Autonomous vehicle (AV) technology has made remarkable strides in the past few years, but the futuristic vision of fully self-driving cars dominating city streets remains just out of reach. While companies like Waymo (a subsidiary of Google), Tesla, and Zoox pilot robotaxi services in select tech hubs and urban markets, widespread adoption is slow and deliberate, hindered by regulatory barriers, safety concerns, and the complexities of scaling the technology.
To help decode the current state of AV, we spoke with Abhishek Nanda, a tech strategist and investor with deep expertise in connected and autonomous vehicle systems. Having worked on Microsoft’s Connected Vehicle Platform and with a career rooted in tech investments and M&A strategy, Nanda offers a sharp, insider perspective on the advances and obstacles shaping the industry.
Self-driving cars have fascinated the public for years. Can you summarize how the technology has evolved in the past decade?
Abhishek Nanda: Yes, so most notably, the biggest breakthroughs have stemmed from advances in artificial intelligence, but hardware and software have also seen steady progress.
SAE International defines six levels of autonomy, from Level 0, no automation, to Level 5, full automation. In the 2010s, we saw major progress in Level 1 and 2 features, like adaptive cruise control and lane centering—basically the early building blocks for autonomy.
Level 3 and 4 are where the car is capable of self-driving under certain conditions, but still needs some degree of human intervention. Most of today’s autonomous systems, like those from Waymo, operate at Level 4. These vehicles can operate without assistance in specific, controlled conditions—like mapped urban areas—but aren’t ready for every environment. Level 5, where a car can handle any situation without human intervention, remains the ultimate goal. Tesla’s recently announced Cybercab might be a step forward, but there’s still a clear gap between current capabilities and where the technology needs to be for full autonomy.
What recent developments have helped push the industry toward Levels 3 and 4?
Nanda: First, as an industry, we’ve made some critical improvements in the hardware capabilities. On the hardware side, better sensors—LiDAR, cameras, and radar—have significantly improved how cars perceive their surroundings. These sensors are now capable of creating highly detailed, real-time 3D imaging of their environment.
We’ve also seen incredible progress in both cloud-based mapping tools and onboard software. High-definition maps, for instance, provide granular data on roads, signs, and traffic patterns. High-definition maps provide granular data about roads, traffic patterns, and infrastructure, while local decision-making algorithms in the vehicles themselves are reaching near-human safety levels—though largely in limited, controlled settings.
Connectivity has also improved dramatically. Telematics and in-car infotainment systems are emerging as valuable tools for predictive maintenance and enhancing the passenger experience. We’re learning how to better equip the cars for autonomy along the way.
Did these trends shape your work on Connected Vehicles at Microsoft?
Nanda: Definitely. During my time with Microsoft, we anticipated the shift toward connectivity and autonomy. In 2016, we launched Microsoft Connected Vehicle Platform (MCVP) to offer services to automakers, fleet operators, and AV companies.
The idea was to be a ‘neutral’ provider, supplying the tools that the entire ecosystem, like cloud infrastructure and advanced mapping, would need as the industry scaled. I believe the bet paid off. Since then, Microsoft has expanded into areas like AVOps (Autonomous Vehicle Operations) and Connected Fleets, addressing more emerging needs in the industry.
AI seems to be a critical technology here. What’s its role, and what’s next for AI in this space?
Nanda: AI powers nearly every aspect of the self-driving stack. It processes inputs from sensors, predicts the behavior of other vehicles and pedestrians, plans paths, and even handles voice commands inside the cabin. That said, different companies have used somewhat different AI techniques to crack the self-driving challenge.
Google’s Waymo, for instance, breaks the problem into discrete tasks and uses specialized AI models for each. Tesla, on the other hand, is relying more on what is called “imitation learning,” feeding AI with massive amounts of driving data to mimic human behavior. Both methods are promising, and the race to Level 5 autonomy will likely hinge on which approach can scale and adapt faster.
From an investor and M&A perspective, where do you see opportunities in the AV space?
Nanda: The core self-driving technology is dominated by big players like Tesla and Waymo, but there’s a large ecosystem of startups working on critical supporting technologies. The development, testing, and ops behind these self-driving systems is still supported by independent companies, which are essential but not always the headline-grabbing aspects of the industry.
There’s a lot of opportunity here for investors to get in early. Take Foretellix, for example. They recently raised $85 million to develop a platform that verifies and validates autonomous systems, ensuring they can handle edge cases like unpredictable pedestrians or unusual traffic conditions. And these will be some of the hardest challenges to address.
What’s next for the self-driving car industry? What challenges do you foresee?
Nanda: The biggest hurdle is cost. High-end sensors like LiDAR can cost tens of thousands of dollars per vehicle, which makes it hard to scale services like robotaxis at a price point that’s accessible to the average consumer. It remains to be seen if companies like Tesla can achieve full autonomy with lower-cost hardware by mostly relying on cameras and using imitation learning to train the AI algorithms.
Then there’s regulation. Governments need to strike a balance between promoting innovation and ensuring public safety. And let’s not forget consumer trust—it’ll take time for people to feel fully comfortable handing over control.
But it’s a technology worth pursuing. Autonomous vehicles could reshape urban life by reducing congestion, improving mobility for people who can’t drive, and even altering how we think about car ownership. The road ahead has its challenges, but the destination is well worth the effort.
