1. 首页
  2. ›NVIDIA Announces $0.25 Per Share Quarterly Dividend; Supply Chain Expansion Locks In Growth

NVIDIA Announces $0.25 Per Share Quarterly Dividend; Supply Chain Expansion Locks In Growth

N
nashnova research
·今天发布·阅读约 13 分钟

Nvidia held its quarterly dividend at $0.25 per share while adding new supply-chain partners and lifting total purchase commitments to $390 billion — the growth bottleneck has shifted from demand to production capacity.

01

The dividend didn't change — what does that signal?

•
Nvidia declared a $0.25 per share quarterly cash dividend, unchanged for a fourth straight quarter. The forward yield sits at roughly 0.64%.
•
Record date and ex-dividend date are December 2; payout is December 23.
•
This means → Nvidia is channeling capital into supply-chain expansion rather than raising dividends — for a company whose profits are surging, a flat payout is itself a capital-allocation signal.
02

Who just joined the supply chain, and why?

•
Interim CEO and CFO Colette Kress disclosed on the FY2027 Q3 earnings call that Nvidia signed new substrate agreements with Taiwan's Kinsus Interconnect Technology and Korea's Samsung Electro-Mechanics, plus an advanced-packaging deal with Taiwan's SPIL.
•
Kinsus is the only entirely new partner in this round. Nvidia had already announced deals with Amkor, Ibiden, KYEC, and Unimicron in October; Samsung Electro-Mechanics and SPIL were on that earlier list as well.
•
In plain terms = building a GPU requires more than just the chip — it needs substrates (the circuit boards that carry the chip) and advanced packaging (the process that bonds multiple chips together). Both are capacity-constrained, so Nvidia is signing up suppliers one by one to close the gap.
03

What do $390 billion in purchase commitments mean?

•
Nvidia's total purchase commitments rose to $390 billion in Q3. Kress called this "a reflection of confidence in the sustainability of long-term demand."
•
The October substrate deals will lift capacity 2.5× over three years. Amkor plans to expand advanced-packaging capacity by over 300% by end-2028. KYEC is building new U.S. capacity backed by an $11 billion prepayment from Nvidia.
•
This means → Nvidia is using upfront cash to lock suppliers onto its capacity roadmap — whoever takes the prepayment gives Nvidia first claim on production lines for years to come.
04

How are the new chips and networking products progressing?

•
Vera Rubin began volume shipments in August. Hyperscale cloud customers brought it into production in roughly one month — half the ramp time required for the prior-generation Grace Blackwell.
•
On the MLPerf Endpoints benchmark, Vera Rubin delivers 10× the performance per watt of the GB300. Vera Rubin Ultra remains on track for next year; the Feynman architecture is slated for 2028.
•
Networking revenue topped $21 billion in the quarter. Scale-out switches using co-packaged optics — CPO, where optical modules are built directly into the switch-chip package — have entered mass production, offering 5× better power efficiency, 15× higher reliability, and 10% lower latency versus pluggable optics.
05

What is DSX POD, and why does it matter?

•
Nvidia introduced the DSX POD, a modular AI-factory reference architecture — essentially a pre-assembled, standardized rack system. Rack assembly and cluster-level integration shift upstream into the supply chain.
•
This means → customers receive near-turnkey cabinets instead of loose components. Nvidia estimates deployment timelines compress from months to weeks.
•
The first DSX PODs ship to U.S. and Central European customers early next year. Partners include Dell, Foxconn, and Schneider Electric. Nvidia says a 1 GW AI factory carries annual revenue potential of nearly $50 billion, and bringing first inference online one quarter early could add over $10 billion in incremental revenue for the operator.
06

Where is the real growth bottleneck now?

•
Kress said Nvidia is working with partners to accelerate "capacity expansion of critical components — wafers, substrates, and advanced packaging."
•
This reflects a fundamental shift: the binding constraint on Nvidia's growth has moved from demand to supply — the problem is no longer finding buyers, but manufacturing fast enough.
•
In plain terms = Kinsus joining the roster, the $390 billion commitment, and the DSX POD rollout all attack the same problem: pushing up the ceiling on how much Nvidia can ship. The production cadence of Vera Rubin Ultra and Feynman will be the key window for judging whether this supply buildout converts into actual deliveries.

市场有风险,内容仅供研究参考,不构成投资建议。

相关标的

4
  • NVIDIANVDA.US-0.4%

    Direct event target: announced quarterly dividend, procurement commitments raised to $390 billion, supply expansion locks in growth.

  • CBOE S&P 500 PutWrite TW^PWT.INDEX-1.1%

    The only newly added substrate supplier, directly secured a new agreement with NVIDIA; capacity expansion brings incremental orders.

  • TT.CRYPTO+2.5%

    Core substrate supplier for NVIDIA; a three-year agreement to boost capacity by 2.5x directly drives revenue growth.

  • Amkor TechnologyAMKR.US0%

    NVIDIA's advanced packaging partner; packaging capacity to expand over 300% by end of 2028, with high order visibility.