When Nvidia CEO Jensen Huang declared last January that the “ChatGPT moment for general robotics is just around the corner,” he tapped into the tech industry’s favorite narrative: that artificial intelligence breakthroughs translate seamlessly into real-world adoption. But as businesses are discovering, the path from AI promise to commercial reality is far more treacherous for physical robots than for software chatbots.
The Kroger Conundrum
Consider the case of Kroger, the US supermarket giant that announced in November it would close three of its eight robotic warehouses. These facilities, powered by UK company Ocado’s technology, represented a significant investment in automation. Yet Kroger simultaneously expanded its relationship with gig economy platforms like Instacart and DoorDash, essentially choosing human “personal shoppers” over robotic ones.
“It’s a bit more like a factory set-up: it needs to have minimum throughput to make money,” explains Tom Andersson, a warehouse automation expert at research company STIQ. “When you do those business cases – because sometimes these projects can take three years in the planning – if your forecast is wrong at that point it will be tricky.”
The Startup Struggle
This disconnect between technical capability and commercial viability extends beyond robotics to the broader AI startup ecosystem. According to analysis from the Financial Times, AI startups face significant disadvantages compared to established enterprise software platforms. They typically spend double what traditional SaaS companies spend on compute and infrastructure while building specialized software on top of standard AI models like ChatGPT or Claude.
“The explosive rise of AI start-ups follows a pattern we’ve seen before: small companies racing to apply new technology to specific business problems, promising huge efficiency gains in their narrow slice of the market,” notes a managing partner at Thoma Bravo. Established platforms like Salesforce, SAP, and Microsoft have decades of industry knowledge, existing software integrations, and regulatory compliance infrastructure that give them a substantial edge.
The Humanoid Hurdle
Even the much-hyped humanoid robots face practical limitations that challenge Huang’s prediction of human-level capabilities this year. First, safety concerns loom large: while a hallucinating chatbot might damage your reputation, a malfunctioning humanoid robot could cause physical harm in a workplace setting.
Second, battery technology remains a significant constraint. As James Pikul, associate professor of mechanical engineering at the University of Wisconsin-Madison has documented, Boston Dynamics’ Spot robot can only operate for about 90 minutes before needing to recharge. Compare that to human workers who routinely handle 10-hour shifts with brief breaks.
The Motional Pivot
Perhaps the most telling example comes from the autonomous vehicle sector. Motional, the joint venture between Hyundai Motor Group and Aptiv, recently rebooted its robotaxi plans after facing financial and operational challenges. The company, which reduced its workforce from 1,400 to less than 600 employees, has shifted from a classic robotics approach to an AI foundation model-based system.
“We saw that there was tremendous potential with all the advancements that were happening within AI; and we also saw that while we had a safe, driverless system, there was a gap to getting to an affordable solution that could generalize and scale globally,” explains Laura Major, President and CEO of Motional. The company now plans to launch a commercial driverless service in Las Vegas by the end of 2026.
The Adoption Equation
What determines whether AI succeeds in workplace adoption? According to FT contributing editor analysis, effective AI must be comfortable, explainable in under a minute, and not make users look incompetent. This echoes historical technology failures like virtual reality – the VR headset market has fallen by an annual 14% according to Counterpoint Research – and highlights generative AI’s current tendency to mix facts with fantasy.
Even in industries where AI promised to automate entry-level work, reality has proven more complex. Despite AI’s potential to automate legal research, junior lawyers’ salaries have soared, and lawyers in the UK and US have faced trouble for citing non-existent cases after using AI tools.
The Regulatory Reality
Beyond commercial considerations, regulatory challenges are mounting. The UK’s media regulator Ofcom has launched a formal investigation into X’s AI chatbot Grok over concerns it’s being used to create and share sexualized deepfakes of women and children. Ofcom can fine X up to �18 million or 10% of its global revenues under the Online Safety Act if it finds violations.
This follows actions by Malaysia and Indonesia, which became the first countries to block access to Grok due to deepfake concerns. U.S. Senator Ron Wyden has called for Grok’s removal from Google and Apple app stores until Elon Musk addresses what he calls “disturbing and likely illegal activities.”
The Bottom Line
What does all this mean for businesses considering AI and robotics investments? The evidence suggests a more nuanced approach than the hype would indicate. While automation will undoubtedly increase across physical settings, successful implementation requires careful consideration of business cases, integration challenges, and practical limitations.
As the Kroger example demonstrates, just because a technology exists and works doesn’t mean it will always be commercial to deploy it. The fact a job can be automated doesn’t necessarily mean it will be automated. For businesses navigating this landscape, the key insight might be this: in the race between robots and reality, reality still sets the pace.