Nvidia plans subscription-fueled journey to $1tr revenue
(2022/03/25)
- Reference: 1648242447
- News link: https://www.theregister.co.uk/2022/03/25/nvidia_trillion_dollar_revenue/
- Source link:
GTC Nvidia has laid out a roadmap of sorts to a trillion dollars in revenue.
That ten-figure revenue projection has no timeline, and is ambitious considering the revenue from the most recent financial year was just $26.9bn, up 61 percent annually. The GPU giant's GTC event this week indicated its path is reliant on extracting repeat revenue from software that requires Nvidia's hardware.
GTC focused heavily on AI and graphics applications that cut across Nvidia's GPU, CPU, data processing, and automotive offerings. It is Nvidia's belief that software in the long run will generate more cash through subscriptions and upgrades than money coming in through one-time hardware shipments. To do this, Nv needs to continue providing a full stack of technology, from the transistors to the development kits and application frameworks.
[1]
Nvidia CFO Colette Kress broke down the $1tr revenue opportunity, saying $100bn could come from gaming, $300bn from chips and systems, $150bn from AI Enterprise software, $150bn from Omniverse Enterprise software, and $300bn from the automotive sector.
[2]
[3]
As an example, Nvidia is helping car makers to put Netflix-style subscription services to enable features like autonomous driving, from which Nv will presumably take some kind of cut. Mercedes Benz and Jaguar Land Rover will put autonomous cars with Nvidia's computers and software on the roads in 2024 and 2025 respectively.
"These services give OEMs an exciting opportunity to transform their business model like we have seen Tesla do with their Autopilot software that has grown in price from less than $5,000 to $12,000 a car today," said Ali Kani, vice president and general manager of Nvidia's automotive division, during an investor day presentation that ran alongside GTC.
[4]
Software represents a vast majority of the $300 billion automotive opportunity, said Colette Kress, CFO at Nvidia at the investor meeting.
"Our software content per vehicle can be in the thousands of dollars over the lifetime of the vehicle compared to the hundreds of dollars for the hardware. And second, software scales with the installed base of vehicles, not annual production," Kress said.
Nvidia's automotive business has three components: the Drive software stack for autonomous driving, in-vehicle hardware, and the data center infrastructure for training and simulation, which were all upgraded at GTC.
[5]
Pre-eminent auto chip players like Renesas and NXP are largely focused on components, but Nvidia's full-stack approach takes the stress off car makers on building full AI systems for cars. Companies like Intel and Qualcomm are also offering auto chips, but largely working with partners, like with the PC and smartphone markets.
Meta revenues
Nvidia is also placing big bets on the metaverse, a graphical 3D universe where users can play, interact and build. Nvidia sees opportunities in the hardware, and tools for creation and collaboration in the graphical universe.
But the metaverse opportunity could be expensive for Nvidia's consumer GPU users.
"The graphics required to deliver a cinematic VR experience in a massive multiplayer, physically accurate, world will likely require three to four orders of magnitude more than the performance of our highest-end GPUs." Jeff Fisher, senior vice president of gaming at Nvidia, during the investor conference.
He also threw in buzzwords about the virtual economies in the metaverse, with NFTs, billions of virtual real estate, and cryptocurrency as the backbone of it all.
"The GPU is offering more value than ever. Based on our data, they are spending $300 more than they paid for the graphics card they replaced," Fisher added.
There are 3 billion gamers globally, and the number is growing every year, and there is no difference in hardware (GeForce hardware) and software (GeForce Now in the cloud) revenue opportunities, CFO Kress said.
"The per user pricing is similar and annualizes at over $100 per year," Kress said.
Nvidia also expects to generate close to $150bn in subscription software through enterprise use of Omniverse, a catch-all branding for Nvidia's metaverse hardware and software offerings.
"We also estimate $150bn software opportunity based on two opportunities. First, a per seat software subscription for professional designers and creators, which we estimate at 45 million [users] and second, a per-robot software subscription for digital twins, based on more than 10 million factories and warehouses," Kress said.
Nvidia made clear that AI and metaverse software offerings run best on the company's GPUs, CPUs and other chips. Nvidia is also creating delivery mechanisms so the code can be delivered to Nvidia's chips that could on-premise, in data centers, cloud servers, or supercomputers.
At GTC, Nvidia also announced the Enterprise AI 2.0, which is "the operating system of AI," Manuvir Das, vice president of enterprise computing at Nvidia, said during the investor conference.
Compared to its predecessor, the new software has expanded hardware support with the ability to run on CPU or GPU, as opposed to only GPUs in the predecessor. The software will also run on both virtualized and bare metal servers running VMWare, Red Hat or other platforms in major public clouds, compared to only VMware support in its predecessor.
"For Nvidia AI Enterprise, we estimate the total available opportunity at $150bn based on the installed base of enterprise servers and our per server software pricing," CFO Kress said.
