AI compute is the 21st century's most critical strategic resource, analogous to oil in the 20th century, and is central to global competition.
The AI market is undergoing a crucial structural shift from training-heavy to inference-heavy workloads, which will diversify the chip market and inevitably challenge NVIDIA's current dominance.
The financial health of the early-stage venture capital market is directly dependent on a successful wave of IPOs from major late-stage tech companies like SpaceX and OpenAI to ease liquidity pressure.
The adoption of Chinese open-source AI models by U.S. enterprises is a significant and accelerating trend, highlighting a complex interplay between global innovation and national interests.
True progress in AI is now measured by its commercialization and application, such as Perplexity's complex agents and the integration of reasoning into robotics, rather than by model training benchmarks alone.
Early Startup Era
Jang references the early days of companies like 'UberCab', highlighting their struggle to raise capital as a baseline for understanding venture market dynamics.
Recent Past (c. 2.5 years)
Cites the rapid growth of generative media startup Fowl, which reached an $8 billion valuation just two and a half years after its founding, as an example of recent market velocity.
Current Year
Identifies the current year as a pivotal moment where the balance of AI compute usage is flipping, with inference workloads growing from one-third to two-thirds of all compute.
Present Day
Describes the current market landscape where NVIDIA holds over 90% of the AI chip market, Perplexity has launched the first mainstream AI agents, and the VC market faces significant liquidity pressure.
Near Future
Predicts that NVIDIA's market dominance will face increasing competition as the shift to inference workloads accelerates.
Future Outlook
Posits that the financial health of the early-stage venture ecosystem depends on a future wave of successful IPOs from major tech companies like SpaceX, OpenAI, and Databricks.
▶The AI Compute EconomyMay 2026
Jang frames AI compute as the 21st century's most critical strategic resource, underpinning global competition. He details the current market structure, with NVIDIA's 90%+ dominance in accelerator chips, and the massive resource requirements of advanced models from labs like OpenAI and Anthropic.
Investors should monitor the training-to-inference compute ratio as a key indicator of market maturation and a potential trigger for a shift in market leadership away from NVIDIA towards more diversified hardware players.
▶Venture Capital and Market LiquidityMay 2026
Jang expresses concern over the liquidity pressure on early-stage venture capital funds. He argues that the health of the entire startup ecosystem is contingent on successful, large-scale IPOs from late-stage private companies like SpaceX, OpenAI, and Databricks to unlock capital.
The valuation and public market performance of a few key 'bellwether' tech companies are disproportionately important, acting as a gatekeeper for capital flow into the next generation of startups.
▶The Commercialization and Application of AIMay 2026
Jang sees the market's pivot to inference-heavy workloads as proof that AI has been successfully commercialized for widespread enterprise use. He points to specific examples like Perplexity's launch of complex AI agents to a mainstream audience and the integration of intelligent reasoning models into robotics as key milestones.
Analysts should shift focus from foundational model capabilities to the practical deployment of end-to-end applications and agents, as this is where Jang indicates value is now being created and measured.
▶Geopolitics of AI DevelopmentMay 2026
Jang highlights the global dimension of AI competition, particularly the rise of Chinese open-source models. He notes their significant adoption within U.S. enterprises, contingent on being hosted domestically, which points to a complex interplay between global innovation and national security concerns.
The proliferation of high-quality open-source models from state-backed competitors like China presents both an opportunity for rapid innovation and a significant supply chain and security risk for Western companies.