large enough for multiple winners .

NVIDIA's position remains strong despite competition concerns. The company's dominance stems from CUDA (Compute Unified Device Architecture), its proprietary software framework that makes its graphics processing units (GPUs) the standard for AI training and inference. While Broadcom and Google are developing alternatives, switching costs are high and NVIDIA's ecosystem advantage is substantial. The real question isn't whether competition exists, but whether the total market for AI chips is growing faster than any individual competitor's share is shrinking .

The evidence suggests it is. With token demand up 15 times year-over-year and major AI companies doubling revenue in months, the absolute size of the AI compute market is expanding rapidly. This rising tide lifts multiple boats, including NVIDIA's.

When macro volatility and political uncertainty dominate headlines, the temptation to panic is strong. But the underlying fundamentals of the AI industry are accelerating, not decelerating. The companies best positioned to capitalize on this moment are those with the compute capacity to serve explosive demand, and the product teams to deploy it effectively. For NVIDIA and OpenAI, that means the next few years could be transformative, regardless of what the media narrative suggests today.