Reframe Systems raises $40 million to scale housing construction using artificial intelligence

Emerald AI has successfully raised $150 million in a re-signed Series A funding round, valuing the company at $1.05 billion. The funding was co-led by Energize Capital and DCVC, and attracted strategic investors from global energy and technology sectors.

The company currently includes 12 Fortune Global 500 enterprises among its investors, who also participate in Emerald AI's Strategic Advisory Board. With the capital raised, the company plans to accelerate the commercial deployment of its solutions worldwide.

Emerald AI's clients include data center operators, artificial intelligence companies, and electric utilities. The company develops software that enables data centers to become flexible electricity consumers. The Emerald Conductor platform manages AI workloads concurrently with the use of local energy resources.

The system automatically regulates energy consumption when grid load periods coincide, while ensuring the protection of critical AI computing tasks. This approach could potentially help data centers gain faster access to power sources and reduce the strain on local infrastructure during peak demand. According to Emerald AI, their technology is capable of freeing up over 100 gigawatts of capacity, referring to the existing US power grid.

The company notes that this capacity can become available before new infrastructure is built, as traditional energy grid projects can take many years to become operational, whereas the electricity demands of AI data centers are growing rapidly.

Emerald AI conducted five demonstration projects in the United States and London. These trials took place in commercial data centers in Arizona, Illinois, and Virginia, as well as in Oregon and London. The company aims to bridge the time gap between demand and availability through software. Partners in these projects included NVIDIA, EPRI, Oracle, Nebius, and National Grid.

Emerald AI has now moved beyond demonstrations and begun commercial deployment. Its software is fully operational in a data center in California, where flexible energy consumption was demonstrated during grid stress. The company also plans additional large-scale deployments, having partnered with Silicon Valley Power for a flexible interconnection program that gives data centers access to additional grid capacity in exchange for providing proven and managed energy flexibility.

Furthermore, Emerald AI is collaborating with Digital Realty and NVIDIA in Virginia on the development of the Vera Rubin AI Research Farm in Manassas, which is expected to have a capacity of around 100 megawatts, and its energy-flexible design has been tested.

Emerald AI CEO, Dr. Varun Sivaraman, stated that AI can help solve its own energy supply problem. He emphasized that the company's demonstrations showed the capability of precisely tuning data center energy consumption. The company's goal is to implement this capability wherever AI infrastructure is developing to strengthen grid reliability and support AI growth.

Energize Capital and DCVC noted that Emerald AI has quickly reached the stage of commercial deployment, and their investment reflects the growing demand for solutions meeting AI's energy requirements. Other participants in the funding round included NVIDIA, Samsung Ventures, Siemens, and GE Vernova. Salesforce Ventures, Aramco Ventures, and RWE also joined the round, strengthening the company's position at the intersection of AI infrastructure and energy, giving it the capital to expand deployments into new markets and assist utilities in managing growing demand.

Twin1 AI has successfully raised $20 million in a seed funding round while simultaneously moving its product out of stealth development. The company creates artificial intelligence-based digital twins designed for knowledge workers.

The funding round was co-led by venture capital firms Bessemer Venture Partners, Tribeca Venture Partners, and Aramco Ventures. The company plans to expand its teams in San Mateo and London, as well as invest in technology development and marketing activities. The Twin1 platform is designed to preserve professional knowledge, judgment, and work context.

The company's goal is to help organizations scale this expertise through artificial intelligence. Twin1's approach prioritizes privacy and human oversight when implementing enterprise AI. This provides the company with capital for developing commercial operations and technologies.

Twin1 provides every specialist with an AI-managed digital twin that evolves alongside their work. These twins have access to authorized emails, meetings, documents, and work systems. Using this context, they can respond to queries and assist in completing business tasks.

The company claims that its system can enhance existing AI agents through deeper professional context, leading to more personalized enterprise AI implementation. Twin1 has developed six levels of privacy and governance control. These mechanisms regulate interactions between humans, digital twins, and AI systems, integrating corporate policies, existing permissions, and human approval requirements, which prevents unauthorized access to confidential knowledge.

Furthermore, Twin1 manages a network that connects individual digital twins across different organizations. This allows for the identification of relevant expertise and coordination of work by gathering information while respecting permissions, yet maintaining human control. This forms a level of coordination between humans and enterprise AI agents. The company also offers a model context protocol server for enterprises, which grants approved AI agents and corporate tools access to controlled context.

