February 4, 2026 was a day of big moves and bold promises across the artificial intelligence industry, with companies ranging from Silicon Valley to global tech giants and security innovators announcing new funding, product lines, and strategic expansions. This news paints a vivid picture of a changing industry where AI is moving from a behind-the-scenes tool to a visible, sometimes autonomous colleague, and even a physical presence in the workplace.
In the San Francisco Bay Area, startup Expert Intelligence made headlines by announcing $5.8 million in seed funding, a round led by Sierra Ventures with participation from TSVC Management and Acorn Pacific Ventures. C&EN. Founded in 2022 by machine learning expert Larin Teverappelma, the company offers a platform designed to automate regulated laboratory workflows, particularly the time-consuming and often subjective process of data interpretation in areas such as drug development and manufacturing.
Theverapperuma explains: “These instruments may be able to generate terabytes or gigabytes of data in minutes, but analysis and interpretation can take days or weeks. This is the real bottleneck. … And the interpretation of the data is often based on intuition and knowledge gathered by human scientists over many years.” The company’s Limited Sample Model is a machine learning system trained on the interpretations of human scientists that promises to reduce analysis time to hours and provide accuracy comparable to that of a master chemist. The tool can learn from relatively small datasets, distinguishing it from the large language models that dominate AI conversations, and is designed to adhere to strict U.S. Food and Drug Administration guidelines.
Expert Intelligence already has over 10 customers, including three major pharmaceutical companies and several food and beverage manufacturers. Tiberappelma’s confidence in the reach of his platform was clear. “If you ate breakfast today, we are confident that some of the ingredients were provided by one of our customers.” The new funding will be used to strengthen the company’s marketing team and accelerate deployment at customer sites, reflecting both the urgency and demand for automation in regulated industries.
But the AI news of the day wasn’t limited to startups. At the Cisco AI Summit, Intel CEO Lip-Bu Tan announced that the company will enter the graphics processing unit (GPU) market, which is currently dominated by Nvidia. According to ReutersIntel’s move is a strategic expansion beyond its traditional focus on CPUs, as it aims to gain market share in both gaming and AI model training, areas where GPUs are essential. The effort will be overseen by Kevork Keshichian, executive vice president and general manager of Intel’s Data Center Group, with support from Eric Demers, a former Qualcomm engineering executive. Tan said Intel’s GPU strategy is still in its early stages and will be shaped by customer demand, but the intent is clear: to challenge Nvidia’s dominance in AI-focused computing.
While hardware giants and startups push the boundaries of what AI can do, the AI & Big Data Expo and Intelligent Automation Conference, which kicked off on the same day, focused on the real-world challenges of integrating AI into enterprise workflows. According to AI newsthe conference’s technical sessions centered around the evolution from passive automation to “agent” systems, i.e. AI tools that can reason, plan, and execute tasks autonomously, rather than simply following a strict pre-programmed script.
Citi’s Amal Makwana described these agent systems as digital coworkers, a sentiment echoed by DeepL’s Scott Ivell and Ire Adewolu, who argued that such AI closes the “automation gap” by closing the distance between intent and execution. But as SS&C Blue Prism’s Brian Halpin pointed out, organizations will need to master traditional automation before deploying agent AI, and this change requires a robust governance framework to deal with the unpredictable and non-deterministic outcomes these systems produce.
Informatica’s Steve Holyer, along with speakers from MuleSoft and Salesforce, emphasized that careful oversight is essential when designing these systems. A governance layer must control how AI agents access and use data to prevent operational failures. SAP’s Andreas Krause went a step further, warning that without trusted, connected enterprise data, AI is doomed to fail. “For GenAI to work in an enterprise context, it needs access to accurate and context-relevant data,” said Kraus. Gigaspaces’ Meni Meller tackled the notorious problem of AI “hallucinations” by proposing search augmented generation (eRAG) combined with a semantic layer to enable models to capture real corporate data in real time.
Physical safety has also become a hot topic, as integrating AI into factories, offices, and public spaces poses different risks than traditional software failures. A panel featuring Edith-Clare Hall from ARIA and Matthew Howard from IEEE RAS discussed the need for safety protocols before robots interact with humans. Perla Maiorino from the Oxford Robotics Institute provided a technical perspective, sharing research on time-of-flight sensors and electronic skins that give robots both self-awareness and environmental awareness, which is important to prevent accidents in manufacturing and logistics.
Observability in software development was also emphasized. Datadog’s Yulia Samoylova said that as AI systems become more autonomous, teams need new ways to observe their internal state and inference processes to ensure reliability. Julian Skeels from Expereo argued that network infrastructure needs to be purpose-built for AI workloads, requiring a sovereign, secure, “always-on” network that can handle high throughput.
But even the best technological solutions can stumble without cultural buy-in. IBM Automation’s Paul Fermor warned against the “illusion of AI readiness” and stressed that strategies must be human-centric to ensure adoption. Jena Miller emphasized this, pointing out that technology won’t bring any benefits if employees don’t trust the tools. Sanofi’s Ravi Jay suggested that leaders need to make operational and ethical decisions early in the process to decide where to build their own solutions and where to buy established platforms.
Meanwhile, in the security space, Artificial Intelligence Technology Solutions Inc. (AITX) announced an expansion order from a leading auction operator, as detailed in a February 4, 2026 SEC filing. The order includes two SARA-licensed RIO 360 units and two RIO Mini units following a successful initial deployment in October 2025. This expansion reflects growing customer confidence in RAD’s autonomous security solutions, which are specifically designed to secure large outdoor auction venues. RAD will present a live demonstration of these solutions at ISC West 2026, highlighting the growing demand for AI-powered security in complex environments.
From laboratory automation and enterprise AI governance to hardware innovation and physical security, developments in February 4, 2026 make it clear that the industry is not only evolving, but accelerating. As AI systems become more capable, there is increasing pressure to ensure that they are deployed safely and ethically, explicitly considering both technical and human realities. The stakes are high and the race is on.
