Huang Ren-hsuan Confirms Nvidia's "Excessive Cash Flow": Will Return 50% to Shareholders in 2026 for Stock Buybacks and Dividends

NVIDIA FY2026 full-year free cash flow is approximately $96.5 billion. Jensen Huang announced that 50% of this will be returned to shareholders through buybacks and dividends. However, based on actual figures, about $41.1 billion has already been returned in FY2026, accounting for only 43%. Even more noteworthy is that as the company simultaneously announced a trillion-dollar revenue target for 2027, the real challenge lies in the tension between “returning value to shareholders” and “expanding capital expenditure.”
(Background: NVIDIA has become a legend in AI! Jensen Huang emphasizes AI, building a trillion-dollar GPU empire.)
(Additional context: NVIDIA’s market cap dominates globally! Huang’s net worth increases by $4 billion daily—how much closer is he to surpassing Elon Musk as the world’s richest person?)

Table of Contents

Toggle

• The acceleration behind the numbers
• The gap between buyback quota and actual execution
• Cold, hard conclusions

At GTC 2026, Jensen Huang said NVIDIA will return 50% of its free cash flow to shareholders this year, adding, “This will be a very large number.” The figure is indeed huge—but a closer look reveals that half of this statement is about what has already happened, and the other half is a check that hasn’t yet been cashed.

In FY2026, NVIDIA’s total free cash flow is about $96.58 billion. The actual buybacks and dividends for the year totaled approximately $41.1 billion, which is 43%, not 50%. The difference isn’t large, but the “50%” figure refers to the “calendar year 2026,” and the mismatch between fiscal and calendar years introduces some ambiguity in the comparison.

The acceleration behind the numbers

In Q4 FY2026, revenue hit $68.1 billion, up 73% year-over-year; for the full year, revenue reached $215.9 billion, up 65%. Maintaining over 60% growth at this scale—even just sustaining it—is already extremely rare. The guidance for Q1 FY2027 is $78 billion (±2%), implying that the quarter-over-quarter growth rate is still accelerating rather than slowing down.

The current market consensus narrative is: NVIDIA’s moat is deep enough, demand is broad enough, and competitors are slow enough. This story may be correct. But the data tells a different story—the real stress test has yet to come.

At GTC 2026, Huang showcased the Blackwell Ultra and next-generation Vera Rubin architectures, setting an ambitious goal: $1 trillion in cumulative orders by 2027. Prototypes like the Kyber framework (144 GPU vertical design) and Groq’s 3-language processor were also unveiled, sending a clear signal—that this company has no intention of letting capital sit idle.

The gap between buyback quota and actual execution

In August 2025, NVIDIA announced a new $60 billion buyback authorization. By the end of Q4 FY2026, $5.85 billion remained available. In other words, after a full fiscal year, only $150 million of the authorization had been used.

This isn’t unusual—large tech companies often set buyback limits rather than specific plans, and extended execution timelines are common. But when a company claims to heavily reward shareholders while also showcasing next-generation hardware roadmaps at GTC, the ultimate destination of capital depends on actions, not declarations.

Cold, hard conclusions

NVIDIA’s fundamentals are solid. The issue is: when a company simultaneously announces a “trillion-dollar revenue goal” and “massive shareholder returns,” these two priorities are inherently at odds in capital allocation. Huang’s approach is to bet on such rapid growth that both can be achieved.

So far, this bet appears correct. But the market is always right—until it isn’t.

If in 2027, signs of convergence in AI capital expenditure emerge, the “50% FCF return” promise will either serve as a cushion for NVIDIA’s stock or be another figure redefined. Continuous monitoring is essential.