US-based AI cloud-computing startup CoreWeave has announced the acquisition of Core Scientific, a leading data centre infrastructure provider, for $9 million in an all-stock transaction.

“This acquisition accelerates our strategy to deploy AI and HPC workloads at scale,” said Michael Intrator, CoreWeave’s CEO and co-founder. 

“Verticalising the ownership of Core Scientific’s high-performance data centre infrastructure enables CoreWeave to significantly enhance operating efficiency and de-risk our future expansion, solidifying our growth trajectory,” he added. 

As a result of this acquisition, CoreWeave will gain ownership of approximately 1.3 gigawatts (GW) of gross power from Core Scientific’s national data centre footprint, with an additional 1 GW or more available for future expansion.

Additionally, it offers the potential for repurposing resources for high-performance computing (HPC) or divesting from the crypto mining sector, supported by approximately 840 gross MW for HPC contracts and 500 gross MW for crypto mining.

According to Reuters, bitcoin miners have leveraged the AI surge to expand their operations beyond cryptocurrency by renting out power and data centre space to accommodate the increasing demand for AI tasks. 

Bernstein analyst Gautam Chhugani told Reuters that the acquisition is anticipated to establish a standard for bitcoin miners seeking to transition to AI, noting that power is the primary limitation for the growth of AI data centres. 

The company anticipates closing the transaction in the fourth quarter of 2025, subject to regulatory approvals and approval by Core Scientific’s stockholders. 

Following CoreWeave’s successful initial public offering (IPO) in March 2025, this acquisition is expected to help CoreWeave streamline its data centre operations, ensuring future revenue growth and improving profitability. 

This deal is expected to yield approximately $500 million in annual cost savings by the end of 2027, primarily through streamlined operations.

French chip startup Arago has secured $26 million in seed capital to create its light-based AI accelerator, known as Jef. 

The company’s goal is to lower AI inference energy consumption by a factor of 10 compared to current GPUs by integrating analogue and digital electronics, silicon photonics, and free-space optics, as reported by the EE Times.

Unlike most photonic computing startups, which experts in optics or photonics typically establish, Arago was founded by individuals specialising in physics and AI, a fact that Arago CEO Nicolas Muller believes provides the team with a unique viewpoint.

Arago has developed a photonic AI accelerator that is designed for high energy efficiency and throughput. By utilising light as the basis for computation, the chip optimises performance per watt and dollar, enabling AI workloads to run with considerably reduced energy and cost requirements.

Based on Arago’s exclusive multi-physics computing core, high-volume AI tasks characterised by 𝑛^3 complexity can be accelerated and performed with extremely low power consumption.

Arago refers to its inference chip, Jef, as a “multiphysics processor” since it integrates features of both analogue and digital electronics alongside silicon photonics and free-space optics. 

Muller believes that Arago faces significant challenges, including the development of a photonic computing chip that operates reliably and is sufficiently high-performing and user-friendly to be integrated into AI technology ecosystems.

“We started from the top, meaning: what does it take to infer a model very efficiently?” Muller told EE Times. “Then we turned that into a data sheet for a product.”

“Historically, companies that were tackling this space were born around 5 to 10 years ago, where the technical and scientific landscape was extremely different,” Muller said. “You had pure optical technologies and different technologies that are based on free space optics. What we built is a technology that overcomes all of those challenges by only using optics at very specific places.”

The company states on its website that contemporary AI models are increasingly limited by the compute and memory capacities of traditional silicon-based accelerators. Improving AI efficiency and deployment necessitates a shift away from conventional transistor technology.

OpenAI, the AI startup behind ChatGPT, has ‘overhauled’ its security operations to protect intellectual property from corporate espionage, the Financial Times reported on July 8. 

The development took place after the company claimed that Chinese AI startup DeepSeek had copied its models using distillation techniques. The report noted that OpenAI is implementing tighter restrictions on sensitive data and strengthening staff vetting processes. 

The company’s policies, referred to as information ‘tenting’, have limited the number of personnel who can access the new algorithms being developed at OpenAI, according to insiders quoted by FT. Besides, employees are only permitted to enter certain rooms by scanning their fingerprints. 

OpenAI safeguards its model weights by implementing a ‘deny-by-default egress policy, ‘ which means that no connections to the internet are permitted unless they are explicitly authorised. In addition, the company has also increased physical security at its data centres. 

Earlier this year, Microsoft security researchers believed that individuals potentially connected to DeepSeek are ‘exfiltrating a significant amount of data’ through OpenAI’s API, as per a report by Bloomberg. 

Furthermore, OpenAI also told FT that it had seen “some evidence of distillation”, which is a technique to improve the performance of an AI model by using outputs from another one.

In April, Business Insider reported that OpenAI required developers seeking access to the company’s advanced AI models to verify their identity with a government ID. 

China’s DeepSeek, a subsidiary of HighFlyer, launched the R1 reasoning model a few months ago. It generated significant industry buzz by being open source and providing capabilities comparable to OpenAI’s o1 reasoning model, but at a fraction of the training cost. Since then, there has been a lot of buzz around the threat that models from China potentially pose.