The Aftershock Hit Asia First

Nvidia reported earnings, and Asian markets reacted swiftly.

SoftBank Group — the Japanese conglomerate run by Masayoshi Son — closed 19.85% higher on Thursday, according to CNBC. That single-day move added over $35 billion to its market cap. One earnings report from an American chipmaker added roughly the GDP of a small nation to a Japanese holding company's valuation overnight.

SoftBank's exposure is substantial. The firm has invested more than $30 billion into OpenAI, with investment gains in that position totaling $45 billion in the fiscal year ended March, per CNBC. It also holds a significant stake in Arm Holdings, whose chip designs power AI servers running Nvidia systems. Arm closed over 15% higher in U.S. trading Thursday.

The gains extended across the region. Taiwan's TSMC — Nvidia's primary manufacturer of advanced AI processors — rose over 2%. Japan's Renesas Electronics, a key Nvidia supplier, closed 8.2% higher. Tokyo Electron, a semiconductor equipment maker, also climbed. Bloomberg reported that Nvidia CEO Jensen Huang's AI and robotics vision sparked the broader Asia tech rally.

Nvidia's Numbers

Nvidia beat Wall Street estimates for the eighth consecutive quarter. Q1 revenue came in at $81.6 billion — above the analyst consensus of roughly $79 billion. Adjusted gross margins hit 75%. The dividend was increased.

Yet the stock sank in extended trading after the results dropped Wednesday night, per CNBC.

Wall Street has already priced in strong performance. The focus now is whether Nvidia can maintain dominance as the AI buildout moves into its next phase.

The Inference Question

The AI chip market is shifting structurally, and the question consuming analysts centers on inference — the phase where AI systems operate in production environments, answering questions and completing tasks.

The initial AI boom focused on training massive models. Nvidia's GPUs dominated that segment entirely. Inference differs significantly. It is less centralized, requires more distributed architecture, and relies more heavily on CPUs alongside GPUs.

That shift explains why Intel, AMD, and Micron have outperformed Nvidia over the past three months, according to CNBC.

On Nvidia's earnings call Wednesday, CJ Muse, managing director at Cantor Fitzgerald, directly asked how the upcoming Vera Rubin chip platform would affect Nvidia's inference market share. Jensen Huang's response, per CNBC: Nvidia's inference share was "growing very quickly" — a statement he repeated five or six times during the call.

Nicolas Gaudois at UBS wrote Wednesday that "server CPU demand is strongly inflecting up" driven by agentic AI — the semi-autonomous bot architecture that relies on traditional processors. Rival custom silicon — Alphabet's tensor processing units, Amazon's Trainium chips, AMD, and newcomer Cerebras — are competing for a share of the inference market Nvidia currently dominates.

The Vera Rubin Wildcard

Nvidia's response is Vera Rubin, its next-generation AI chip system scheduled to ship later this year. Huang told analysts that production capacity could be at maximum before the chip even ships.

"My sense is that we'll be supply constrained throughout the entire life of Vera Rubin," Huang said, per CNBC.

If supply constraints hold, demand would remain so far ahead of production that competition would have limited near-term impact. Customers would accept whatever Nvidia can produce.

But supply constraints create openings. When Nvidia chips are unavailable, companies begin evaluating alternatives — precisely the window AMD, Amazon, and Google are targeting.

Capital Spending Pressure

Projected 2026 capital spending by AI hyperscalers — Alphabet, Amazon, Meta, and Microsoft — has risen from $531 billion in December to $725 billion, according to BNP Paribas as cited by Ynetnews. In five months, expected spending increased by nearly $200 billion.

Nvidia currently holds roughly 85% of the AI chip market, per Ynetnews. The market is growing faster than most competitors can realistically capture share in the near term.

But the companies investing $725 billion have strong incentive to avoid permanent dependence on a single vendor.

What Comes Next

Most coverage frames this as a straightforward victory — Nvidia beat, markets rallied, AI demand remains robust.

The substantive story is a company at peak dominance facing its first genuine architectural challenge. The shift to inference is not peripheral. Huang spent considerable time on his earnings call addressing it. When a CEO emphasizes a single point repeatedly, that's the area worth monitoring.

If Vera Rubin ships on schedule and remains supply-constrained, Huang's assessment holds and competitors remain marginalized. If Vera Rubin encounters delays, or if the inference market fragments faster than expected, the companies spending $725 billion on AI infrastructure will have options. The boom in AI continues, but Nvidia's period of uncontested dominance may be ending.