Cognition Valued at $48B After Revenue Doubles to $900M

Key Takeaways

• Cognition is valued at $48 billion following a major funding round led by Andreessen Horowitz and Accel.
• Cognition's annual revenue has doubled to over $900 million in just three months.
• Enterprise customers like Citi and Mercedes-Benz use Cognition's Devin agent for complex coding tasks.
• Cognition released the SWE-2 model, which costs 81 percent less than the previous version while matching performance.
• CoreWeave launched Physical AI Field Engineering to help clients deploy AI models using their own data.
• Renesas opened a new Physical AI and Robotics Lab in Beijing and added 64-bit MPUs to its RZ/G Series.
• NVIDIA's NIM optimizations increased Nemotron 3 Ultra throughput by 2.5x to 1,997 tokens per second.
• Vulcan Infrastructure and Power raised $39 million to expand its AI data center operations.
• E-commerce sales are expected to grow faster than overall retail sales during the upcoming holiday season.
• 80% of asset managers now use AI in product development, though only 26% use it across multiple workflow stages.

Cognition Valued at 48 Billion After Major Funding

Cognition, the company that makes the AI coding agent Devin, has raised new money that values it at 48 billion dollars. Major investors Andreessen Horowitz and Accel joined existing backers to support this growth. The company says its annual revenue has doubled to over 900 million dollars in just three months. Although Devin is not as famous as other tools like Claude Code, large companies like Citi and Mercedes-Benz use it heavily for complex coding tasks. Enterprise customers report that Devin can finish projects in days that usually take months.

Cognition Releases SWE-2 Model at Lower Cost

Cognition has released a new coding model called SWE-2 that performs nearly as well as top industry models but costs much less to run. The model scores 50.0 percent on a benchmark test, which is very close to the previous version SWE-1.7. It was built using Kimi K3, a large open model from Moonshot AI, and uses a new training method to save money. SWE-2 medium version costs 81 percent less on average while matching the performance of the older model. This new tool is now available for users on Devin Desktop, the command line, the web, and Fusion.

CoreWeave Launches Physical AI Field Engineering

CoreWeave has started a new service called Physical AI Field Engineering to help customers use their own data for real-world AI projects. Engineers from CoreWeave work directly with client teams in industries like automotive, aerospace, and robotics to build accurate models. For example, they helped the Aston Martin F1 team process race radio data quickly to make fast strategy decisions. This service uses CoreWeave tools like Weights and Biases and ARIA to ensure the AI works safely and reliably. The goal is to turn private engineering data into useful production AI systems.

CoreWeave Expands Into Physical AI With Field Engineers

CoreWeave is expanding its business by sending physical AI engineers to work inside customer teams. This new service helps companies in sectors like automotive and robotics deploy AI models using their own proprietary data. The engineers focus on building and maintaining systems that handle real-world industrial workflows rather than just providing cloud storage. This move allows CoreWeave to move beyond simple infrastructure and offer hands-on integration for high-stakes operations. The company aims to bridge the gap between technical AI skills and deep physical engineering knowledge.

Renesas Adds 64 Bit MPUs and Opens AI Lab

Renesas Electronics has added new 64-bit microprocessors to its RZ/G Series for use in human-machine interfaces and IoT devices. The company also joined the top tier of the Autoware Foundation to bring open-source autonomous driving software to its R Car platforms. Additionally, Renesas opened a new Physical AI and Robotics Lab in Beijing to focus on real-world AI applications. These steps connect the company more closely with industrial AI, automotive systems, and edge computing workloads. Investors are watching to see how these changes affect the company's future growth and design wins.

Alessio Vinassa prioritizes AI security and governance

Alessio Vinassa, a tech entrepreneur with over fifteen years of experience, shares his investment approach for emerging technologies. He believes artificial intelligence and cybersecurity must work together because AI systems handle sensitive data and automate critical decisions. Vinassa argues that leaders should use AI to support their judgment rather than replace it, ensuring human oversight remains in place. He evaluates companies based on their ability to solve real problems, adopt new tools consistently, and maintain strong governance structures. For Vinassa, good governance allows businesses to scale innovation responsibly without slowing down their progress.

NVIDIA NIM optimizations boost Nemotron 3 Ultra users

NVIDIA released updates to its NIM platform that help serve more users on the Nemotron 3 Ultra large language model. The optimized serving stack uses advanced techniques like cache reuse and speculative decoding to improve system performance. Testing on a system with four B200 GPUs shows the optimized version delivers 1,997 tokens per second compared to 718 for the baseline. This improvement represents a 2.5x increase in throughput, allowing applications to handle significantly more concurrent users while staying responsive. The NIM Certified version also includes regular updates and commercial support for enterprise deployments.

Vulcan Infrastructure raises $39 million for AI data centers

Vulcan Infrastructure and Power has closed a $39 million investment to pay off debt and expand its AI data center operations. The funding was led by a group of private investors, and the company's existing shareholders also participated in the deal. Vulcan plans to use these funds to grow its data center business and develop new services powered by artificial intelligence. The company aims to strengthen its position in the growing market for AI infrastructure through this financial boost.

E-commerce sales to grow faster than retail thanks to AI

E-commerce sales are expected to grow faster than overall retail sales during the upcoming holiday season due to increased use of artificial intelligence. Deloitte forecasts that total holiday retail sales will rise between 4% and 4.8% to reach between $1.7 trillion and $1.71 trillion. In contrast, e-commerce sales are predicted to grow between 7.5% and 8.4% to reach between $316.1 billion and $318.9 billion. Consumers are using digital tools like AI to research products and complete purchases more easily. Additionally, a projected rise in disposable personal income of 4.5% to 5.2% is expected to support these sales figures.

Asset managers prioritize private markets and AI adoption

A new survey by Nasdaq reveals that private markets are the top priority for asset managers across all regions. The 2026 Global Voice of the Issuer Study found that 80% of asset managers now use artificial intelligence in their product development processes. However, only 26% of these managers actively use AI across multiple stages of their workflow. The findings highlight a growing focus on private markets and the increasing role of AI in the asset management industry. This trend shows how technology and investment strategies are converging to meet client demands.

Sources

• Devin Creator Cognition Valued At $48 Billion After Backing From a16z, Accel & Others
• Cognition Releases SWE-2, Says It Performs Close To Frontier At 70% Lower Cost
• CoreWeave Launches Physical AI Field Engineering to Turn Proprietary Data Into Production AI
• What Does CoreWeave (CRWV) Gain From Its Physical AI Field Engineering Launch?
• Do New 64 Bit MPUs Change The Bull Case For Renesas (TSE:6723)?
• Alessio Vinassa Unveils an Emerging Technology Investment Approach Shaped by Financial Challenges
• How Full-Stack NIM Optimizations Deliver 2.5x More Users on Nemotron 3 Ultra
• Dealroom.co | Vulcan closes $39M investment to clear debt and fund AI data center push
• E-commerce to outpace overall holiday sales growth as shoppers lean on AI
• Private Markets and AI Emerge as Defining Priorities for Asset Managers as Regional Priorities Shape Growth Strategies