enCharge Raises $100M+ to Advance AI with Analog Chips

EnCharge AI Secures $100 Million in Series B Funding

EnCharge AI, a semiconductor company focused on developing analog memory chips specifically for artificial intelligence applications, has successfully completed a Series B funding round, exceeding $100 million. This investment was spearheaded by Tiger Global and is intended to accelerate the company’s growth trajectory.

Addressing the Cost of AI

The substantial funding arrives during a period of unprecedented interest in AI technologies. However, the considerable expense associated with both building and maintaining AI services remains a significant concern. EnCharge, originating from research at Princeton University, proposes that its analog memory chips will offer a solution.

These chips are designed for integration into a variety of devices, including laptops, desktop computers, mobile handsets, and wearable technology. The company anticipates that its technology will not only enhance AI processing speeds but also contribute to a reduction in overall costs.

Energy Efficiency and Market Entry

Based in Santa Clara, EnCharge asserts that its AI accelerators consume 20 times less energy when processing workloads compared to currently available chips. The company projects the initial availability of these chips later in the current year.

Strategic Importance and Government Support

This funding round is particularly noteworthy given the U.S. government’s emphasis on bolstering domestic innovation in hardware and infrastructure, with a specific focus on semiconductor technology. Successful execution of its plans could position EnCharge as a crucial component of this national strategy.

Funding Details and Valuation

This Series B represents a new infusion of capital for the company. A previous funding announcement from December 2023 was separate from this current round. Initial reports in May of the previous year indicated EnCharge’s intention to raise at least $70 million to facilitate business expansion.

During a discussion with TechCrunch, Naveen Verma, CEO and co-founder of EnCharge, declined to reveal the company’s current valuation. He also refuted data from PitchBook suggesting a $438 million post-money valuation from an October funding event.

Investor Landscape

While Verma did not disclose specific customer details, the diverse group of investors participating in the round provides insights into potential collaborations.

Alongside Tiger Global, the round includes investments from Maverick Silicon, Capital TEN (Taiwan), SIP Global Partners, Zero Infinity Partners, CTBC VC, Vanderbilt University, and Morgan Creek Digital. Returning investors such as RTX Ventures, Anzu Partners, Scout Ventures, AlleyCorp, ACVC, and S5V also participated.

Corporate and Government Backing

Several corporations have also invested in EnCharge, including Samsung Ventures and HH-CTBC – a joint venture between Hon Hai Technology Group (Foxconn) and CTBC VC. Previous backing came from the VentureTech Alliance.

Furthermore, the company has received support from government-affiliated investors such as In-Q-Tel (associated with the CIA) and Constellation Technology (a clean energy manufacturer). EnCharge has also been awarded grants from U.S. agencies including DARPA and the Department of Defense.

Collaboration with TSMC

Verma confirmed that EnCharge is maintaining a close working relationship with TSMC, which will be responsible for the manufacturing of the company’s initial chips.

He highlighted TSMC’s long-standing interest in his research, dating back to the early stages of EnCharge’s research and development. He emphasized that TSMC’s provision of access to advanced silicon technology is an exceptional circumstance.

A Distinct Approach: EnCharge and Analog Computing

EnCharge is pursuing a unique strategy within the artificial intelligence landscape, centering its efforts on analog computing. Currently, the majority of investment and development is directed towards processing chips utilized for AI training and inference on servers, benefiting companies like Nvidia and AMD.

A recent research paper from IBM highlights the core distinction of EnCharge’s methodology. IBM’s researchers posit that analog chips exhibit “no separation between compute and memory,” resulting in significant cost efficiencies when contrasted with conventional designs.

Similar to EnCharge, IBM’s findings suggest these chips are presently well-suited for inference tasks, though their efficacy for training remains limited. EnCharge’s chips are specifically designed for deploying existing AI models “at the edge,” with ongoing research aimed at broadening their applicability.

While both IBM and EnCharge are exploring analog solutions, a key advancement for EnCharge lies in its chip design, specifically its resilience to noise. This is a critical factor in ensuring reliable performance.

Verma explains that managing noise across billions of transistors is essential for functionality. However, traditional approaches can compromise efficiency. “The breakthrough we achieved was discovering how to render analog processing insensitive to noise,” he stated.

The company leverages a readily available component within the standard supply chain – geometry-dependent metal wires – which can be precisely controlled to achieve this noise resilience.

EnCharge distinguishes itself as a full-stack company, developing both the hardware and the necessary software infrastructure.

The expertise of the founding team – Verma, COO Echere Iroaga, and CTO Kailash Gopalakrishnan – who previously worked at Macom and IBM respectively, is a significant asset. However, navigating the competitive landscape remains a challenge, with companies like Mythic and Sagence also active in the analog chip space.

Jimmy Kan, an investment partner at Anzu Partners, notes the extensive exploration his firm has undertaken in this area. “We’ve examined over 50 companies in this space since 2017, and likely more than 50 since then.”

Kan emphasizes Anzu’s focus on identifying truly differentiated AI compute technologies. “We sought an AI compute technology that was genuinely innovative, rather than incremental, and not something Nvidia could readily replicate,” he added. “We are very impressed with the advancements EnCharge has demonstrated.”

EnCharge’s trajectory contrasts with the typical development path of many deep tech startups.

The past quarter-century has witnessed a surge in venture capital available to support startups aiming to become the next major tech giants – Google, Microsoft, Apple, Meta, or Amazon. This has expanded the overall startup ecosystem.

This expansion includes a growing number of “deep tech” ventures, often founded by talented individuals seeking funding for promising ideas before fully developed, market-ready products. Quantum computing serves as a prime example of a “deep tech” category.

EnCharge could have followed this conventional path, emerging from Princeton earlier and pursuing funding to develop its chip innovation.

However, the startup deliberately delayed its independent launch. It wasn’t until 2022, nearly a decade after the initial research began at Princeton, that EnCharge publicly emerged and began securing commercial partnerships while continuing technological development.

Verma explains that certain innovations benefit from early venture backing, but fundamentally new technologies require thorough understanding and risk mitigation. “Taking venture funding shifts your priorities… The focus moves from technological understanding to customer needs,” he said.

Stay informed with TechCrunch’s AI newsletter! Subscribe here to receive it weekly in your inbox.