[6]Chip designers made bank in 2021 amid global shortage
[7]Nvidia CEO: We're open to Intel making our chips
[8]It takes big business to make Nvidia's Omniverse tangible
[9]How Nvidia is overcoming slowdown issues in GPU clusters
Nvidia is wrapping its full-stack offerings through "AI factories" like the EOS supercomputer, which will be up in a few months. EOS is meant to be a showcase of its hardware, which includes its CPU and networking stack. Nvidia at GTC announced the [10]Hopper H100 GPU , which will be in the EOS AI supercomputer.
Nvidia's software assets are built on the closed-source CUDA framework, which Nvidia considers a crowning jewel and a starting point on which Nvidia's GPUs are built. Nvidia's one-stop shop approach differs from that of open approaches by Intel and AMD with OpenCL, Intel's OpenAPI and AMD's ROCm, given the reliance on partners for system-integration.
Nvidia even further tightened its grasp on CUDA, making further chip [11]enhancements that ensure code written on it executes the fastest on its GPUs.
At GTC, the company announced more than 60 updates to its CUDA software libraries, including frameworks for quantum computing, 6G networks, robotics, cybersecurity, and drug discovery.
"With each new SDK, new science, new applications and new industries can tap into the power of Nvidia computing. These SDKs tackle the immense complexity at the intersection of computing algorithms and science," CEO Jensen Huang during a keynote on Tuesday.
The closed hardware approach creates a dilemma for potential buyers -- either go with Nvidia software and hardware or take alternative approaches. This consideration has already played out with some cloud buyers, with Google opting a "build" approach with its own AI chips, and Facebook [12]creating its metaverse with the "buy" approach with Nvidia's GPUs and AMD CPUs.
Nvidia also created faster communication lanes for third parties to send code to execute on its GPUs. The company opened up NVLink-C2C die-to-die interconnect so outside chips can connect to the its GPUs, CPUs, data-processing units and other chips. The new [13]Grace CPU Superchip , also announced at GTC, has CPUs connected using the NVLink-C2C interconnect. ®
Get our [14]Tech Resources
[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2022/03/24/fabless_chip_designers_made_bank/
[7] https://www.theregister.com/2022/03/24/nvidia_intel_chips/
[8] https://www.theregister.com/2022/03/23/nvidia_omniverse_gtc/
[9] https://www.theregister.com/2022/03/23/nvidia_slowdown_gpu/
[10] https://www.theregister.com/2022/03/22/nvidias_h100_gpu/
[11] https://www.theregister.com/2022/03/23/nvidia_parallelism_cuda/
[12] https://www.theregister.com/2022/01/25/meta_supercomputer_metaverse/
[13] https://www.theregister.com/2022/03/22/nvidia_reveals_epyccrushing_144core_grace/
[14] https://whitepapers.theregister.com/
That ten-figure revenue projection has no timeline, and is ambitious considering the revenue from the most recent financial year was just $26.9bn, up 61 percent annually. The GPU giant's GTC event this week indicated its path is reliant on extracting repeat revenue from software that requires Nvidia's hardware.
GTC focused heavily on AI and graphics applications that cut across Nvidia's GPU, CPU, data processing, and automotive offerings. It is Nvidia's belief that software in the long run will generate more cash through subscriptions and upgrades than money coming in through one-time hardware shipments. To do this, Nv needs to continue providing a full stack of technology, from the transistors to the development kits and application frameworks.
[1]
Nvidia CFO Colette Kress broke down the $1tr revenue opportunity, saying $100bn could come from gaming, $300bn from chips and systems, $150bn from AI Enterprise software, $150bn from Omniverse Enterprise software, and $300bn from the automotive sector.
[2]
[3]
As an example, Nvidia is helping car makers to put Netflix-style subscription services to enable features like autonomous driving, from which Nv will presumably take some kind of cut. Mercedes Benz and Jaguar Land Rover will put autonomous cars with Nvidia's computers and software on the roads in 2024 and 2025 respectively.
"These services give OEMs an exciting opportunity to transform their business model like we have seen Tesla do with their Autopilot software that has grown in price from less than $5,000 to $12,000 a car today," said Ali Kani, vice president and general manager of Nvidia's automotive division, during an investor day presentation that ran alongside GTC.
[4]
Software represents a vast majority of the $300 billion automotive opportunity, said Colette Kress, CFO at Nvidia at the investor meeting.
"Our software content per vehicle can be in the thousands of dollars over the lifetime of the vehicle compared to the hundreds of dollars for the hardware. And second, software scales with the installed base of vehicles, not annual production," Kress said.
Nvidia's automotive business has three components: the Drive software stack for autonomous driving, in-vehicle hardware, and the data center infrastructure for training and simulation, which were all upgraded at GTC.
[5]
Pre-eminent auto chip players like Renesas and NXP are largely focused on components, but Nvidia's full-stack approach takes the stress off car makers on building full AI systems for cars. Companies like Intel and Qualcomm are also offering auto chips, but largely working with partners, like with the PC and smartphone markets.
Meta revenues
Nvidia is also placing big bets on the metaverse, a graphical 3D universe where users can play, interact and build. Nvidia sees opportunities in the hardware, and tools for creation and collaboration in the graphical universe.