The system can also initiate actions based on information contained within individual twins. Twin1 calls this approach the foundation for sovereign enterprise AI. The platform has already been implemented by partners in various industries, including legal, financial, and energy companies. Clients mentioned include Linklaters, Orrick, Dechert, Customers Bank, and Aegis Energy; the company reports that some clients have automated between 30% and 50% of communication work.

Twin1 was founded in 2025 by individuals such as Lewis Liu, Tom Cahn, Huiting Liu, and Jonathan Budd. The founders previously worked on enterprise AI through Eigen Technologies. Several investors from Eigen also participated in the new funding round. This support reflects continued confidence in the team's experience in enterprise technologies. CEO Lewis Liu emphasized that human expertise remains central to knowledge-based organizations, stating that AI should augment individual knowledge, not generate generic outputs.

Twin1 aims to preserve professional judgment while increasing the reach of each employee, believing this will help expertise accumulate across different organizations. Bessemer Venture Partners noted that corporate knowledge remains fragmented across organizations and believes Twin1 can provide a contextual layer between teams and systems. Tribeca Venture Partners described the platform as a coordination layer for enterprise AI, while Aramco Ventures highlighted the team's experience working with regulated corporate environments. The next phase of work will focus on increasing adoption in knowledge-intensive industries. The company positions individual digital twins as a new interface for enterprise AI, and its success will depend on the balance between automation, privacy, governance, and human agency.

Velaura AI has successfully raised $110 million in a Series A funding round, valuing the company at over one billion dollars. The round was led by Seligman Ventures, with participation from Capricorn Investment Group, Prosperity7 Ventures, and existing investors including Mayfield, Maverick Silicon, MARA, Premji Invest, and Samsung Catalyst Fund. StepStone Group also participated in the financing.

The capital raised will accelerate the development of the company's portfolio of AI computing power and advance its commercial activities. Velaura is targeting two growing areas: high energy-efficiency data center computing and Physical AI applications.

The company believes that energy consumption is becoming a serious constraint for artificial intelligence infrastructure. Although major cloud providers are actively investing in AI data centers, securing electricity and cooling capacity remains an increasingly complex challenge.

Velaura's core technology is the Titan Core silicon platform. This platform provides patented intellectual property and digital chip design capabilities. The company claims that Titan Core can deliver two to four times higher performance per watt.

The technology is oriented towards mathematical operations used in AI accelerators. It has already been implemented at a commercial scale, deployed in over 30 million ASICs utilizing leading semiconductor manufacturing processes. Velaura also demonstrates data on yield quality and production reliability.

The technology is planned to be applied in AI accelerators, and the architecture will be expanded for Physical AI applications. These applications include intelligent robots, drones, and autonomous systems that operate under strict power and heat dissipation constraints.

Rajiv Kemani, co-founder and CEO of Velaura, noted that further progress in AI will require improving the economic efficiency of computing. He added that the company aims to build a silicon and software foundation for this transition.

As AI infrastructure consumes more electricity, data center operators face the need to improve computing efficiency. Higher energy consumption also entails additional cooling requirements, which can increase the cost of deploying AI capabilities.

Velaura directs its solution at eliminating these limitations at the silicon level, focusing on increasing performance without a proportional increase in energy consumption. Furthermore, the company sees potential beyond traditional data centers, as Physical AI systems require efficient computation to operate in the real world, and robots cannot rely on the same resources as large data centers.

Velaura's leadership team includes executives and engineers from major technology companies such as Apple, NVIDIA, Google, Qualcomm, and Marvell. Their experience spans the development of low-power and high-performance semiconductor platforms and contributions to products shipped to billions of devices.

The new funding will be used to accelerate the development of Titan Core, as well as to expand engineering and customer teams. Additional resources will support deeper collaboration with strategic partners and customers in the fields of AI infrastructure and Physical AI development.

This funding reflects the growing investor interest in energy-efficient AI infrastructure, as power availability is increasingly viewed as the main barrier to AI expansion. Velaura's approach may allow for more computational power to be placed within existing energy constraints and reduce the thermal load on AI infrastructure.

The company positions its technology as suitable for both hyperscale and edge applications, giving it a presence across several segments of the growing AI computing market. Its headquarters in Silicon Valley place it within a dense semiconductor ecosystem.