How did RISC-V become the third largest processor architecture?

While the current economy and event business aren't strong, the RISC-V Summit drew about 1,000 people to San Jose, Calif., this week to hear the latest news on the open source processor. RISC-V International CEO Calista Redmond said the numbers were down from last year, but she attributed part of that to conference attendance budgets hit by the economic downturn and lingering COVID concerns.

The nonprofit's membership has been steadily growing, from a group of energetic academics to some of the largest tech companies like Google and Nvidia. Over the past decade, the organization has established RISC-V as a viable alternative to proprietary Arm and Intel-based processors. Last year, Deloitte predicted that RISC-V chip revenue would grow from $400 million in 2021 to $1 billion in 2024.

RISC-V International's membership grew by more than 26% year-on-year, with more than 3,180 members in 70 countries/regions. The RISC-V community has approved six extensions and specifications in the past year and expects to approve 10 more specifications and extensions in the next 120 days.

RISC-V International now has 81 technology groups. More than 30 such teams will be added in 2022 alone, focusing on fast-growing markets for RISC-V such as security, system-on-chip infrastructure, automotive and AI/ML.

One of the advantages of RISC-V is that it is exempt from sanctions. As an open source platform, RISC-V is not subject to export restrictions. Deloitte said this makes it attractive to companies that are already affected by or fear being affected by these restrictions, especially in China. Deloitte said several companies are planning to use it for different storage, graphics and machine learning applications. Even Intel is working with RISC-V player SiFive. Arm claims it has more features and more support options for developers. But Redmond noted that competitors could be more expensive.

Redmond noted that because RISC-V is royalty-free, it's not easy to do the accounting and figure out the exact number of RISC-V chips that have been shipped. But it's estimated that there are about 10 billion cores on the market. It's small compared to x86 and Arm, but it's growing rapidly and the RISC-V opportunity is spreading.

Redmond noted that Ventana, SiFive and Andes all have ambitions to penetrate the world's data centers with RISC-V processors that can handle specialized tasks that other processors cannot.

The following is an edited transcript of the interview:

VentureBeat: How many people are coming this week?

Calista Redmond: There are over a thousand of us. We don't have final numbers yet.

VentureBeat: Is this similar to the past?

Redmond: In terms of composition, I think there are some similarities. We saw a little bit lower than we would have liked, but I think that's primarily due to some of the budget constraints that the company had going into the fourth quarter and probably some of the concerns and caution around COVID. We require everyone to be vaccinated to attend our events. We're trying to maintain a lot of protocols this way.

VentureBeat: Was there a year where you canceled events?

Redmond: We held it online for a year. Last year we were with DAC, and then the year before that we were meeting online. 2019 was the last independent event we held here. I think our numbers are very consistent with what others in the industry are seeing in terms of in-person events and the distribution of in-person and virtual.

VentureBeat: In terms of the big theme, is RISC-V still making progress on other [chip architectures]?

Redmond: Oh, of course. Absolutely. It’s a combination of a lot of new, uncharted and green territory. More consideration is given to some of these existing workloads. This year, you're seeing a lot of announcements that span the deeper performance end of the computing spectrum, not just one end. and further forays into some industries. The car just kept taking off. In HPC, we're seeing a lot of traction.

A lot of people are excited about data centers. The data center is a tough area to conquer. This is also an area where there are both existing workloads and some new capabilities being introduced into the data center. Hyperscale, other large implementations, bring acceleration. AI, ML, those types of capabilities, and the continued expansion and coordination of IOT and other types of endpoint technologies.

VentureBeat: Ventana, SiFive, they seem to be interested in challenging the big guys and which are the largest processors.

Redmond: A lot of our members are focused on getting some of these high-performance aspects. You've named some of them. Andes is also on that list. They have some higher performance stuff now. I don't know that all those high performers are challenging the incumbents, but there are so many opportunities. There is room for everyone.

