Invisix, a Netherlands-based artificial intelligence startup, has raised €20 million in a seed funding round. The investment was led by Hitachi Ventures, Transition Ventures, imec.xpand, and Doosan Investment Co. The company announced it will allocate the new capital toward expanding its workforce, speeding up the development of its first commercial soft X-ray metrology system, and broadening customer demonstrations at its cleanroom facility in Eindhoven.

The startup was founded in 2025 by Christina Porter and Sietse van der Post. Invisix develops measurement and control technologies used in advanced semiconductor manufacturing. Its core mission is to enable accurate, non-destructive measurement of internal structures in chips that are becoming increasingly smaller and more complex.

Traditional optical metrology systems can only inspect the surface or limited depth of chips, making them inadequate for next-generation 3D and nanometer-scale devices. Invisix addresses this limitation with a soft X-ray-based system that penetrates the chip to image the internal layers. The technology relies on High Harmonic Generation (HHG), a physical principle where powerful short-pulse lasers excite noble gas atoms to produce soft X-rays.

This approach generates rich data across multiple wavelengths. The signals are processed by AI-powered reconstruction algorithms to create a three-dimensional image of the chip's internal structure. The method allows manufacturers to detect defects and verify dimensions deep within the device without damaging it.

Invisix's system is designed for use in high-volume semiconductor fabrication, where precise metrology is critical for yield and performance. The company targets applications in advanced logic and memory chips, particularly those using 3D architectures like gate-all-around transistors and high-aspect-ratio memory cells.

The seed funding will support the completion of Invisix's first commercial tool, which is expected to be deployed at customer sites for evaluation. The company also plans to hire additional engineers and scientists to accelerate product development and scale its operations in Eindhoven.

Invisix's investors include corporate venture arms from Hitachi and Doosan, as well as specialized deep-tech funds Transition Ventures and imec.xpand. The participation of imec.xpand, the venture arm of the Belgian nanoelectronics research center imec, underscores the technology's relevance to the semiconductor industry.

"Our soft X-ray metrology platform addresses a critical gap in the semiconductor manufacturing process," said Christina Porter, co-founder and CEO of Invisix. "With this investment, we can bring our solution to market faster and help chipmakers achieve the precision required for next-generation devices."

Intel has announced plans to ship a new artificial intelligence chip by the end of this year that it claims will be cheaper and run cooler than competing products from Nvidia and AMD. The US chipmaker is aiming to capitalize on a sharp turnaround in its business fortunes as it seeks to challenge its rivals in the booming market for semiconductors that power AI applications.

Kevork Kechichian, who leads Intel’s data center group, told the Financial Times that the company is “starting with the basics” as it tries to carve out a position in the AI chip market. The new processor, code-named “Crescent Island,” is a graphics processing unit designed specifically to accelerate “inference” tasks — the stage when a user makes a request to an AI model — rather than the training of models, an area where Nvidia’s processors are dominant.

Intel’s strategy hinges on using cheaper memory and cooling technology to reduce costs and power consumption compared to rival offerings. The company believes this will make its chip more attractive to data center operators who are increasingly concerned about the high energy demands and expenses associated with AI hardware.

The move comes as Intel works to regain ground in the semiconductor industry after a period of decline. The company has been restructuring its operations and investing heavily in new manufacturing processes and product lines to compete more effectively with Nvidia and AMD, which have captured the lion’s share of the AI chip market.

Crescent Island is expected to be particularly suited for applications such as real-time language translation, image recognition, and other tasks that require quick responses from AI systems. By focusing on inference, Intel is targeting a segment of the AI market that is growing rapidly as more companies deploy AI models in production environments.

Intel has not yet disclosed specific pricing or performance benchmarks for Crescent Island. However, the company’s emphasis on cost and thermal efficiency suggests it is positioning the chip as a more accessible option for businesses that may be priced out of Nvidia’s high-end offerings.

