Loading page…
Nvidia’s quarterly results show its AI infrastructure business still expanding rapidly. Jalapeño offers an early, narrow efficiency comparison—and signals that more AI companies want custom hardware options.
Listen to this story
OpenAI’s Jalapeño project adds a diversification angle to an Nvidia-led infrastructure market, but its disclosed performance gap is narrow in scope. Nvidia management expects 70% fiscal 2028 revenue growth—well above the 44% analyst consensus—after a strong June quarter, while OpenAI reported 1.5–1.9× more work per watt and lower latency than selected Nvidia systems.
Nvidia posted $96.22 billion in June-quarter revenue and $53.95 billion in net income, with both more than doubling year over year.
Colette Kress’s 70% fiscal 2028 growth forecast is management’s expectation, not revenue already delivered.
OpenAI’s efficiency result is company-supplied and limited to selected Nvidia systems and three publicly available models.
Nvidia expects fiscal 2028 revenue growth of 70% after its revenue more than doubled in the June quarter. Yet OpenAI is developing Jalapeño, a custom AI chip that it says outperformed selected Nvidia systems on efficiency in its tests, setting current sales momentum against a push for alternatives in parts of AI hardware.
Nvidia reported $96.22 billion in second-quarter revenue for the three months ended in June, more than double the prior-year figure. Net income rose by more than 100% year over year to $53.95 billion, or $2.22 a share. Its results beat expectations, and its revenue forecast also topped estimates.
Chief Financial Officer Colette Kress put Nvidia’s fiscal 2028 growth expectation well above the analyst consensus cited by CNBC. The forecast is management’s expectation, rather than a result already delivered.
Nvidia CFO Colette Kress said the company expects 70% fiscal 2028 revenue growth, compared with a 44% analyst consensus expectation.
OpenAI has begun developing Jalapeño, its own AI chip. The company said it delivered 1.5 to 1.9 times more AI work per watt than selected Nvidia systems across three publicly available models, and that it recorded lower latency in those tests. Those are company-supplied results from a defined set of comparisons.
The effort is not unique to OpenAI. CNBC identified Google, AWS and Meta as other companies developing their own AI chips, and characterized OpenAI as a potential Nvidia competitor in some areas. That competitive pressure is directed at parts of the infrastructure market while Nvidia continues to project rapid revenue growth.
Jalapeño’s disclosed comparison is limited to selected Nvidia systems and three models, so it does not settle how custom chips will fare across broader workloads. The immediate picture is clearer: Nvidia’s reported demand remains strong, while major AI companies are building more hardware options of their own.
Loading discussion...
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.