AI & RoboticsNews

Astera Labs raises $150M for chips that deliver AI in the cloud

Check out all the on-demand sessions from the Intelligent Security Summit here.


Astera Labs provides chips for connecting AI systems in the cloud, and today it has raised $150 million at a $3.15 billion valuation. It makes sure AI processor are fed enough data.

In addition to closing the fourth round of funding, Santa Clara, California-based Astera Labs expanded its board of directors, added new product lines to its portfolio, and grew its presence globally with new locations and an increased employee base.

These milestones represent the strong traction for its silicon, software, and system-level solutions that help its customers realize the vision of AI and machine learning in the cloud, said Sanjay Gajendra, chief business officer and cofounder, in an interview with VentureBeat.

Astera Labs got started about five years ago by a trio of former Texas Instruments engineers. Now it has more than 180 people.

Event

Intelligent Security Summit On-Demand

Learn the critical role of AI & ML in cybersecurity and industry specific case studies. Watch on-demand sessions today.


Watch Here

“We saw this big train of AI and machine learning that seem to be everywhere. The one problem that we saw was underrepresented was the connectivity problem,” Gajendra said. “That means even though there were several companies that were building big chips, no one really spoke about how to interconnect them to achieve a system that supported the new workloads. So that was the context and the spark that got us going.”

Astera Labs is providing CXL chip solutions, among other things.

In five years, the company has earned hundreds of design wins for its chips. It focuses on memory controller chips that handle CXL, PCIe, and Ethernet technologies. Compute Express Link (CXL) is an open standard for high-speed central processing unit-to-device and CPU-to-memory connections. It is designed for high-performance datacenter computers.

The company focusing on more efficient ways to interconnect various kinds of peripherals, accelerators, memory devices, and other electronics to a CPU and a GPU in a cloud server.

“This is similar to what USB did for laptops,” said Gajendra. “We are creating our Leo product line with multiple products for tasks such as memory expansion, memory pooling and memory sharing.”

Fidelity Management and Research led the round, and other investors include Atreides Management, Intel Capital and Sutter Hill Ventures.

Cloud servers need more of this kind of memory chip because of expontial growth in AI models, which drives the need for more processing and more memory to store data for the processors. As CPU cores rise, the bandwidth for memory has been decreasing, and memory channels have been limited.

The Leo Memory Connectivity Platform for CXL is a smart memory controller with a scalable memory interface that lets it support multiple CPUs in cloud servers. It seamlessly operates with major CPUs, GPUs and memory vendors. Astera Labs has a trio of product lines announced.

“Astera Labs continues to exceed expectations and more than execute on its vision to innovate
exponentially and be the Cloud industry’s trusted connectivity partner,” said Stefan Dyckerhoff,
managing director at Sutter Hills Ventures and board member of Astera Labs, in a statement. “I’m extremely impressed by the company’s ability to assert itself as the leader in highly competitive markets
and rapidly grow its value. I continue to be confident in its long-term growth potential.”

Astera Labs expands board

Astera Labs has raised $150 million at a $3.15 billion valuation.

Astera Labs has elected to its board of directors Alexis Black Bjorlin, vice president of infrastructure at Meta, and Michael Hurlston, president and CEO of Synaptics.

Black Bjorlin leads the infrastructure hardware teams at Meta to drive technology innovation at scale for Meta’s infrastructure. She brings to Astera Labs expertise in hardware, semiconductors, and systems, and in scaling operations.

Hurlston has served as president and CEO of Synaptics since 2019.

“I am pleased to welcome these visionary leaders to the Astera Labs’ Board of Directors,” said Manuel Alba, chairman of the board at Astera Labs, in a statement. “Their passion for transformative technology
combined with expertise in semiconductors, cloud and the data center industry will add valued insights to the board and help Astera Labs continue to scale exponentially.”

To support its rapidly growing product portfolio and operations, Astera Labs is hiring for several
positions in North America and Asia. The team includes industry veterans with deep experience
in connectivity products, protocols, customer-centric design, and high-volume manufacturing.

“Astera Labs continues to consistently surpass every milestone for a technology start-up, and we are now deep into the next stage of evolution for our company as we make strategic investments to scale for growth,” said Jitendra Mohan, CEO of, Astera Labs, in a statement. “This latest funding round is a testament that we are not only invested in the right growth markets such as cloud, artificial intelligence/machine learning, and hyperscale infrastructure, but that we are also able to consistently execute and deliver breakthrough connectivity products that are critical to our customers and partners.”

The company is also establishing new research and development centers in Toronto and Vancouver in Canada.

The funding round was done to support the continued growth of the company and help it get ready for becoming a public company someday, Gajendra said.

As far as funding rounds go, Gajendra believes the company caught the investment wave just in time before the downturn started affecting valuations of all companies.

“If you have the right, right ingredients, I truly believe it’s possible to raise funds,” he said.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.


Author: Dean Takahashi
Source: Venturebeat

Related posts
AI & RoboticsNews

Nvidia and DataStax just made generative AI smarter and leaner — here’s how

AI & RoboticsNews

OpenAI opens up its most powerful model, o1, to third-party developers

AI & RoboticsNews

UAE’s Falcon 3 challenges open-source leaders amid surging demand for small AI models

DefenseNews

Army, Navy conduct key hypersonic missile test

Sign up for our Newsletter and
stay informed!