SINGAPORE, April 17 – Taiwan Semiconductor Manufacturing Co (TSMC) lifted investor sentiment across Asian and European chip stocks on Thursday by issuing a positive forecast, helping ease nerves after bleak warnings from Nvidia and ASML had shaken the market. TSMC, a key supplier for Apple and Nvidia, reported better-than-expected quarterly earnings and reaffirmed its full-year targets for both revenue and capital expenditure. The upbeat outlook came after markets closed in Taiwan, pushing up share prices of tech companies in Japan and Europe. TSMC’s Frankfurt-listed shares rose 5.5% in early trading.

A day earlier, chip stocks took a hit after Nvidia warned of a $5.5 billion loss due to new U.S. export restrictions on its AI chips for China, and ASML expressed concerns over its business outlook. Despite potential risks from U.S. tariffs, TSMC executives said they had not observed any shift in customer behavior and expected strong demand for artificial intelligence to continue driving business growth. As the world’s top contract chipmaker, TSMC plays a critical role in producing advanced AI chips.

Gary Tan of Allspring Global Investments said TSMC was well-positioned to pass on price increases through the supply chain. Although his fund holds TSMC shares, it remains underweight on Taiwan. While TSMC’s outlook offers some relief for chip investors, uncertainties in the sector persist as earnings season unfolds. Mark Hackett of Nationwide noted that company executives might use tariff-related uncertainty as a reason to withhold future guidance.

Concerns over tariffs have already impacted TSMC’s Taipei-listed shares, which have fallen over 20% in 2025 — the worst start to a year in three decades. Its U.S.-listed shares are down 23%. Goldman Sachs reported that foreign investors have sold $8.66 billion worth of TSMC shares this year, following net purchases of $2 billion in 2024 and $10.4 billion in 2023.

Taiwan-related Risks: The broader selloff in TSMC and other Taiwanese stocks highlights investor unease as they respond to unpredictable U.S. trade policy under President Trump. Taiwan was previously set to face a 32% tariff, which was paused to allow for bilateral talks. The first direct discussions between Taiwan and the U.S. on tariffs took place last Friday.

TSMC, which recently committed an additional $100 billion to expand its operations in the U.S., is pivotal to the American chip ecosystem. By relocating more production to U.S. soil, the company could help reduce supply chain vulnerabilities for clients such as Qualcomm and AMD. Sam Konrad of Jupiter Asset Management noted that geopolitical instability surrounding Taiwan poses significant risks to U.S. tech companies, yet those risks appear to be more reflected in Taiwanese stock valuations than in their American counterparts.

Disclaimer: This image is taken from Reuters.

Meta CEO Mark Zuckerberg defended the company's acquisition strategy in court, testifying as the first witness in a major antitrust trial that could reshape the future of the $1.4 trillion company. The U.S. Federal Trade Commission (FTC) has accused Meta of creating an illegal monopoly by acquiring rivals Instagram and WhatsApp to eliminate competition.

Zuckerberg argued that the purchases of Instagram and WhatsApp were aimed at enhancing innovation and improving user experience. He acknowledged that the platform's focus has shifted more toward content discovery beyond personal connections, but emphasized that relationships with friends and family remain central to Meta's mission. If the case succeeds, Meta could be forced to sell Instagram and WhatsApp, marking one of the most significant antitrust actions against a tech company in decades.

The FTC’s lead attorney, Daniel Matheson, argued in his opening statement that Meta's dominance through Facebook, Instagram, and WhatsApp has left consumers with limited choices. He claimed that the acquisitions were intended to prevent competition, effectively dominating the “personal social networking” market. Matheson stated that Meta’s strategy has been to neutralize competitors rather than compete with them, leaving users with no reasonable alternatives.

Meta's legal team pushed back, arguing that the social networking market is broader than the FTC suggests and that the company competes with a variety of platforms like TikTok, Snapchat, YouTube, and Apple’s iMessage. They also pointed out that both Instagram and WhatsApp were approved by regulators when acquired. They criticized the FTC’s case as relying heavily on old emails from a decade ago.