VentureBeat: I saw David Patterson's post about fallacies. That's pretty cool. He is very capable of explaining these things. The neutral Swiss argument still seems to be very important.

Redmond: We're still there. I don't think that's the most important thing these days. We've been there for a few years.

VentureBeat: But architecture is a neutral concept—

Redmond: I don’t know what this has to do with Switzerland. When you think about global standards, it's effectively public domain everywhere, regardless of any merging entity behind it. Some standards may not even have a consolidated entity behind them. At RISC-V, we are the stewards of this specific global standard and its related extensions and specifications. Once you bring something into the public domain, it is no longer held by any one company or country. RISC-V, which is in the public domain as a global standard, does not belong to Switzerland. As far as RISC-V entities are concerned, people like Switzerland, but it has nothing to do with technical deliverables.

VentureBeat: But you're drawing attention to the fact that there are so many issues associated with other closed architectures that have some very geographical issues.

Redmond: Proprietary structures held by a single company. RISC-V is not – the architecture is not owned by one company. The RISC-V architecture contributes to the public domain. It's similar to Creative Commons or other global standards like Ethernet. Anyone, anywhere can describe their design as Ethernet. This has nothing to do with Switzerland. It’s about how we in the public sphere are tied to proprietary architecture with destiny tied to specific companies.

VentureBeat: That seems to explain one reason for the momentum behind it. But are there a lot of other things that could explain some of the momentum and why it's growing?

Redmond: Huge opportunity. If you're making a business case for where you want to go strategically, you want to have the design flexibility and freedom to get there. This huge opportunity is coming true. There are already over 10 billion cores on the market. Many of these implementations are opportunities in the green field. Is this a new opportunity? Are there opportunities to take existing designs and enter adjacent markets?

When you build a business case, it's about the huge opportunity in a high-growth space. We see this in semiconductors. It's time to go public. Can you speed up this path? Well, you've accelerated past a lot of barriers to entry. When you start with an open global standard like RISC-V, the initial investment cost of entering the field is easier to overcome. That's one of those areas.

The next area to consider in the business case is, can you differentiate? Can you become extremely competitive compared to your peers? Well, you can't do that if you're locked into proprietary constraints. You can do this by taking advantage of the design flexibility and freedom you get in RISC-V. By the way, in addition to being able to do what you want, it's also designed for this. It's designed so you put software acceleration first, giving you tight control over power consumption, code size, and various design elements. Vector is a great example of what we introduced as a very competitive differentiator and you can do things on RISC-V that you can't do elsewhere. If you're looking to grab some explosive opportunities, you'll want to be fast and as competitive as possible.

VentureBeat: For open source standardized things, people often say that this road is slower, and using something that is known and proprietary can speed up your time to market. It sounds like you would think this is a fallacy. In a sense, no one owning it might be a recipe for no one taking responsibility.

Raymond: No, it's a vested interest for everyone. The way it works, you don't sit back and wait for the community to develop your product. You're accelerating and engaging in hard problems that matter to you. All these building blocks are there. You take those, as a separate organization or entity or company or researcher, or what have you, and say, "Okay, I'm going to take this nugget and increase it through my competitive differentiation, and then I Go to market." You don't have to spend millions of dollars negotiating for years to get that building block. Architecture licenses can be a barrier to entry into new fields. I'm not sure this is a slower path unless there's a foundational nugget you're waiting for.

Now, that's not to say we're 100% done with all elements. We have a lot of work going on. For those elements that are of strategic importance to you, this is an opportunity to start it yourself and support it until eternity, or participate in a working group to get approved extensions or specifications to be built modularly into your design. You can do both. You can make your own proprietary extension for now, then come back and introduce another part in your next generation.

But what you might also appreciate is that you want the design to have long-term strategic durability. This means you don’t want your product or solution to become obsolete in the next generation. It's this compatibility and interoperability and all the other great features that you want to persist over time because you're not just using your product in isolation either. You're also trying to fit into the larger ecosystem.