The artificial intelligence boom is driving demand for artificial intelligence chips, which are purpose-built to train and operate artificial intelligence models. And the big players, from VCs to startups, are trying to get in on the ground floor.
SoftBank’s Masayoshi Son is According to reports is seeking to raise $100 billion for a chip initiative that would rival tech giant Nvidia. OpenAI, meanwhile, is he said to be in talks with investment firms to start a business making artificial intelligence chips.
AI chip launch Axelera has kept a comparatively low profile. But it managed to win over backers, including Samsung, in part by focusing on a niche in the burgeoning AI chip market: chips that run AI on extreme devices.
“There is no doubt that the AI industry has the potential to transform many sectors,” Fabrizio Del Maffeo, one of Axelera’s co-founders and CEO, told TechCrunch. “However, to truly realize the value of AI, organizations need a solution that delivers high performance and efficiency while balancing cost.”
Axelera – based in the Netherlands, with a workforce of around 180 people spread across offices in Belgium, Switzerland, Italy and the UK – designs AI-powered chips and systems for applications such as security, retail, the automotive and robotics industry supplying B2B edge computing manufacturing partners and internet of things products.
Axelera was born out of an effort led by Del Maffeo and a team at Imec, the Belgium-based technology lab, along with Evangelos Eleftheriou and a team of Zurich-based IBM researchers, to create a high-performance chip architecture AI. The founding team incubated much of Axelera within the Bitfury Group, a blockchain company specializing in Bitcoin hardware.
The defining features of Axelera’s AI hardware stack are RISC-V instruction set architecture (ISA) and in-memory computing.
ISAs are a chip-based technical specification that describes how software controls the chip’s hardware. Chip designers typically license an existing ISA from a major chip manufacturer such as Arm or Intel, but RISC-V presents an open, royalty-free alternative. In terms of in-memory computing, it refers to performing computations on a system PLUNGER to reduce the latency introduced by storage devices.
Axelera is not the first to try its hand at a memory-based and/or RISC-V architecture for AI chips.
NeuroBlade develops chips that combine computing and memory into a single block of hardware for data processing. MemVerge, GigaSpaces, Hazelcast and H20.ai also offer in-memory hardware solutions for artificial intelligence and data analytics applications. Elsewhere, Tenstorrent, backed by Hyundai Motor Group and Samsung, sells AI processors and other related IP built around RISC-V.
Axelera tried to differentiate itself by providing both hardware chips and software to manage and develop AI models on this hardware. And from the looks of it, the strategy seems to be working for that.
Axelera announced Thursday that it has closed a $68 million Series B funding round that brings its total raised to $120 million. Contributors to the round include the European Innovation Council Fund, the Innovation Industries Strategic Partnership Fund, Invest-NL and the Samsung Catalyst Fund.
The new cash will be used to expand into new markets ahead of full production of Axelera’s flagship Metis AI platform in the second half of 2024, according to Del Maffeo. Axelera also has an eye on the data center chip market, with preliminary plans to fund chip R&D aimed at high-performance computing use cases.
“Metis went into full production in the second quarter and will be delivered in volume in the third quarter,” said Del Maffeo. “Axelera AI is now developing a new generation of products for computer vision, large language models and large multimodal models. This new product family will be introduced later this year and will go into full production in 2025.”
The challenge will be to send its AI chips to scale — and compete with countless others in the AI chip race. Many opponents have terrible support. a Crunchbase report as of June finds that VC-backed chip startups have raised nearly $5.3 billion in just 175 deals so far this year.
But the payoff could be significant. According to Statista and Market.us data, the The AI chip market could reach $67 billion in revenue by 2027. Axelera has little chance of unseating incumbents like Nvidia anytime soon, if ever. (Nvidia has one is appreciated 70% to 95% share of the AI chip market, per Mizuho Securities.) But capturing even a fraction of the market would be a major win.
“The funding supports our mission to democratize access to artificial intelligence, from the edge to the cloud,” Del Maffeo said, adding that Axelera has “dozens” of enterprise customers. “By expanding our product lines beyond the edge computing market, we are able to address industry challenges in AI inference and support current and future AI processing needs.”
