Foundation Models

Twitter

Facebook

Copy Link

Key Facts

Year2021

Coined ByStanford HAI

Key PlayersOpenAI, Google, Anthropic, DeepSeek

Top News

Xiaomi commits $8.7 billion to AI as MiMo-V2-Pro model drives stock surge and investor optimism

Xiaomi announced a massive $8.7 billion AI investment over the next three years, following the successful launch of its MiMo-V2-Pro large language model. The Chinese tech giant's stock jumped, making it the top performer on the Hang Seng Tech Index as CEO Lei Jun outlined ambitious plans to compete in the rapidly evolving AI agent market.

Technology2 sources

Xiaomi commits $8.7 billion to AI as MiMo-V2-Pro model drives stock surge and investor optimism

Technology2 sources

Roche and Nvidia deploy 3,500+ Blackwell GPUs to accelerate AI drug discovery and diagnostics

Swiss drugmaker Roche has deployed over 3,500 Nvidia Blackwell GPUs across the U.S. and Europe, creating the largest pharmaceutical AI infrastructure in the industry. The expansion aims to accelerate drug and diagnostics development while reducing timelines and costs through advanced AI computing capacity.

Technology3 sources

Roche and Nvidia deploy 3,500+ Blackwell GPUs to accelerate AI drug discovery and diagnostics

Technology3 sources

Firefly Neuroscience stock surges 47.7% on AI-powered ADHD subtype diagnosis breakthrough

Firefly Neuroscience shares soared 47.7% to $2.57 on Friday after the company announced research advances using AI-powered EEG technology to distinguish between ADHD subtypes. The breakthrough targets a U.S. treatment market estimated at over $10 billion and could shift diagnosis from symptom-based assessments to objective neural biomarkers.

Health2 sources

Firefly Neuroscience stock surges 47.7% on AI-powered ADHD subtype diagnosis breakthrough

Health2 sources

Canva Magic Layers Turns AI Images Into Editable Designs, Tackling a Major Creative Bottleneck

Canva has launched Magic Layers, an AI-powered feature that transforms flat AI-generated images into fully editable designs with separate layers. The tool addresses one of the biggest frustrations in AI content creation: the inability to make precise edits without regenerating entire images. Available now in public beta across the US, UK, Canada, and Australia, it works on any PNG or JPEG file.

Technology9 sources

Canva Magic Layers Turns AI Images Into Editable Designs, Tackling a Major Creative Bottleneck

Technology9 sources

Rhoda AI raises $450 million to train robots using internet videos at $1.7 billion valuation

Rhoda AI emerged from stealth with $450 million in Series A funding at a $1.7 billion valuation, unveiling a robot intelligence platform that learns from hundreds of millions of internet videos. The startup aims to solve robotics' longstanding challenge: enabling machines to handle unpredictable real-world conditions in manufacturing and logistics, not just controlled lab environments.

Technology5 sources

Rhoda AI raises $450 million to train robots using internet videos at $1.7 billion valuation

Technology5 sources

Microsoft releases Phi-4 multimodal model that knows when thinking wastes time

Microsoft unveiled Phi-4-reasoning-vision-15B, a compact 15-billion-parameter model trained on just 200 billion tokens—five times less than competing systems. The open-source AI model intelligently activates chain-of-thought reasoning for complex math and science problems while staying silent on simple tasks like image captioning, delivering competitive performance with a fraction of the compute cost.

Technology2 sources

Microsoft releases Phi-4 multimodal model that knows when thinking wastes time

Technology2 sources

Stanford's Merlin AI delivers consistent diagnoses across hospitals using 3D CT scans

Stanford researchers developed Merlin AI, a vision-language model that analyzes 3D abdominal CT scans with 81% accuracy in predicting diagnoses. Trained on over 15,000 scans and nearly one million diagnostic codes, the NIH-funded tool outperformed specialized models across 750 tasks and demonstrated consistent performance across multiple hospital sites, offering a potential solution to the growing radiologist shortage.

Science2 sources

Stanford's Merlin AI delivers consistent diagnoses across hospitals using 3D CT scans

Science2 sources

Alibaba's Qwen AI chief Junyang Lin steps down, triggering task force formation and concerns

Junyang Lin, the technical architect behind Alibaba's Qwen AI platform, abruptly stepped down just one day after the company unveiled its Qwen 3.5 open-weight small models. His departure, along with at least two other key team members, has rattled the developer community and prompted CEO Eddie Wu to form a special task force to accelerate foundation model development.

