India is seeking proposals for the construction of a nuclear power plant core in the western state of Rajasthan, which will be one of the largest utilizing indigenous reactors. The state-owned company Nuclear Power Corp. of India Ltd. announced on social media that the contract for the so-called nuclear island package for a 2.8 gigawatt power station is valued at 28,000 crore rupees ($2.9 billion).
The tender was issued by the joint venture between the company and India's leading power producer, NTPC Ltd., named Anushakti Vidhyut Nigam Ltd. The Mahi Banswara project will be the first nuclear facility in the country not wholly owned by Nuclear Power Corp., which currently manages all of India's atomic power capacity amounting to 8.8 gigawatts.
Anushakti Vidhyut plans to construct four 700-megawatt reactors using locally developed compressed heavy water technology. Prime Minister Narendra Modi's government has opened this strategically sensitive sector by allowing private firms to build and operate nuclear power plants, with the ambitious goal of increasing the country's nuclear generation capacity elevenfold by 2047. Currently, this sector provides only 3% of India's electricity.
Businessman Gautam Adani has already announced plans to create 10 gigawatts of nuclear capacity, and Mukesh Ambani of Reliance Industries Ltd. signed a preliminary agreement to invest 2 trillion rupees in nuclear projects in the coastal state of Maharashtra. The tender covers the design and construction of the nuclear island—the part of the plant where the reactor and associated systems are located, in addition to other vital infrastructure. The deadline for submitting bids is September 30.
The total cost of the Anushakti project is estimated at 42,000 crore rupees. According to project partner NTPC, the first reactor will begin operation in 2031, and the remaining three units will be commissioned within the next five years. This project is also the first for NTPC in the nuclear energy sector, as the company aims to decarbonize its coal-based fleet. The New Delhi-based utility intends to build 30 gigawatts of nuclear capacity by 2047, which would account for 30% of India's target capacity by then.
Whale has successfully closed a Series C funding round, securing an additional $40 million. This brings the total funding in Series C to $100 million. The fundraising comes amid growing corporate interest in artificial intelligence platforms capable of integrating digital tools with real business processes.
The leaders of this round were CMB International and SMBC Asia Rising Fund. They were joined by Krungsri Finnovate, Singtel Innov8, Hyundai Motor Group, and Charisma Partners. Previously involved in the round were funds such as Bosch Ventures, MTR Lab, MDI Ventures, Gentree Fund, and Linear Capital.
The company plans to use the raised funds to strengthen international growth, enhance relationships with corporate clients, and deepen integration into local infrastructure. In the near term, Whale aims to increase its presence in North America and the Asia-Pacific regions.
Founded by Jerry Ye, Whale developed an AI Operating System designed to assist organizations in understanding and automating physical production environments. The company's proprietary Business World Model interprets data collected via cameras, sensors, and audio devices.
Investors are demonstrating confidence in AI tools that genuinely impact business operations, rather than being merely impressive autonomous software. Whale intends to direct capital toward recruiting new specialists, expanding its presence across various regions, and improving the reliability of its platform, especially given the demand for such solutions.
Whale currently collaborates with enterprises in sectors such as retail, automotive, manufacturing, finance, food processing, healthcare, and telecommunications. The company's primary objective is to help a greater number of organizations automate workflows, ensure regulatory compliance, and gain deep insights into their business operations.
CMB International emphasized Whale's capability to execute complex corporate AI projects across different industries and markets, believing that the company's technological foundation ensures long-term global growth. SMBC Asia Rising Fund noted Whale's ability to structure data from the physical environment into actionable business analytics. This partnership is expected to combine Whale's technologies with the international client network of SMBC Group to accelerate solution adoption across various sectors.
Krungsri Finnovate gained capabilities for comprehensive operational support in the ASEAN market from Whale's AI platform, while Hyundai Motor Group believes the technology can be applied in sales, marketing, customer interaction, and service within the Group's business.
Fireworks AI successfully raised $1.5 billion in a Series D funding round, boosting its valuation to $17.5 billion. This growth is driven by high demand for platforms that allow companies to build and run their own artificial intelligence models independently.