One crucial piece of evidence for the FTC is a 2012 email in which Zuckerberg discussed buying Instagram to "neutralize a competitor." Zuckerberg acknowledged these emails but insisted they were part of broader discussions and misinterpreted. He claimed that Meta made significant investments in Instagram and WhatsApp after their acquisitions, supporting their growth. The FTC also argued that Meta intentionally slowed Instagram's development after the acquisition to protect Facebook's market position. Matheson labeled this a "rational business decision" from Meta's perspective, but it contradicted antitrust law’s goal of preserving competition.

The trial continues with Zuckerberg scheduled to return to the stand. Additional testimony from figures like former COO Sheryl Sandberg and Instagram co-founder Kevin Systrom is expected in the coming weeks. If the court rules in favor of the FTC, Meta could be forced to undo its acquisitions of Instagram and WhatsApp, disrupting the company’s operations and possibly affecting its advertising revenue, with Instagram alone projected to contribute over half of Meta’s U.S. ad revenue by 2025. The trial, overseen by U.S. District Judge James Boasberg, is expected to last for eight weeks, with several high-profile figures from Meta and other tech companies scheduled to testify.
Disclaimer: This image is taken from Bloomberg.

OpenAI is planning to implement a mandatory ID verification process for organizations seeking access to upcoming AI models, according to a support page on its website. This new system, called "Verified Organisation," will allow developers to unlock access to the most advanced models and features on OpenAI’s platform. To verify an organization, a government-issued ID from one of the countries supported by OpenAI’s API is required. Each ID can only be used for verifying one organization every 90 days, and not all organizations will be eligible for verification.

OpenAI explained that the verification process is being introduced to reduce unsafe usage of AI while still making advanced models available to the larger developer community. The company also mentioned that the move could help prevent intellectual property theft. A Bloomberg report earlier this year revealed that OpenAI was investigating whether a group connected to the China-based AI lab DeepSeek had unlawfully accessed large amounts of data through OpenAI’s API, possibly violating the company's terms by using it to train models.

The verification process is essential for gaining access to advanced models and additional features on the OpenAI platform, enabling developers to use the latest AI technologies. If verification isn’t available, developers can still use existing models on the platform. Furthermore, models that require verification may become accessible to all users in the future, even without verification. OpenAI supports IDs from over 200 countries, and users can check if their country is supported by starting the verification process.

Disclaimer: This image is taken from Bloomberg.

OpenAI has launched a significant upgrade to ChatGPT’s memory feature, allowing the chatbot to remember previous interactions and deliver more personalized responses. This update is part of OpenAI’s broader goal to enhance the AI’s ability to adapt to individual users over time. Announcing the update on X, CEO Sam Altman noted that the memory feature is optional and can be turned on by users. “We’ve made big improvements to memory in ChatGPT – it can now recall all your past conversations! It’s a surprisingly great feature and shows the potential of AI systems that evolve with you to become more helpful and personalized,” Altman said.

At present, the feature is only available to ChatGPT Pro users, though it will soon roll out to 'Plus' subscribers as well. However, the upgraded memory is not yet accessible in some regions, including the European Economic Area (EEA), the UK, Iceland, Liechtenstein, Switzerland, and Norway. To address privacy concerns, OpenAI stressed that users retain full control over memory settings. They can disable the feature at any time or use a temporary chat mode that doesn’t save any conversations. “For chats without memory usage, simply use the temporary chat option,” the company advised.

OpenAI added that the memory now not only stores previously saved details but also uses past chats to provide more relevant and helpful replies. Future interactions build on this context, making the experience more seamless and customized. This update underscores OpenAI’s continued efforts to advance personalization and contextual awareness in AI—an area also being explored by companies like Google, Amazon, and Meta.
Disclaimer: This image is taken from Reuters.