This Week in NLP #372
Keep up with what happened in NLP in the week ending Friday 2nd January 2026.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Meta has acquired Singapore-based AI startup Manus for over US$2B, with plans to sever its Chinese ties and integrate the company’s technology. [Gizmodo]
Nvidia is in advanced talks to acquire AI21 Labs for $2-3 billion, primarily targeting its talented workforce of around 200 AI experts. [CTech]
Nvidia is positioning itself as a crucial software company by developing extensive GPU-accelerating libraries and frameworks that can be used across multiple industries beyond AI. [The Register]
OpenAI is offering its employees an unprecedented average of US$1.5m in stock compensation, significantly higher than other tech companies before their public offerings. [Cryptopolitan]
TSMC plans to start mass producing advanced 2-nanometre chips by the end of the year, promising improved computing power and energy efficiency for various technologies. [Mint]
Now read on for everything else that happened in NLP this week.
Making News:
What makes this newsletter different?
This Week’s Topics:
If you’re reading this in a desktop web browser, you can access a navigation menu at the left-hand edge of this window.
The Generative AI Wars
Google Chrome may require a subscription for advanced AI browsing features that autonomously perform tasks on users’ behalf. [BGR]
Microsoft CEO Satya Nadella is restructuring leadership and accelerating AI development to compete with Google and OpenAI in the rapidly evolving tech landscape. [India Today]
OpenAI is exploring ways to integrate targeted advertisements into ChatGPT, potentially giving sponsored results preferential treatment while trying to maintain user trust. [Futurism]
AI Supremacy
China has launched three state-backed venture capital funds totalling US$21B to boost innovation in hard technology sectors amid global tensions and technological competition. [WebProNews]
Sovereign AI
India will invest Rs 10,300 crore to expand AI computing power to 38,000 GPUs, supporting startups and researchers through the IndiaAI Mission. [siliconindia]
Japan is set to quadruple its spending support for semiconductors and AI to ¥1.23 trillion in the upcoming fiscal year. [Bloomberg]
Supercomputer technology in Mexico is primarily concentrated in academic institutions, with 10-12 high-performance computing centers applied across various scientific disciplines. [teleSUR]
Türkiye is positioning itself as a strategic contributor to global AI development in 2025, focusing on digital autonomy, technological innovation, and inclusive technological leadership across defense, data, and diplomatic domains. [Daily Sabah]
The UAE has positioned itself as a global leader in AI adoption, launching initiatives like the AI Readiness Index and ‘AI for Development’ while embedding AI across government, infrastructure, and cultural preservation efforts. [Aletihad]
US tech giants have committed US$67.5B to expand data centre and AI infrastructure in India, targeting a rapidly growing digital market. [EnterpriseAI.com]
Hype Bubble?
A GeekWire survey of Seattle-area VCs suggests an AI market with signs of excess but not a catastrophic bubble, emphasizing real technological value. [GeekWire]
Microsoft expects 1.3 billion AI agents by 2028, but businesses may find these tools underwhelming and struggle to extract meaningful productivity gains. [ITPro]
A TechCrunch survey of enterprise VCs suggests 2026 will be the year enterprises start to meaningfully adopt and see value from AI technologies. [TechCrunch]
Big Iron
Goldman Sachs is co-leading financing for a 5-gigawatt Texas AI power project with Newmark Group in south Dallas. [Bloomberg]
Starcloud successfully trained an AI model in space and plans to deploy a massive constellation of satellites to provide orbital data center services. [GeekWire]
xAI has purchased a third building near Memphis to expand its AI infrastructure, aiming to boost compute power to nearly 2 gigawatts. [Yahoo Finance]
And tech companies are building onsite power generation facilities using natural gas to meet the massive and growing electricity demands of AI data centers, bypassing grid constraints. [WebProNews]
Hot Chips
ByteDance plans to invest US$14B in Nvidia AI chips in 2026, reflecting surging computing demands across its apps and services. [South China Morning Post]
China is imposing a new requirement for chipmakers to use at least 50% domestically produced equipment, aiming to reduce dependence on foreign semiconductor technology. [The Sri Lanka Guardian]
IDC warns of potential PC market downturn in 2026 due to AI-driven memory shortage, with possible price increases and reduced shipments. [Engadget]
Kioxia has seen its stock surge 540% in 2025, driven by surging demand for AI memory chips from tech giants like Apple and Microsoft. [Bloomberg]
Nvidia is seeking to increase H200 AI chip production at TSMC to meet strong Chinese tech company demand, despite uncertain regulatory approval for exports. [Reuters]
Nvidia faced a massive GPU smuggling operation in the US, with federal prosecutors uncovering a network attempting to illegally export US$160m worth of AI chips to China. [CNBC]
Samsung staff leaked advanced semiconductor manufacturing technology to ChangXin Memory Technologies, causing tens of trillions of won in estimated losses for South Korean companies. [Cryptopolitan]
