Summary
The second day of TechEx North America focused on the practical challenges of using artificial intelligence in large companies. While many businesses are excited about AI, many projects fail to move past the testing phase, a problem experts called the "AI graveyard." Speakers at the event shared strategies to help businesses scale their technology, improve security, and use AI in physical machines like robots. The discussions highlighted that success with AI requires more than just good software; it needs strong data foundations and careful planning.
Main Impact
The most significant takeaway from the event was the growing "velocity gap" between AI development and corporate security. Companies are adopting new AI tools so quickly that security teams are struggling to keep up. This creates a major risk where sensitive company data might be used in unapproved tools, a trend known as "shadow AI." To solve this, experts suggested using a "zero trust" approach, which means every person and every AI agent must prove their identity and have limited access to data by default.
Key Details
What Happened
During the sessions at the San Jose McEnery Convention Center, experts looked at why AI projects often stall. A common problem is the "personal copilot" effect. An AI tool might work perfectly for one person or a small team, but it often fails when a company tries to roll it out to thousands of employees. This happens because the infrastructure and data needed for a whole department are much more complex than what is needed for a single user. The event also introduced a new track focused on Physical AI, which looks at how AI can control hardware and robots in the real world.
Important Numbers and Facts
The conference featured several hands-on learning opportunities, including a Google Hackathon and workshops led by Nvidia. These sessions taught developers how to build "agentic AI," which are systems that can take actions and improve themselves over time. Attendees also discussed the costs of AI, specifically how "token-based" pricing affects a company's budget. This type of pricing charges businesses based on how much data the AI processes, which can become expensive if not managed correctly. The next major TechEx event is scheduled to take place in Amsterdam this September.
Background and Context
For the past few years, businesses have been in a rush to test AI. However, 2026 is becoming the year of reality. Many organizations have spent a lot of money on AI pilots that never turned into useful tools. This matters because investors and business leaders are now looking for a real return on investment. They want to see AI actually saving time or making money, rather than just being a cool experiment. Understanding the "buy versus build" debate is also a big part of this context, as companies decide whether to create their own AI systems or pay for existing services.
Public or Industry Reaction
The reaction from the crowd was a mix of caution and excitement. While the talk of "stalled projects" was serious, there was a lot of energy around the robotics displays. Humanoid robots were a major attraction on the show floor, showing that people are still very interested in the future of automation. Developers at the event were particularly happy with the "Learning Hub," where they could use tools like Google Colab to write code and see AI agents work in real-time. Industry leaders noted that while software coding has seen the most benefit from AI so far, physical industries like manufacturing are next in line.
What This Means Going Forward
Going forward, the focus will shift from simple chat tools to more complex "agentic" systems. These are AI tools that can work in the background to complete specific business tasks without constant human help. However, this will require companies to fix their data first. If the data is messy, the AI will not work. Businesses will also need to focus on "durable ROI," meaning they need to find ways for AI to provide value for a long time, not just for a few months. Security will remain the biggest hurdle, as teams try to close the gap between fast innovation and safe operations.
Final Take
The message from TechEx is clear: the honeymoon phase of AI is over, and the hard work of integration has begun. Success in the near future will not come from the flashiest technology, but from the most disciplined companies. Those who focus on strong security, clean data, and practical scaling will avoid the "AI graveyard" and see real benefits from their investments. As the industry moves toward physical AI and autonomous agents, the gap between those who can manage the technology and those who cannot will only grow wider.
Frequently Asked Questions
What is the "AI graveyard"?
The AI graveyard refers to artificial intelligence projects that work well during a small test or pilot phase but fail to work effectively when a company tries to use them across the entire business.
What is "shadow AI"?
Shadow AI happens when employees use AI tools or software that have not been approved by their company's IT or security department. This can lead to security risks and the leaking of private company data.
How does "zero trust" help with AI security?
Zero trust is a security model that assumes no user or machine should be trusted by default. It requires every AI agent and person to verify their identity before they can access any part of the company's network.
