NVIDIA is firing on all cylin­ders, achie­ving record reve­nues in Gaming, Data Cen­ter and over­all,” said Jen­sen Huang, foun­der and CEO of NVIDIA. “The new NVIDIA GeFor­ce RTX GPU pro­vi­des our lar­gest-ever genera­tio­nal leap and demand is over­whel­ming. NVIDIA RTX has made ray tra­cing the new stan­dard in gaming.

We are con­ti­nuing to rai­se the bar with NVIDIA AI. Our A100 com­pu­te plat­form is ram­ping fast, with the top cloud com­pa­nies deploy­ing it glo­bal­ly. We swept the indus­try AI infe­rence bench­mark, and our cus­to­mers are moving some of the world’s most popu­lar AI ser­vices into pro­duc­tion, powe­red by NVIDIA technology.

We announ­ced the NVIDIA DPU pro­gramm­a­ble data cen­ter pro­ces­sor, and the plan­ned acqui­si­ti­on of Arm, creator of the world’s most popu­lar CPU. We are posi­tio­ning NVIDIA for the age of AI, when com­pu­ting will extend from the cloud to tril­li­ons of devices.”

NVIDIA paid $99 mil­li­on in quar­ter­ly cash divi­dends in the third quar­ter. It will pay its next quar­ter­ly cash divi­dend of $0.16 per share on Decem­ber 29, 2020, to all share­hol­ders of record on Decem­ber 4, 2020.

Q3 Fis­cal 2021 Sum­ma­ry

($ in mil­li­ons, except
ear­nings per share)
Q3 FY21 Q2 FY21 Q3 FY20 Q/Q Y/Y
Reve­nue $4,726   $3,866   $3,014   Up 22% Up 57%
Gross mar­gin   62.6 %   58.8 %   63.6 % Up 380 bps Down 100 bps
Ope­ra­ting expenses $1,562   $1,624   $989   Down 4% Up 58%
Ope­ra­ting income $1,398   $651   $927   Up 115% Up 51%
Net inco­me $1,336   $622   $899   Up 115% Up 49%
Dilu­t­ed ear­nings per share $2.12   $0.99   $1.45   Up 114% Up 46%


($ in mil­li­ons, except
ear­nings per share)
Q3 FY21 Q2 FY21 Q3 FY20 Q/Q Y/Y
Reve­nue $4,726   $3,866   $3,014   Up 22% Up 57%
Gross mar­gin   65.5 %   66.0 %   64.1 % Down 50 bps Up 140 bps
Ope­ra­ting expenses $1,101   $1,035   $774   Up 6% Up 42%
Ope­ra­ting income $1,993   $1,516   $1,156   Up 31% Up 72%
Net inco­me $1,834   $1,366   $1,103   Up 34% Up 66%
Dilu­t­ed ear­nings per share $2.91   $2.18   $1.78   Up 33% Up 63%

NVIDIA’s out­look for the fourth quar­ter of fis­cal 2021 is as follows:

  • Reve­nue is expec­ted to be $4.80 bil­li­on, plus or minus 2 percent.
  • GAAP and non-GAAP gross mar­gins are expec­ted to be 62.8 per­cent and 65.5 per­cent, respec­tively, plus or minus 50 basis points.
  • GAAP and non-GAAP ope­ra­ting expen­ses are expec­ted to be appro­xi­mate­ly $1.64 bil­li­on and $1.18 bil­li­on, respectively.
  • GAAP and non-GAAP other inco­me and expen­se are both expec­ted to be an expen­se of appro­xi­mate­ly $55 million.
  • GAAP and non-GAAP tax rates are both expec­ted to be 8 per­cent, plus or minus 1 per­cent, exclu­ding any dis­cre­te items. GAAP dis­cre­te items inclu­de excess tax bene­fits or defi­ci­en­ci­es rela­ted to stock-based com­pen­sa­ti­on, which are expec­ted to gene­ra­te varia­bi­li­ty on a quar­ter-by-quar­ter basis.


During the third quar­ter, NVIDIA announ­ced a defi­ni­ti­ve agree­ment to acqui­re Arm Limi­ted from Soft­Bank Capi­tal Limi­ted and SVF Hold­co (UK) Limi­ted in a tran­sac­tion valued at $40 bil­li­on. The tran­sac­tion will com­bi­ne NVIDIA’s lea­ding AI com­pu­ting plat­form with Arm’s vast eco­sys­tem to crea­te the pre­mier com­pu­ting com­pa­ny for the age of AI. The tran­sac­tion ― which is expec­ted to be immedia­te­ly accre­ti­ve to NVIDIA’s non-GAAP gross mar­gin and non-GAAP ear­nings per share ― is expec­ted to clo­se in the first quar­ter of calen­dar 2022. 

