Cognition Revenue, Valuation, Funding & Investors

Reflection AI is a Brooklyn, New York-based artificial intelligence company that develops open-weight frontier foundation models and autonomous software agents. It was founded in 2024 by former Google DeepMind researchers Misha Laskin and Ioannis Antonoglou, and its first product, Asimov, is a code-research agent that helps engineering teams understand and work across large codebases. The company has since repositioned around open frontier models and end-to-end AI systems sold to governments, large enterprises and public sector institutions that want full control over their AI stack. Reflection emerged from stealth in March 2025 with $130 million across a $25 million seed round led by Sequoia Capital and CRV and a $105 million Series A led by Lightspeed Venture Partners and CRV. It raised $2 billion at an $8 billion valuation in October 2025 in a round led by Nvidia, and in April 2026 closed a $2.5 billion Series C at a $25 billion pre-money valuation. In June 2026 it signed a compute agreement with SpaceX worth up to $6.3 billion, and it had about 320 employees in 2026.

Modal Labs is a New York-headquartered cloud infrastructure company that runs compute-heavy artificial intelligence workloads for developers. Its platform offers inference, sandboxes for coding agents, training and batch jobs on GPUs that start in seconds, with usage-based pricing and capacity pooled across many cloud providers. Erik Bernhardsson and Akshat Bubna founded the company in 2021, and the team works from New York, Stockholm and San Francisco. Customers use it for generative AI inference, model fine-tuning, computational biotech and media processing.

Lovable is a Stockholm-based software platform that uses artificial intelligence to generate full-stack web applications from natural language descriptions. Launched in 2024, it supports building apps with built-in authentication, databases, and deployment to platforms like Vercel. The tool targets developers and non-technical users in sectors such as technology, finance, and healthcare, enabling rapid prototyping of customer-facing products like dashboards and e-commerce sites.

OpenAI is an AI research and deployment company that develops large language models, including the GPT series, and consumer-facing products such as ChatGPT. Founded in 2015 as a nonprofit research lab, the organization transitioned to a capped-profit structure in 2019 to attract the capital needed for large-scale AI development. The company offers API access to its models for enterprise and developer use, alongside image generation through DALL-E and code-assistance tools. Headquartered in San Francisco, OpenAI has become one of the most prominent players in the generative AI space, backed by a multibillion-dollar investments from Microsoft, Amazon, NVIDIA, SoftBank and many more.

Anthropic is an AI safety company developing large language models and researching methods to make artificial intelligence more reliable and steerable. Anthropic was founded in 2021 by former OpenAI researchers including Dario and Daniela Amodei. The company's flagship product is Claude, a family of AI assistants deployed across consumer applications and enterprise APIs. Research priorities center on constitutional AI, mechanistic interpretability, and alignment techniques. Major investors include Google, Spark Capital, and Salesforce Ventures.

C3.ai Inc is an enterprise artificial intelligence company. The company provides software-as-a-service applications that enable customers to rapidly develop, deploy, and operate large-scale Enterprise AI applications across any infrastructure. It provides solutions under three divisions, namely, The C3 AI Platform, which is an end-to-end application development and runtime environment for designing, developing, and deploying AI applications; C3 AI Applications, which is a portfolio of pre-built, extensible, industry-specific, and application-specific Enterprise AI applications; and C3 Generative AI, which combines the utility of large language models. Geographically, the company derives revenue from North America, Europe, the Middle East and Africa, Asia Pacific, and the Rest of the World.

Project Prometheus is an artificial intelligence company developing AI systems for the physical economy, with applications across engineering and manufacturing in sectors such as aerospace, automotive, and computing. Founded in 2025 and headquartered in San Francisco with offices in London and Zurich, the company is co-led by Amazon founder Jeff Bezos and Vik Bajaj, a chemist and physicist formerly of Google X and co-founder of Verily. Unlike large language models that learn primarily from digital text, Prometheus focuses on AI that learns from real-world experimentation and physical trial and error, aiming to optimize design, production, and performance in industrial settings. The company launched with $6.2 billion in initial funding and subsequently closed a $10 billion round at a valuation of approximately $38 billion, with participation from investors including JPMorgan Chase and BlackRock. Its team has grown to over 120 employees, drawing talent from OpenAI, Google DeepMind, and Meta, and the company acquired agentic AI startup General Agents in 2025.

Cohere is a Canadian artificial intelligence company headquartered in Toronto that builds large language models and enterprise AI products for regulated industries. It was founded in 2019 by Aidan Gomez, Ivan Zhang and Nick Frosst. Its products include the Command family of generative models, Embed and Rerank retrieval models, and the North platform for business use, which can run in private clouds, on premises or in isolated virtual private clouds. Customers come mainly from finance, healthcare, manufacturing, energy and the public sector. The company has offices in Toronto, Montreal, New York, San Francisco, London, Paris and Seoul. It raised a US$500 million round at a US$6.8 billion valuation in August 2025 and extended it by US$100 million in September 2025. Cohere bought the Vancouver start-up Cognosys in May 2025 and the biopharma AI company Reliant AI in May 2026.

