Now led by Jim Keller, the company has built a new leadership team and a new strategy. It has tremendous potential. Now it must execute.
It is hard to believe the difference a year makes. In 2021, there were over 100 public and venture-backed startups with the same mission, to compete with NVIDIA in producing the fast chips needed to create and run artificial intelligence (AI). Fast forward to 2023, and now many companies are struggling to gain market traction or acquire enough capital to keep going. Part of the problem is undoubtedly the global economy; many AI adopters and investors do not have the resources or courage to give new chips a chance. But the real culprit is NVIDIA; they are proving a lot harder to beat than many companies and their investors ever imagined. So why is Tenstorrent, a Toronto-based startup, any different? Why should we believe that Tenstorrent could succeed where so many are struggling and even failing? This paper will explore what distinguishes Tenstorrent from the scores of other startups from leadership, strategy, and technology perspectives.
If an AI startup isn’t scared, it doesn’t get it.
Do we really need yet another AI hardware startup? Over the last five years, the industry has been flooded with over 100 such firms. Some have already closed shop, realizing that NVIDIA technology for data center AI is tough to beat. Consequently, investors have become far more cautious. Realistically, all these companies are trying to vie for a spot as a second source for NVIDIA in AI data center processing for training and inference processing. Could they win? In our opinion, they should be thrilled if they could get a combined 10% of the data center AI pie over the next three years. Yes, NVIDIA is THAT good.
Into this storm enters Tenstorrent, the Toronto-based AI Hardware startup with offices in the Bay Area, Austin, and Tokyo, Japan. Over the last year, the company has begun to expand from early-phase research and development into becoming a real company on a mission, with marketing, sales, support, and functional area execs adding to the engineering talent the company has been recruiting. The company has now grown to over 280 employees.
From a leadership perspective, the legendary CPU designer (including at Apple, AMD, Tesla, and Intel) and early angel investor Jim Keller has recently assumed the CEO role. At the same time, founder Ljubisa Bajic has stepped back into an advisor role. The company has hired David Bennett as CCO, Keith Witek as COO, Matthew Mattina as VP of Machine Learning, Wei-Han Lien as Chief CPU Architect (also from Apple), and Olof Johansson as VP of OS and Infrastructure, and Mamoru Nakano to head sales in Japan. Now, Raja Koduri, ex-Intel and AMD, has just joined the Board of Directors.
The Emerging Compute Landscape: A Rich Opportunity
The data center is evolving rapidly, with the move to cloud-based resources coinciding with the move to accelerators for HPC and AI. NVIDIA’s data center revenue has recently received another shot in the Arm from the explosion of interest in ChatGPT and the new AI battle between Microsoft and Google for the future of search and the applicability of AI to productivity applications. Last quarter, NVIDIA data center sales reached $3.8B, growing at 31%. Markets and Markets projects that the global data center accelerator market size could reach $64.0 billion by 2027, growing at a CAGR of 24.7% during the forecast period.
Meanwhile, the market for CPUs is starting to fracture considerably, with AMD growing from 10.7% to 17.6% in the data center in the latest quarter at Intel’s expense. Meanwhile, Arm continues to grow, and Gartner projects to achieve some 19% of server shipments in 2026.
Now, along comes open-source RISC-V, which is penetrating the microcontroller market and has ambitions in the data center as well. Semico Research projects RISC-V will account for 25 billion AI SoCs By 2027. That’s a lot of silicon.
Looking at AMD and Intel, AI Accelerator SoCs will increasingly become integrated on die with CPU cores to perform scalar and management operations. Tenstorrent believes that it should develop its own RISC-V cores instead of relying on a third party. If Tenstorrent can deliver superior RISC-V cores, it could create a second revenue stream. So, when one combines the $64B in AI accelerators with the 25 billion RISC-V SoCs, the Tenstorrent opportunity becomes more evident.
The Tenstorrent Strategy
Jim Keller has worked closely with infrastructure buyers for years; he knows what the customer wants. Simply put, they want an open platform for AI compute that is easy to deploy at scale and provides lower TCO than current alternatives. Tenstorrent has an AI chip today, a roadmap for the future that promises those benefits, and the credibility and track record of Mr. Keller to pull it off.
Tenstorrent is different from the rest of the field and may have a higher probability of success. First, the company has a software strategy that encourages open-source community innovation. Second, Tenstorrent is the only startup with an AI accelerator and RISC-V CPU designs and ambitions. Finally, Tenstorrent has attracted a world-class engineering team, and the company is now led by perhaps the industry’s best-known CPU designer, Jim Keller.
While we wish we had better insights into the eventual product mix and revenues Tenstorrent might realize, we believe that the RISC-V and AI Accelerators technology has performance and TCO advantages over the company’s many competitors. AMD, Intel, and NVIDIA have already embraced the idea of combining GPU and CPU; all three point to the memory and throughput advantages a combination can offer. These advantages are significant to meeting large language models’ needs for training and inference processing. Tenstorrent is the only startup we know of that can engage at this level of technology integration.
And having what could be the best RISC-V cores available to large buyers as IP or chiplets could create another revenue stream or an attractive exit strategy. Consequently, when anyone asks us who looks good to be able to compete in this new world, we invariably point to Tenstorrent as a likely winner.
For more information about Tenstorrent, please see our full report on our site here.
And for a lively conversation with Raja Koduri on what he thinks about Tenstorrent and RISC-V, check out this interview.