But the metaverse opportunity could be expensive for Nvidia's consumer GPU users.
"The graphics required to deliver a cinematic VR experience in a massive multiplayer, physically accurate, world will likely require three to four orders of magnitude more than the performance of our highest-end GPUs." Jeff Fisher, senior vice president of gaming at Nvidia, during the investor conference.
He also threw in buzzwords about the virtual economies in the metaverse, with NFTs, billions of virtual real estate, and cryptocurrency as the backbone of it all.
"The GPU is offering more value than ever. Based on our data, they are spending $300 more than they paid for the graphics card they replaced," Fisher added.
There are 3 billion gamers globally, and the number is growing every year, and there is no difference in hardware (GeForce hardware) and software (GeForce Now in the cloud) revenue opportunities, CFO Kress said.
"The per user pricing is similar and annualizes at over $100 per year," Kress said.
Nvidia also expects to generate close to $150bn in subscription software through enterprise use of Omniverse, a catch-all branding for Nvidia's metaverse hardware and software offerings.
"We also estimate $150bn software opportunity based on two opportunities. First, a per seat software subscription for professional designers and creators, which we estimate at 45 million [users] and second, a per-robot software subscription for digital twins, based on more than 10 million factories and warehouses," Kress said.
Nvidia made clear that AI and metaverse software offerings run best on the company's GPUs, CPUs and other chips. Nvidia is also creating delivery mechanisms so the code can be delivered to Nvidia's chips that could on-premise, in data centers, cloud servers, or supercomputers.
At GTC, Nvidia also announced the Enterprise AI 2.0, which is "the operating system of AI," Manuvir Das, vice president of enterprise computing at Nvidia, said during the investor conference.
Compared to its predecessor, the new software has expanded hardware support with the ability to run on CPU or GPU, as opposed to only GPUs in the predecessor. The software will also run on both virtualized and bare metal servers running VMWare, Red Hat or other platforms in major public clouds, compared to only VMware support in its predecessor.
"For Nvidia AI Enterprise, we estimate the total available opportunity at $150bn based on the installed base of enterprise servers and our per server software pricing," CFO Kress said.
[6]Chip designers made bank in 2021 amid global shortage
[7]Nvidia CEO: We're open to Intel making our chips
[8]It takes big business to make Nvidia's Omniverse tangible
[9]How Nvidia is overcoming slowdown issues in GPU clusters
Nvidia is wrapping its full-stack offerings through "AI factories" like the EOS supercomputer, which will be up in a few months. EOS is meant to be a showcase of its hardware, which includes its CPU and networking stack. Nvidia at GTC announced the [10]Hopper H100 GPU , which will be in the EOS AI supercomputer.
Nvidia's software assets are built on the closed-source CUDA framework, which Nvidia considers a crowning jewel and a starting point on which Nvidia's GPUs are built. Nvidia's one-stop shop approach differs from that of open approaches by Intel and AMD with OpenCL, Intel's OpenAPI and AMD's ROCm, given the reliance on partners for system-integration.
Nvidia even further tightened its grasp on CUDA, making further chip [11]enhancements that ensure code written on it executes the fastest on its GPUs.
At GTC, the company announced more than 60 updates to its CUDA software libraries, including frameworks for quantum computing, 6G networks, robotics, cybersecurity, and drug discovery.
"With each new SDK, new science, new applications and new industries can tap into the power of Nvidia computing. These SDKs tackle the immense complexity at the intersection of computing algorithms and science," CEO Jensen Huang during a keynote on Tuesday.
The closed hardware approach creates a dilemma for potential buyers -- either go with Nvidia software and hardware or take alternative approaches. This consideration has already played out with some cloud buyers, with Google opting a "build" approach with its own AI chips, and Facebook [12]creating its metaverse with the "buy" approach with Nvidia's GPUs and AMD CPUs.
Nvidia also created faster communication lanes for third parties to send code to execute on its GPUs. The company opened up NVLink-C2C die-to-die interconnect so outside chips can connect to the its GPUs, CPUs, data-processing units and other chips. The new [13]Grace CPU Superchip , also announced at GTC, has CPUs connected using the NVLink-C2C interconnect. ®
Get our [14]Tech Resources
[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Yj5JodvHwf0ImyOJBqFTTAAAAJQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2022/03/24/fabless_chip_designers_made_bank/
[7] https://www.theregister.com/2022/03/24/nvidia_intel_chips/
[8] https://www.theregister.com/2022/03/23/nvidia_omniverse_gtc/
[9] https://www.theregister.com/2022/03/23/nvidia_slowdown_gpu/
[10] https://www.theregister.com/2022/03/22/nvidias_h100_gpu/
[11] https://www.theregister.com/2022/03/23/nvidia_parallelism_cuda/
[12] https://www.theregister.com/2022/01/25/meta_supercomputer_metaverse/
[13] https://www.theregister.com/2022/03/22/nvidia_reveals_epyccrushing_144core_grace/
[14] https://whitepapers.theregister.com/