Warm Bodies
Yoshua Bengio warns that AI is rapidly displacing jobs across industries, from cognitive to trade roles, with potentially devastating societal consequences within five years. [Breitbart]
A survey by TechCrunch reveals enterprise VCs expect AI to significantly impact workforce dynamics, potentially leading to job automation and layoffs by 2026. [TechCrunch]
Meta’s Mark Zuckerberg predicts AI will replace mid-level software engineers by 2025, signaling a potential profound shift in the tech industry. [ITPro]
Meanwhile, Big Tech companies hired over 32,000 employees in India in 2025, marking an 18% increase driven by demand for specialized AI, cloud, and cybersecurity talent. [News 24]
Consumer AI
AI wearables transformed from passive trackers to intelligent companions in 2025, offering personalized, context-aware assistance across devices like smart rings, glasses, and pendants. [Technowize]
Samsung is partnering with Perplexity AI to enhance Bixby’s intelligence, transforming it from a simple voice assistant to a more capable research tool on Galaxy devices. [Android Headlines]
It’s Only a Model
Alibaba’s Qwen-Image-2512 offers an open-source, commercially viable alternative to Google’s Gemini 3 Pro Image, providing enterprise-grade image generation with flexible deployment options. [VentureBeat]
Fal has released FLUX.2 [dev] Turbo, a lightning-fast, cost-efficient, open-weight image generation model that outperforms rivals while maintaining high-quality outputs. [VentureBeat]
Google Health has released MedASR, an open-weights medical speech-to-text model trained on 5,000 hours of clinical audio for specialized healthcare transcription tasks. [Marktechpost Media]
iMini AI integrated Kling 2.5 Turbo and Wan 2.6 to enhance its AI video creation platform with advanced multimodal capabilities. [EIN Presswire]
MiniMax released M2.1, an open-source AI model with enhanced multilingual programming capabilities and improved performance across web, mobile, and design tasks. [TechRepublic]
Other LLM Sightings
AvidTrak launched Avidia, an AI-powered data analyst that provides instant insights from call tracking and marketing data through natural language queries. [EIN Presswire]
CrafterCMS has released an open-source MCP Client Plugin that enables AI-powered, context-aware digital experiences through standardized LLM integration. [EIN Presswire]
Darkhorse Insurance Brokers integrated Rhea, an AI agent that helps recover abandoned insurance quotes through contextual messaging channels. [EIN Presswire]
Homesage.ai launched an AI-powered platform that helps hard money lenders process property deals faster and more accurately through automated analytics tools. [EIN Presswire]
LovedByAI launched a WordPress plugin that helps small businesses optimize their websites for AI search visibility at an affordable price. [EIN Presswire]
Renaissant has launched a voice-based AI clerk to improve logistics efficiency and address security concerns in shipment coordination. [WisBusiness]
TokTak launched a global AI content platform at toktak.ai, enabling users to generate and publish social content across channels with a single product link. [EIN Presswire]
Vanguard is leveraging AI to create a hyper-personalized digital advisor that can serve millions of clients efficiently. [Fortune]
Zipline launched AI capabilities for retail operations, enabling stores to replicate top-performing team strategies and improve decision-making across their organization. [EIN Presswire]
Risks and Responses
Bernie Sanders criticized AI’s potential economic and social impacts, calling for a moratorium on datacenters and questioning tech moguls’ motives in developing the technology. [The Guardian]
Geoff Hinton criticized Trump’s hands-off approach to AI regulation, warning of potential risks including the possibility of AI taking over the world. [The Daily Beast]
LangChain discovered a critical AI vulnerability that exposes millions of applications to potential theft and code injection risks. [eWeek]
Microsoft researchers found that AI tools like Copilot may diminish critical thinking skills and lead to long-term reliance on generative AI in workplace tasks. [ITPro]
OpenAI is hiring a $555,000-per-year ‘head of preparedness’ to manage potential risks and mitigate harmful capabilities in its rapidly expanding AI products. [Gizmodo]
OrgLogic launched a governance-first approach to responsible AI use in HR, emphasizing executive oversight, ethical judgment, and structured people systems. [PRWeb]
A survey by Pew Research reveals growing American distrust in tech companies due to privacy concerns, political entanglements, and fears about data misuse. [WebProNews]
Regulation
Britain is positioning itself as a global leader in AI innovation through a flexible, principles-based regulatory approach that encourages experimentation and learning. [UKTN]
California has enacted a new law requiring AI companies to disclose catastrophic risk plans and provide whistleblower protections, aiming to enhance transparency and accountability in the rapidly evolving AI industry. [ZDNet]
California’s attempt to regulate data center energy usage in 2025 resulted in a toothless law requiring a report by 2027, after facing opposition from Big Tech and business groups. [The Mercury News]
China is enforcing strict AI regulations to prevent chatbots from potentially challenging Communist Party control. [The Wall Street Journal]
