AI companies, including OpenAI, are shifting away from the ‘bigger is better’ philosophy for training models. Instead, they are developing techniques that allow algorithms to ‘think’ in more human-like ways. These methods aim to address challenges such as massive energy consumption, hardware failures, and data scarcity that have hindered advancements in large language models.
OpenAI’s new model, o1, uses a technique called ‘test-time compute’, allowing it to consider multiple answers and choose the best option during use. This approach improves performance in complex tasks, like problem-solving and decision-making, without needing extensive pre-training. Noam Brown, an OpenAI researcher, revealed that even brief ‘thinking’ boosts the model’s capabilities significantly.
The industry-wide shift has broader implications for AI hardware, especially as Nvidia’s chips have been critical to AI training. Experts predict a move towards distributed cloud-based servers for inference tasks, potentially reshaping the demand landscape for chips. Prominent investors, such as Sequoia and Andreessen Horowitz, are monitoring these changes closely as they may impact investments in AI infrastructure.
German optical tech firm Carl Zeiss AG has inaugurated its first global capability centre (GCC) in Bengaluru, India, and plans to double its local workforce to 5,000 over the next three years. The new centre will focus on cloud computing, cybersecurity, and network operations, alongside software development for Carl Zeiss’s medical tech division. This move highlights India’s transformation from an outsourcing destination to a strategic base supporting global operations.
Beyond the GCC, Zeiss is expanding its presence in Bengaluru with a new manufacturing plant slated to open in 2025. This facility, the company’s largest investment outside Germany, will be its fifth in India, contributing to its workforce growth. The India unit, also involved in R&D and sales, is projected to reach a revenue of 22 billion rupees for the year ending September 2025—a 19% increase.
India’s GCC sector is booming, with Karnataka’s government aiming to double GCCs in the state by 2029. Industry reports expect the Indian GCC market to reach up to $105 billion by 2030, reflecting the country’s increasing role in global business support.
Vietnam’s semiconductor industry is gaining momentum as foreign companies invest in chip testing and packaging facilities, shifting some production away from China. Amid trade tensions between the US and China, several global players, including South Korea’s Hana Micron and US-based Amkor Technology, are expanding operations in Vietnam to diversify their production bases. Hana Micron has committed over $930 million to boost its packaging capacity, while Amkor is investing $1.6 billion to establish its largest packaging plant, transferring some machinery from its Chinese facilities.
The rise in investment is set to increase Vietnam’s global share in chip assembling, testing, and packaging, with estimates suggesting a rise from 1% in 2022 to around 8-9% by 2032. Domestic companies are also stepping up. Vietnamese tech firm FPT plans to start a testing facility near Hanoi next year, investing up to $30 million, while Sovico Group is seeking partnerships for a chip plant in Danang.
Vietnam’s strategic push into the semiconductor sector has been encouraged by the US, viewing the country as a potential alternative to China for supply chains. The Biden administration’s support, especially as trade tensions grow, has bolstered Vietnam’s role in this industry. With domestic and foreign investments combined, Vietnam is poised to strengthen its position as a key player in the global semiconductor back-end market.
Looking forward, Vietnam is ambitiously aiming to develop its front-end chipmaking capabilities, planning to have its first foundry operational by 2030. Viettel, a state-owned firm, is set to lead this initiative, indicating Vietnam’s broader goal of advancing its semiconductor industry and reducing reliance on foreign production bases.
EU has committed €133 million ($142 million) to support pilot production facilities for photonic semiconductors in the Netherlands, according to the Dutch economy ministry. This initiative forms part of a larger €380 million fund under the EU’s Chips Joint Undertaking, a public-private partnership designed to bolster Europe’s semiconductor industry. Photonic semiconductors, which use light instead of electrons for calculations, promise enhanced speed and energy efficiency and are increasingly essential in fields like data centres and automotive technology.
Dutch economy minister Dirk Beljaarts emphasised photonics as a “technology of strategic importance” for Europe’s economic competitiveness. By building strong domestic capabilities in research, innovation, and supply chains, the EU aims to reduce dependence on global tech rivals. The move follows calls from European industry leaders for significant EU investment to keep pace with advancements in Asia and the US.
The Dutch pilot facilities, slated to begin construction in 2025, will involve Eindhoven and Twente universities alongside the TNO research institute, with co-investment from companies utilising the new infrastructure.
Taiwan Semiconductor Manufacturing Co. (TSMC) confirmed that its investment plans in the United States will continue unchanged, following the election of Donald Trump as the next US president. TSMC, a leading global chipmaker and supplier to tech giants like Apple and Nvidia, is investing $65 billion in new semiconductor factories in Arizona.
Despite Trump’s previous comments accusing Taiwan of harming the US semiconductor industry, TSMC has recently secured a $6.6 billion subsidy from the US Commerce Department to support advanced chip production in Phoenix. TSMC’s US unit, along with other firms like GlobalFoundries, is expected to receive additional support under the Biden administration’s Chips and Science Act.
TSMC shares have remained resilient, bolstered by strong demand for AI technology, with its American Depositary Receipts rising 4.1% on Thursday as Nvidia’s stock surged, helping drive investor confidence.
