An $80 million check is a bet on a new kind of power cord for the AI boom. In November 2025, Amperesand, a Singapore-founded deep-tech startup, closed a Series A round co-led by Walden Catalyst Ventures and Temasek [Yahoo Finance/Business Wire, Nov 2025]. The capital is earmarked for a single, concrete goal: deploying 30 megawatts of commercial power systems in 2026, with a focus on hyperscale AI data centers [Yahoo Finance/Business Wire, Nov 2025]. The company is not selling compute. It is selling the electricity that makes compute possible, and it claims its hardware can deliver it faster, in less space, and with greater reliability than the century-old transformer technology it aims to replace.
The Wedge: Silicon-Carbide vs. The Grid
Amperesand’s product is a medium-voltage solid-state transformer (SST). In essence, it is a power conversion system that takes high-voltage electricity from the grid and steps it down to the usable low-voltage power that feeds server racks. The traditional method uses large, heavy, oil-filled transformers and accompanying switchgear. Amperesand’s alternative uses silicon-carbide power modules and proprietary high-frequency transformers to perform the same function in a fraction of the space [Bloomberg, Jul 2025]. The company claims its systems achieve a conversion efficiency of over 98.5%, occupy 50% less physical footprint, and can be powered up five times faster than conventional setups [Amperesand, retrieved 2026]. For a hyperscaler racing to bring a new data hall online to serve AI workloads, those are not minor spec-sheet improvements. They are potential determinants of speed to revenue and capital efficiency.
A Spinout With Institutional Backing
The company’s origins are as deliberate as its technology. Amperesand was incubated by Xora Innovation, Temasek’s deep-tech investment platform, and spun out of the Energy Research Institute at Nanyang Technological University (NTU) in 2023 [LinkedIn, retrieved 2026]. It was the first spinout under a S$75 million pilot program between NTU, the National University of Singapore, and Temasek [The Business Times, retrieved 2026]. This institutional pedigree is reflected in its cap table. The Series A saw participation from new investors Industry Ventures, Acclimate Ventures, and SG Growth Capital, while existing backers Xora Innovation, Material Impact, TDK Ventures, and Foothill Ventures also re-upped [Yahoo Finance/Business Wire, Nov 2025]. Total disclosed funding now sits at approximately $122.45 million [CB Insights, retrieved 2026].
The leadership team carries a heavy industrial pedigree. CEO Brian Dow, cited in the Series A announcement, leads the company [Yahoo Finance/Business Wire, Nov 2025]. The founding team includes veterans from ABB, General Electric, Siemens, and Vestas [Dealroom, retrieved 2026]. Gary Lawrence, identified in some sources as CEO, is a former global president for the power and grid segment at Schneider Electric with prior roles at ABB [AI Online, Jan 2025]. Phil Inagaki, Managing Partner of Xora Innovation, is co-founder and chairs the board [ZoomInfo, retrieved 2026]. This collective resume suggests a group built to navigate the sales cycles and technical specifications of utility and infrastructure buyers.
The Commercial Push and Open Questions
The planned 30 MW deployment for 2026 represents the company’s first volume shipments and its transition from pilot to commercial scale [Dataintelo.com, retrieved 2026]. The target customer is unambiguous: hyperscale AI data centers. A November 2026 account from Cornell’s business school noted the company was working with hyperscalers on product pilots, though it did not name them [Cornell SC Johnson College of Business, Nov 2026]. The absence of named customer announcements is typical for deep-tech hardware companies in early deployment, but it remains the critical variable for 2026.
The company’s hiring posture indicates a scaling operation. Recent job postings, active into late 2026, show a focus on applications engineering, test engineering management, and product compliance roles in locations including Singapore, San Francisco, and Reno, Nevada [Greenhouse, Oct 2026]. These are not research positions. They are roles for deploying, supporting, and certifying industrial equipment.
The Risk in a Crowded Field
Amperesand’s bet is substantial, but the path is lined with execution risks inherent to capital-intensive hardware. The competitive landscape for advanced power conversion is not empty, though sources did not name direct competitors. The primary challenge is not just technical performance, but commercial validation at scale.
- Proving Reliability. Claiming 99.999% uptime via proprietary redundancy is one thing [Amperesand, retrieved 2026]. Proving it over thousands of operating hours in a mission-critical data center is another. The sales cycle for core electrical infrastructure is measured in years, not months.
- Manufacturing Scale. Going from pilot production to volume manufacturing of complex power electronics is a notorious bottleneck. The 30 MW target is a first test of this capability.
- Economic Moats. With 18 patent families cited, the company is building an IP moat [Amperesand, retrieved 2026]. The durability of that advantage against large incumbents and other well-funded startups will depend on continuous innovation and manufacturing cost reduction.
The company’s answer to these risks appears to be its combination of academic IP, seasoned industrial leadership, and patient, deep-pocketed capital from firms like Temasek and Walden Catalyst. The $80 million Series A is not seed money for R&D. It is deployment capital. Walden Catalyst and Temasek are betting that Brian Dow and his team can convert a laboratory-validated technology into a certified, shipped product that meets the urgent power-density demands of the world’s largest cloud companies.
The next twelve months will be measured in megawatts deployed and logos landed. If Amperesand hits its 30 MW target with hyperscale customers, a Series B at a meaningfully higher valuation becomes a straightforward conversation. If deployment timelines slip or performance in the field falters, the conversation changes. For now, the bet is clear: the AI infrastructure stack is being rewritten, and the power layer is next. Can a Singapore spinout become the default plug for the next generation of data centers?
Funding History
| Round | Date | Amount | Lead Investor(s) |
|---|---|---|---|
| Seed | Sep 2023 | $12.45M | Xora Innovation [Caplight, retrieved 2026] |
| Series A | Nov 2025 | $80M | Walden Catalyst Ventures, Temasek [Yahoo Finance/Business Wire, Nov 2025] |
| Total Disclosed | ~$122.45M |
Sources
- [Yahoo Finance/Business Wire, Nov 2025] Amperesand Raises $80M Series A Co-led by Walden Catalyst Ventures and Temasek | https://finance.yahoo.com/news/amperesand-raises-80m-series-co-140000819.html
- [Amperesand, retrieved 2026] Company Website | https://www.amperesand.io/
- [LinkedIn, retrieved 2026] Amperesand Company Page | https://www.linkedin.com/company/amperesandpteltd
- [Bloomberg, Jul 2025] Amperesand Pte Ltd Company Profile | https://www.bloomberg.com/profile/company/2370850D:SP
- [CB Insights, retrieved 2026] Amperesand Company Profile | https://www.cbinsights.com/company/amperesand
- [The Business Times, retrieved 2026] Article on NTU, NUS, and Temasek spinout program | https://www.businesstimes.com.sg
- [Dealroom, retrieved 2026] Amperesand Company Profile | https://dealroom.co/companies/amperesand/
- [AI Online, Jan 2025] Article on Gary Lawrence | https://ai-online.com
- [ZoomInfo, retrieved 2026] Phil Inagaki Profile | https://www.zoominfo.com
- [Cornell SC Johnson College of Business, Nov 2026] A summer powering the future of AI infrastructure | https://business.cornell.edu/centers/2026/11/03/powering-the-future-of-ai-infrastructure/
- [Greenhouse, Oct 2026] Amperesand Careers Page | https://job-boards.greenhouse.io/amperesand/jobs/4448323009
- [Caplight, retrieved 2026] Amperesand Funding Details | https://caplight.com
- [Dataintelo.com, retrieved 2026] Article on Amperesand Series A | https://dataintelo.com