Moonshot AI is a Beijing-based foundation-model company best known for the Kimi family of long-context and multimodal artificial-intelligence assistants. The company develops proprietary large language models, consumer chat products, application-programming interfaces and tools for reasoning over long documents. Its offerings address consumers, software developers and enterprises that use generative AI for research, content, coding and knowledge work. Moonshot monetizes through model access, application subscriptions and usage-based API services while investing heavily in model training and inference infrastructure. Kimi became one of China’s most visible independent AI assistants as the company competed on context length, reasoning and multimodal capabilities. HKIC’s 2025 annual report places Moonshot among its selected AI portfolio companies and dates the investment on its official portfolio timeline.

Safe Superintelligence is a Palo Alto-based AI firm building systems with embedded safety protocols. Launched in 2024, the company develops risk assessment frameworks and mitigation tools to ensure secure deployment of advanced models in enterprise settings.

Z.ai is an advanced artificial intelligence platform known for its GLM model series, multimodal creation tools, and cost-efficient agentic coding environments like ZCode. It provides high-performance reasoning, real-time web search capabilities, and versatile utilities for building full-stack applications and data dashboards.

DeepSeek is a Chinese frontier AI lab that develops open-source large language models. DeepSeek conducts research in machine learning, natural language processing and related AI applications. DeepSeek has earned global recognition by challenging the hegemony of U.S. AI incumbents (like OpenAI and Google) through technological breakthroughs in efficiency and reasoning. DeepSeek-R1: A powerful reasoning model that rivals premium Western models. Its unique capability to mimic human thought and break down mathematical or programming tasks incrementally sent shockwaves through the tech industry.DeepSeek-V3 & V4: State-of-the-art general language models capable of running incredibly fast while consuming significantly less computational power than peer models.

Mistral AI is a Paris-headquartered developer of open-source large language models outperforming GPT-3.5 benchmarks. The company releases models like Mistral 7B, Mixtral 8x7B with 46.7 percent MT-Bench scores, and Pixtral 12B for vision-language tasks. Enterprises deploy these via La Plateforme API supporting 80 billion parameter models for coding, translation, and summarization. Mistral AI partners with Microsoft Azure and NVIDIA for inference hardware. Founded in 2023, it serves clients in automotive and finance across Europe and the U.S.

ElevenLabs is a London and Warsaw-headquartered AI company providing voice synthesis, cloning, dubbing, and generation tools. Its platform supports text-to-speech in over 29 languages, developer APIs for voice agents, and enterprise solutions for content localization. Features include instant voice cloning from short audio samples, emotional tone preservation, and audio editing capabilities used by creators and businesses worldwide.

Founded in 2012, Snowflake is a fully managed platform that consolidates data hosted on different public clouds for centralized analytics and governance. Snowflake’s cloud-native architecture allows users to independently scale the compute and storage layers, providing customers with optimized performance at lower costs. The company’s data lake and data warehouse products support a variety of use cases, including business analytics, data engineering, and artificial intelligence. Snowflake is widely used by Fortune 2000 companies in financial services, media, and retail sectors.

Perplexity AI is a San Francisco software company founded in 2022 that develops a web-connected answer engine. Its product combines large language models with real-time retrieval and presents responses with links to underlying sources. Consumers can use a free service or pay for advanced models, deeper research and higher usage limits. Enterprise customers buy search across public and private information, administrative controls and collaboration features. Perplexity also offers Sonar application-programming interfaces so developers can add grounded search and answer generation to other products. Its Comet browser extends the service into agent-assisted browsing and task execution. The company therefore earns subscription and usage-based revenue from consumer, enterprise and developer products rather than advertising-led ranked links.

FluidStack delivers a leading AI cloud platform with supercomputing infrastructure for AI labs, governments, and enterprises, serving clients like Mistral, Meta, and Black Forest Labs.

IonQ Inc sells access to several quantum computers of various qubit capacities and is in the process of researching and developing technologies for quantum computers with increasing computational capabilities. The company currently makes access to its quantum computers available via cloud platforms and also to select customers via its own cloud service. This cloud-based approach enables the broad availability of quantum-computing-as-a-service (QCaaS). The company derives its revenue from its quantum-computing-as-a-service arrangements, consulting services related to co-developing algorithms on company's quantum computing systems and contracts associated with the design, development, and construction of specialized quantum computing systems together with related services.