At a time when most artificial intelligence chip startups are struggling, Hailo received $120 million in financing to continue to compete with Nvidia

The funding environment for artificial intelligence chip startups, once as sunny as mid-July, is starting to look cloudy as Nvidia establishes its dominance.

According to a recent report, U.S. chip companies raised only $881 million from January to September 2023, down from $1.79 billion in the first three quarters of 2022. Artificial intelligence chip company Mythic ran out of cash in 2022 and was almost forced to cease operations. And once well-capitalized rival Graphcore is now facing mounting losses.

But one startup appears to be succeeding in the ultra-competitive and increasingly crowded world of artificial intelligence chips.

Hailo was co-founded in 2017 by Orr Danon and Avi Baum, the former chief technology officer for wireless connectivity at microprocessor company Texas Instruments, which designs specialized chips to run artificial intelligence workloads on edge devices. Hailo’s chips perform AI tasks with lower memory usage and power consumption than typical processors, making them a strong candidate for compact, offline and battery-powered devices such as cars, smart cameras and robots.

“I co-founded Hailo with a mission to deliver high-performance AI at scale outside of the data center world,” Danon told TechCrunch. “Our processors are used for tasks such as object detection, semantic segmentation, and AI-driven image and video enhancement. More recently, they have been used to run large language models (LLMs) on edge devices such as PCs and infotainment electronic control units. ).”

Many AI chip startups have yet to secure a major contract, let alone dozens or hundreds of contracts. Danon claims that Hailo currently has more than 300 customers in industries such as automotive, security, retail, industrial automation, medical equipment and defense.

Betting on Hailo’s future, a group of financial backers including Israeli businessman Alfred Akirov, car importer Delek Motors and venture capital platform OurCrowd invested $1.2 in Hailo this week. billion, which is an extension of the company’s Series C financing. Danon said the new capital will “enable Hailo to capitalize on all opportunities in the pipeline” while “laying the foundation for long-term growth.”

“We are strategically positioned to bring artificial intelligence to edge devices, significantly expanding the reach and impact of this remarkable new technology,” Danon said.

Now, you might be wondering, are startups like Hailo real Any chance against chip giants like Nvidia and, to a lesser extent, Arm, Intel and AMD? Christos Kozyrakis, a professor of electrical engineering and computer science at Stanford University, is one expert who agrees — he believes accelerator chips like Hailo will become “absolutely necessary” as artificial intelligence becomes more common.

“The energy efficiency gap between CPUs and accelerators is too big to ignore,” Kozyrakis told TechCrunch. “You can use accelerators to make critical tasks, such as artificial intelligence, more efficient, and one or two processors for programmability.”

Kozyrakis Do Longevity poses a challenge to Hailo’s leadership — for example, if the artificial intelligence model architecture its chips are designed to run efficiently loses popularity. Kozyrakis said software support could also become an issue if a large number of developers are unwilling to learn to use the tools built around the Hailo chip.

“Most of the challenges involving custom chips exist in the software ecosystem,” Kozyrakis said. “This is where Nvidia has a huge advantage over other companies in AI, for example, because they’ve been investing in the software they’ve architected for more than 15 years.”

But with $340 million in the bank and about 250 employees, Danon is confident in Hailo’s future — at least in the short term. He believes the startup’s technology can solve many of the challenges companies face with cloud-based AI inference, particularly latency, cost and scalability.

“Traditional AI models rely on cloud-based infrastructure and often face latency issues and other challenges,” Danon said. “They don’t have access to real-time insights and alerts, and their reliance on the network jeopardizes reliability and integration with the cloud, raising data privacy concerns. Hailo is addressing these challenges by providing solutions that run independently of the cloud, enabling them to Able to handle more AI processing.”

Curious about Danone’s perspective, I asked about generative AI and its heavy reliance on the cloud and remote data centers. Surely Hailo sees the current top-down, cloud-centric model (like how OpenAI operates) as an existential threat?

Instead, Danon said, generative AI is driving new demand for Hailo hardware.

“In recent years, we have seen a surge in demand for edge AI applications across most industries, from airport security to food packaging,” he said. “The new wave of generative artificial intelligence is further driving this demand, as we are not only seeing clients in the computing and automotive industries requesting locally processed LLMs, but also in areas such as industrial automation, security and more.”

How about that.