With the new capital, Fireworks plans to strengthen its computing power, hire more engineers, and solidify partnerships with leading cloud service providers. Furthermore, the company announced annual revenue exceeding $1 billion, representing a fivefold increase compared to the previous year.
The round was led by Atreides Management, Index Ventures, and TCV, with support from major players such as NVIDIA, Lightspeed, Evantic Capital, Bessemer, Menlo Ventures, Insight Partners, Ontario Teachers’ Pension Plan, and Lone Pine Capital. Fireworks was founded in 2022 by Lin Xiao and six former Meta engineers. The company helps organizations fine-tune open AI models, aiming to become a platform for creating custom intelligence that meets specific needs rather than relying entirely on closed models.
Fireworks AI forecasts that the next generation of enterprise artificial intelligence will focus on models developed for specific companies. Since every business has a unique set of data, workflows, and knowledge, these can become specialized AI skills. The company found that about 95% of processed tokens in its system come from user models, not standard base models. Fireworks offers a portfolio of over 200 text, image, and multimodal models.
The company's services are used by major clients such as Uber, Shopify, Doximity, Elastic, GitLab, and MongoDB. Fast-growing AI firms like Harvey and Cursor also rely on Fireworks' infrastructure. In addition to clients, the figures are impressive: earlier this year, Fireworks strengthened its collaboration with Microsoft, giving Microsoft customers access to AI models on the Fireworks platform running on the compute resources of over 20 different infrastructure providers. Gavin Baker from Atreides Management notes a trend where more and more businesses will begin combining advanced AI models with customizable open-source options tailored to their needs. The daily volume of tokens passing through the platform has increased from 15 trillion to over 40 trillion.
The new funds will be directed towards developing global computing infrastructure, increasing the engineering staff, and strengthening ties with cloud providers. The company currently employs 200 people and expects its staff to triple by the end of the year amid sustained demand. Fireworks asserts that its specialized models match or exceed closed models on a range of business-specific tasks at a significantly lower cost. Fireworks is betting on the growing interest of enterprises in gaining more control over AI performance, costs, and data security.
The social network X has introduced new guidelines to penalize accounts that use copied content from viral posts. These changes aim to redirect revenue generated by plagiarized publications to the original creators.
Nikita Bier, the network's product chief, announced the updates on the platform this Thursday, July 16. X will start employing an advanced version of Grok, the artificial intelligence developed by Elon Musk's xAI, to locate duplicated texts and videos with greater speed and accuracy.
The company claims that the new mechanism can identify stolen material three times faster and more accurately compared to the previous version. Bier clarified that the system will maintain copy recognition even if the user adds watermarks or other visual elements to the content.
Repeated use of copied content can result in the user's exclusion from the monetization program. Detection will also extend to phrases that are constantly repeated on the network, such as the cited example of 'Twitter is like the smoking area of the internet.' Recently, about 1.5 million stolen posts were identified, and X projects that the new rules will allow directing over US$1 million—equivalent to approximately R$5.5 million in direct conversion—that would previously have gone to infringing accounts.
Furthermore, X will strengthen the fight against so-called 'engagement baits,' which are posts created solely to artificially inflate a account's revenue through likes, replies, or shares. Profiles caught posting this type of message three times may be removed from the monetization program.
Bier mentioned examples of such promises, such as 'I will follow back everyone who replies to this post.' After the third infraction, the case may be forwarded to the X policy team for potential account suspension. He reported that on that day alone, about 4 thousand accounts were removed from the monetization program.
In a continuous effort to encourage the creation of original material, Bier had already signaled the intention to decrease the dissemination of recycled posts by presenting new video editing features on the platform. This is because many of the content that goes viral on X are, in fact, copies of existing material on other social networks, such as Instagram and Facebook, especially in the video segment.
In parallel, X has strengthened its blocking and bot identification mechanisms. Bier recalled that in April, the platform's security tools were able to detect and suspend, on average, 208 automated accounts per minute, as reported by TechCrunch.