NVIDIA also announ­ced plans to build a world-class AI lab in Cam­bridge, Eng­land ― inclu­ding a power­ful AI super­com­pu­ter based on NVIDIA and Arm tech­no­lo­gy ― and pro­vi­de rese­arch fel­low­ships and part­ners­hips with local insti­tu­ti­ons and AI trai­ning cour­ses. Sepa­r­ate­ly, it plans to build Cambridge‑1, the U.K.’s most power­ful AI super­com­pu­ter, based on an NVIDIA DGX Super­POD™ sys­tem and desi­gned for AI rese­arch in health­ca­re and drug discovery.

NVIDIA also achie­ved pro­gress sin­ce its pre­vious ear­nings announ­ce­ment in the­se areas:

Data Cen­ter

  • Third-quar­ter reve­nue was a record $1.90 bil­li­on, up 8 per­cent from the pre­vious quar­ter and up 162 per­cent from a year earlier.
  • Shared news that Amazon Web Ser­vices and Ora­cle Cloud Infra­st­ruc­tu­re announ­ced gene­ral avai­la­bi­li­ty of cloud com­pu­ting instan­ces based on the NVIDIA A100GPU, fol­lowing Goog­le Cloud Plat­form and Micro­soft Azure.
  • Announ­ced the NVIDIA DGX Super­POD Solu­ti­on for Enter­pri­se ― the world’s first turn­key AI infra­st­ruc­tu­re ― which is expec­ted to be instal­led by yea­rend in Korea, the U.K., India and Sweden.
  • Announ­ced that five super­com­pu­ters backed by EuroHPC ― inclu­ding “Leo­nar­do,” the world’s fas­test AI super­com­pu­ter built by the Ita­li­an inter-uni­ver­si­ty con­sor­ti­um CINECA ― will use NVIDIA’s data cen­ter acce­le­ra­tors or net­wor­king.
  • Intro­du­ced the NVIDIA BlueField‑2 DPU (data pro­ces­sing unit) ― sup­por­ted by NVIDIA DOCA™, a novel data-cen­ter-infra­st­ruc­tu­re-on-a-chip archi­tec­tu­re ― to bring bre­akthrough net­wor­king, sto­rage and secu­ri­ty per­for­mance to every data center.
  • Announ­ced a broad part­ners­hip with VMware to crea­te an end-to-end enter­pri­se plat­form for AI and a new archi­tec­tu­re for data cen­ter, cloud and edge using NVIDIA DPUs, bene­fi­t­ing 300,000-plus VMware customers.
  • Unvei­led NVIDIA Maxi­ne™, an AI video-strea­ming plat­form that enhan­ces strea­ming qua­li­ty and offers such AI-powe­red fea­tures as gaze cor­rec­tion, super-reso­lu­ti­on, noi­se can­cel­la­ti­on and face relighting.
  • Intro­du­ced the NVIDIA RTX A6000 and NVIDIA A40 GPUs, built on the NVIDIA Ampere archi­tec­tu­re and fea­turing new RT Cores, Ten­sor Cores and CUDA® cores.
  • Exten­ded its lead on MLPerf per­for­mance bench­marks for infe­rence, win­ning every test across all six app­li­ca­ti­on are­as for data cen­ter and edge com­pu­ting systems.
  • Announ­ced a part­ners­hip with GSK to inte­gra­te com­pu­ting plat­forms for ima­ging, geno­mics and AI into the drug and vac­ci­ne dis­co­very process.
  • Intro­du­ced at SC20, three power­ful advan­ces in AI tech­no­lo­gy: the NVIDIA® A100 80GB GPU, powe­ring the NVIDIA HGXAI super­com­pu­ting plat­form with twice the memo­ry of its pre­de­ces­sor; the NVIDIA DGX Sta­ti­on™ A100, the world’s only petas­ca­le work­group ser­ver, for machi­ne lear­ning and data sci­ence workloads; and the next genera­ti­on of NVIDIA® Mel­lanox® Infi­ni­Band, for the fas­test net­wor­king performance.