Cloud monitoring firm Datadog raised its annual revenue and profit forecasts on Thursday, driven by increasing demand for its AI-backed cybersecurity products. The New York-based company now expects full-year revenue of about $2.66 billion, up from its previous projection of $2.62 to $2.63 billion, with analysts having anticipated $2.63 billion. Datadog also raised its adjusted profit forecast to between $1.75 and $1.77 per share, surpassing earlier estimates of $1.62 to $1.66.
The company’s performance has been bolstered by the growing adoption of AI applications by its customers, who are increasingly deploying these tools in live production environments. As AI apps run in the cloud, Datadog stands to benefit from the ongoing migration to cloud services, which drives demand for its monitoring software. For the quarter ending September 30, Datadog reported revenue of $690 million, beating the expected $664.3 million, and posted an adjusted profit of 46 cents per share, exceeding analysts’ predictions of 40 cents.
Despite the strong results and optimistic growth outlook, Datadog’s stock saw some volatility, rising 4.1% before later paring its gains as investors reacted to high expectations for the company’s performance.
French AI startup LightOn launched an initial public offering (IPO) on the Euronext Growth market in Paris, with its debut trading expected later this month. The company, known for its large language model (LLM) software used by businesses and the French government, will be Europe’s first publicly listed generative AI startup, a significant milestone as France aims to position itself as a leader in AI within Europe.
LightOn’s co-CEOs Igor Carron and Laurent Daudet emphasised that the IPO provides a ‘unique opportunity’ for investors to support a growing French tech company with a track record of success both in France and internationally. Shares are priced at 10.35 euros, valuing the company at around 50 million euros, and LightOn aims to raise roughly 10.4 million euros through the capital increase. The subscription period will run until November 20, with shares expected to trade beginning 26 November.
This move aligns with France’s broader push to close the innovation gap with the US and the UK, with ambitions for 100 tech ‘unicorns’ by 2030. LightOn’s listing could signal an opening for more European AI firms to seek public funding, offering investors access to an evolving tech market in the region.
Amazon is reportedly in advanced talks for a second multi-billion dollar investment in the AI startup Anthropic, building on its previous $4 billion commitment made in 2023. This new investment would not only bolster Amazon’s growing ties with Anthropic but also help enhance its strategic position in the highly competitive AI sector. Anthropic, which is using Amazon Web Services (AWS) to power its AI model training, has become a key player in the AI race.
In addition to providing financial backing, Amazon has reportedly asked Anthropic to utilise its servers, which are powered by Amazon’s custom-designed chips. However, sources note that Anthropic has a preference for using Nvidia-designed chips, which are widely recognised as the industry standard for AI processing. This dynamic highlights the ongoing competition between Amazon and Nvidia in the AI hardware space, as both tech giants vie for dominance in the rapidly expanding market.
Anthropic, founded by former OpenAI executives Dario and Daniela Amodei, has attracted significant interest from other major players in the tech industry. The startup secured a $500 million investment from Google’s parent company, Alphabet, last year, with Alphabet pledging an additional $1.5 billion over time. Despite these investments, both Amazon and Anthropic have declined to comment on the specifics of the latest talks regarding the new investment, underscoring the confidential nature of these high-stakes negotiations.
GlobalWafers has expressed optimism that the US Chips and Science Act will continue to provide strong support for chip manufacturers under the new administration. This landmark act, aimed at boosting domestic semiconductor production, offers financial incentives to encourage companies to invest in US facilities—a vital step toward securing supply chains and reducing reliance on foreign manufacturing.
In a recent statement, GlobalWafers noted that programs of this scale and duration are typically supported across different US administrations, given their importance to economic and national security. The company sees the CHIPS Act as essential for driving investments in semiconductor production and also for advancing technological innovation within the industry. They anticipate that the act’s stability under a Trump administration will allow businesses to plan long-term investments in US operations without interruption.
By fostering consistent investment in chip manufacturing, GlobalWafers believes the CHIPS Act will help ensure a robust, self-reliant US semiconductor ecosystem. The program’s continuation is seen as crucial for sustaining growth in the industry, creating jobs, and advancing the global competitiveness of the US in semiconductor technology.
Apple has launched the AI-powered version of iOS 18.2, bringing new features to public beta users. These include tools like Genmoji, an AI-powered emoji generator, Image Playground for creating AI-generated images and ChatGPT integration with Siri. The update also introduces Visual Intelligence, which uses the iPhone 16’s camera for real-time object and place identification. These features were previously available only to developers but are now open to select users, with some requiring a waitlist for access.
The new capabilities, grouped under “Apple Intelligence,” promise to enhance Siri’s functionality, including offering writing and proofreading assistance across apps. Apple plans to extend these AI tools to third-party developers, potentially increasing their impact. For now, users can experiment with ChatGPT for text generation and image creation, as well as generate custom emojis and edit rough sketches using Image Wand.
While Apple Intelligence is enabled by default on some devices, others will need to sign up for the new features. However, there’s a waitlist for certain tools, which Apple is managing to ensure a safe rollout. Concerns over misuse, like the creation of NSFW emojis, have contributed to the cautious approach.
The update follows AI improvements introduced in iOS 18.1, such as enhanced writing tools and updated Siri features. Alongside iOS 18.2, Apple also released public betas for iPadOS 18.2, macOS Sequoia 15.2, and tvOS 18.2, continuing to expand its AI-driven ecosystem.