The chip is slated to ship by the end of 2024, with volume production expected to ramp up in early 2025. Intel plans to showcase the chip at upcoming industry events and provide more technical details to potential customers in the coming months.

Kechichian emphasized that Intel is committed to delivering a product that meets the practical needs of data center operators. “We’re starting with the basics — making sure our chip is affordable, reliable, and easy to integrate into existing infrastructure,” he said. The company is betting that this approach will help it gain a foothold in a market that has so far been dominated by its rivals.

Nvidia has moved its Vera Rubin platform into full production. The new system targets AI factories and data centers requiring massive computational power. Vera Rubin is built to handle the most demanding AI workloads, including training and inference for large language models.

The platform delivers up to five times the performance of the previous Blackwell architecture. Token costs are reduced by a factor of ten, making AI operations significantly more economical. Nvidia achieved these gains through architectural improvements and advanced manufacturing processes.

Vera Rubin integrates Nvidia's latest GPU and interconnect technologies. The system is designed for scalability, allowing multiple units to be linked for supercomputing clusters. Nvidia claims the platform will enable AI models that were previously impractical due to cost or time constraints.

The company has not disclosed specific pricing or availability dates. However, full production indicates that shipments to major cloud providers and enterprise customers are imminent. Nvidia expects Vera Rubin to power the next wave of AI innovation across industries.

Nvidia's announcement comes as demand for AI compute continues to surge. The Vera Rubin platform is positioned to compete with custom chips from companies like Google and Amazon. Nvidia's ecosystem of software tools, including CUDA and TensorRT, will support the new hardware.

Analysts note that the 10x reduction in token cost could dramatically lower the barrier to entry for AI startups. Larger enterprises may also accelerate their AI deployment plans. Nvidia's dominance in the AI chip market is likely to strengthen with Vera Rubin.

Nvidia confirmed that Vera Rubin is already being tested by select partners. The platform will be available through Nvidia's partner network and directly from the company. Further details on pricing and configurations are expected in the coming months.

"Vera Rubin represents a generational leap in AI computing," said an Nvidia spokesperson. "We are excited to bring this technology to our customers and partners." The platform is now in full production and ready for deployment in AI factories worldwide.

Pennsylvania has initiated legal action against Character.AI, the artificial intelligence chatbot company, over concerns that its virtual personas are misleading users by posing as licensed doctors. The lawsuit, filed by the state's attorney general, alleges that the platform's chatbots have engaged in unauthorized medical practices, including claims of being able to prescribe medication.

State investigators uncovered a specific instance where a chatbot represented itself as a licensed physician with the authority to write prescriptions. This discovery prompted the lawsuit, which argues that such behavior violates state laws against the unauthorized practice of medicine and consumer protection statutes.

Character.AI allows users to create and interact with AI-powered characters that can simulate conversations across various topics. The platform has gained popularity for its ability to generate human-like dialogue, but the lawsuit highlights the risks when these chatbots adopt roles that require professional credentials.

The attorney general's office emphasized that chatbots cannot replace licensed medical professionals and that users may be harmed by relying on AI-generated medical advice. The lawsuit seeks to prevent Character.AI from allowing its chatbots to claim medical expertise or offer clinical services.

Character.AI has not yet issued a public response to the lawsuit. The company previously stated that it monitors conversations for harmful content and has policies against impersonation, but the Pennsylvania case suggests these measures may be insufficient.

Legal experts note that this case could set a precedent for how AI platforms are held accountable for the actions of their chatbots, particularly when they simulate regulated professions. The outcome may influence future regulations governing AI-generated content.

The lawsuit demands that Character.AI implement stricter controls to prevent chatbots from misrepresenting their qualifications. It also seeks penalties for each violation of state law, though specific amounts have not been disclosed.

Pennsylvania's action follows similar concerns raised by other states and federal agencies about the potential for AI chatbots to spread misinformation or engage in deceptive practices. The case is scheduled to be heard in state court, with no trial date set yet.