David Sacks, Trump’s AI czar, has worried tech lobbyists by pushing an aggressive executive order that may derail efforts to create a federal AI regulation framework. [Politico]
Korea’s new AI Basic Act leaves tech industry uncertain and unprepared, with vague guidelines and potential compliance challenges threatening innovation and service launches. [Korea JoongAng Daily]
Conversational AI
Aarav Solutions launched an AI-powered CPQ Chatbot on Odoo, enabling intelligent digital sales with enterprise-level pricing control and customer-friendly ordering. [EIN Presswire]
Crafter Software pre-announced CrafterQ, an AI chatbot platform designed to help businesses deploy custom, secure, and brand-aligned digital agents across various channels. [EIN Presswire]
Be Real
China’s draft rules prohibit AI companions from simulating relatives or replacing social interactions for elderly users. [The Register]
Joi AI, a Cyprus-based company, offers sexualized chatbots based on various characters, highlighting a potential lasting byproduct of the generative AI surge despite broader industry uncertainties. [Wired]
Voice News
AI-powered music and voice technologies are transforming audio creation, enabling faster, cheaper production while raising ethical concerns about ownership and authenticity. [News9 LIVE]
Audfly launched FocusAura, a dual-directional voice interaction solution creating private audio zones for AI avatars, kiosks, and robotics at CES 2026. [EIN Presswire]
OpenAI is preparing to launch an audio-first personal device with advanced conversational AI capabilities, reflecting the tech industry’s shift towards audio interfaces. [TechCrunch]
SaraMusic AI has launched an AI-powered 24/7 vocal coaching platform, offering affordable, personalized voice training for aspiring singers and speakers. [EIN Presswire]
Translation
MachineTranslation.com pioneered consensus-based AI translation, reducing errors by 22% by leveraging multiple independent AI engines simultaneously. [Technology.Org]
AI in Journalism
Al Jazeera’s new AI-integrated newsroom, built with Google Cloud, represents a sophisticated system for shaping journalistic context and narrative through technological infrastructure. [Ynet]
Health Tech
AI therapists are emerging as a potential solution to the global mental-health crisis, offering accessible support but raising significant ethical, privacy, and effectiveness concerns. [MIT Technology Review]
AIdMD Technologies has launched an AI-powered healthcare platform designed to reduce administrative burden while maintaining physician oversight and workflow efficiency. [EIN Presswire]
PureHealth launched ‘Nada’, an AI-powered digital assistant that helps doctors document medical notes during patient consultations while prioritizing patient engagement. [Business Wire]
Legal Tech
Vocodia is launching an AI-powered platform to support law firms with marketing, client intake, and lead generation in the mass tort legal market. [GlobeNewswire]
Ed Tech
Squirrel Ai set a Guinness World Record by demonstrating its AI learning system’s superior performance compared to traditional teaching methods across multiple student metrics. [EIN Presswire]
The Emergent Learning Research Institute is partnering with Lincoln Parish Schools to launch Project Read.AI, an innovative AI-powered initiative to enhance PreK-3 reading skills. [Magnoliareporter.com]
Funding
Ambassador raised US$7m to expand its AI-powered customer feedback software platform, launching HiroAI to help companies analyze and improve customer engagement. [GeekWire]
BriefCatch raised US$6m in Series A funding from Full In to expand its AI-powered legal writing technology platform. [LawSites]
MiniMax Group led six Hong Kong listings worth HKUS$16.7B on Wednesday, highlighting a strong year for AI and chipmaker IPOs in the city. [Reuters]
Shanghai Biren Technology raised HKUS$5.58B in its Hong Kong IPO, with strong institutional and retail investor demand for the AI chip startup. [Reuters]
SoftBank completed a US$22.5B investment in OpenAI, bringing its total ownership to approximately 11% and fulfilling its initial commitment. [SoftBank]
Traini raised US$7.5m to develop an AI-powered smart collar that translates pet emotions and behaviours through advanced multimodal generative technology. [Business Wire]
Zhipu AI launched an initial public offering in Hong Kong to raise HKUS$4.35B, positioning itself as the first Chinese AI LLM developer to list. [Yahoo Finance]
There’s More
India will host the AI Impact Summit 2026, bringing together global leaders and tech executives to build consensus on AI’s inclusive and democratic development. [BusinessToday.in]
Walmart’s CEO Doug McMillon predicts AI will transform every job, emphasizing the need for proactive workforce reskilling and adaptation. [Inc.]
The Back Cover
Here’s Clemens Wenners and Sofia Dragt with a cover of Peter Gabriel’s Don’t Give Up.
If you know of a great cover version of a back-catalogue classic, drop an email to us at backcover@language-technology.com and we’ll consider it for inclusion here.
Got this from someone else?
Did you find this newsletter useful? if so, you might forward it to a friend; or you could email us at news@language-technology.com to tell us what you want more of.
Did you hate this newsletter? if so, you could forward it to an enemy; or you could email us at news@language-technology.com to tell us why – make it better!