Business12 sources

Alibaba's Qwen AI chief Junyang Lin steps down, triggering task force formation and concerns

Business12 sources

Meta creates new AI unit in Reality Labs to accelerate model development

Meta has formed a new applied AI engineering organization within its Reality Labs division to speed up model development and support superintelligence initiatives. Led by VP Maher Saba, the unit will focus on building interfaces, generating real-world data, and providing evaluations to strengthen Meta's AI models. The move comes as Meta commits billions to AI infrastructure alongside a $50 million annual content licensing deal with News Corp.

Business2 sources

Meta creates new AI unit in Reality Labs to accelerate model development

Business2 sources

Rebel Startups Challenge SK and LG in South Korea's 'AI Squid Game' for National AI Dominance

South Korea launched a state-backed AI competition dubbed 'AI Squid Game' where rebel startups Motif Technologies and Upstage face off against industrial giants SK Group and LG Group. The Ministry of Science and ICT oversees this multi-stage contest to select two best homegrown AI models, marking a shift in government strategy from backing conglomerates to supporting agile tech entrepreneurs in the global AI race.

Technology2 sources

Rebel Startups Challenge SK and LG in South Korea's 'AI Squid Game' for National AI Dominance

Technology2 sources

Google Gemini 3.1 Pro doubles reasoning score, beats rivals in key AI benchmarks

Google released Gemini 3.1 Pro with dramatically improved problem-solving and reasoning capabilities, more than doubling its predecessor's performance on abstract reasoning tests. The newest and more powerful large language model scored 77.1% on ARC-AGI-2 and achieved a record 44.4% on Humanity's Last Exam, outperforming OpenAI and Anthropic in most benchmarks as accelerating competition among major tech companies intensifies.

Technology27 sources

Google Gemini 3.1 Pro doubles reasoning score, beats rivals in key AI benchmarks

Technology27 sources

Fei-Fei Li's World Labs raises $1 billion to advance spatial intelligence and world models

World Labs, founded by AI pioneer Fei-Fei Li, has closed a $1 billion funding round backed by Nvidia, AMD, and Autodesk. The startup is advancing spatial intelligence through generative AI world models that create interactive 3D virtual environments. Autodesk's $200 million investment signals a strategic partnership to integrate world models into 3D workflows for entertainment, robotics, and design.

Startups7 sources

Fei-Fei Li's World Labs raises $1 billion to advance spatial intelligence and world models

Startups7 sources

NPCI partners with Nvidia to build sovereign AI infrastructure for India's digital payments

NPCI has partnered with Nvidia to develop a payments-native AI foundation model aligned with India's regulatory and data sovereignty requirements. The collaboration will use Nvidia Nemotron to create a foundational AI layer for the payments ecosystem, building on the UPI Help Assistant pilot that supports grievance resolution at scale.

Technology2 sources

NPCI partners with Nvidia to build sovereign AI infrastructure for India's digital payments

Technology2 sources

Sarvam AI Launches Indus Chat App as Indian Startup Takes on ChatGPT and Gemini

Bengaluru-based Sarvam AI has launched its Indus chat app, powered by a 105-billion-parameter model built specifically for India's 22 official languages. The startup trained its AI models for Indian market with just 40 researchers and 4,000 GPUs over six months, positioning itself as a frugal alternative to global rivals like ChatGPT and Gemini while advancing India's sovereign AI ambitions.

Technology9 sources

Sarvam AI Launches Indus Chat App as Indian Startup Takes on ChatGPT and Gemini

Technology9 sources

Apple's AI strategy bets on restraint while Big Tech burns $700 billion on infrastructure

While Google, Meta, Microsoft, Amazon, and Alphabet collectively spend over $700 billion on AI infrastructure, Apple allocates just $12.7 billion annually. The spending gap reveals a fundamental disagreement about AI's future: will foundation models become commodities, or sources of lasting advantage? Apple's approach may signal strategic discipline or a costly miscalculation.

Business2 sources

Apple's AI strategy bets on restraint while Big Tech burns $700 billion on infrastructure

Business2 sources

Prev 1…567…12 Next