Tesla Lost Its Top Powertrain Engineer. The Robotaxi Timeline Should Worry You.

Drew Baglino spent 18 years at Tesla. He was one of only four named executive officers at the company. Every motor, every battery cell architecture, every thermal system that makes a Tesla actually work passed through his orbit as SVP of Powertrain and Energy. And in April 2024, he walked out the door.

This wasn't some quiet retirement. It happened the same week Tesla laid off more than 10% of its global workforce, right as Elon Musk publicly declared the company was going "all in" on the robotaxi concept. That timing tells you something. The stock price has been trying to figure out what ever since.

I've watched enough senior engineering departures in my career to recognize the pattern: when the person who understands the physics of your product leaves during a strategic pivot, you don't just lose institutional knowledge. You lose the ability to reality-check ambitious timelines. And Tesla's timelines were already the most aggressive in the industry.

This Isn't Just Another Executive Departure

Most coverage of Baglino's exit focused on optics. A top executive leaving during mass layoffs looks bad. Sure. But that's the shallow read.

The real problem is what Baglino represented inside Tesla's engineering org. Battery chemistry, motor design, thermal management, energy storage. These aren't side projects. They're the foundation every Tesla vehicle sits on, including any future robotaxi.

Here's the thing nobody's saying about robotaxis: they aren't just software on wheels. A robotaxi needs to run continuously for 16+ hours a day, manage thermal loads from always-on autonomous compute hardware, and maintain battery health across hundreds of thousands of miles with no owner babying the charge cycles. That's a powertrain and thermal engineering problem just as much as it's an autonomy problem.

Baglino was deeply embedded in Tesla's next-generation vehicle platform and battery architecture. At the 2023 Investor Day, he presented extensively on manufacturing innovations and next-gen powertrain efficiency targets. That work directly underpins whether a fleet vehicle's economics make any sense at all.

Losing him doesn't kill the robotaxi program. But it slows it down. Full stop.

The Stock Is Pricing In a Fantasy

Tesla's valuation has always been a bet on the future. That's not new. What's new is how specific that bet has gotten.

After Musk's pivot announcement, the narrative shifted hard. Tesla is no longer primarily valued as an automaker. A huge chunk of its market cap is now tied to the expectation that Tesla will operate a robotaxi network — essentially becoming an autonomous Uber with better margins. Morgan Stanley analyst Adam Jonas has been one of the most vocal proponents of this thesis, assigning substantial value to Tesla's "mobility-as-a-service" potential in his price targets.

But let's actually spell out what that valuation assumes. Tesla has to solve Full Self-Driving at a level requiring zero human intervention. It has to build a purpose-built robotaxi vehicle. It has to stand up a fleet management and maintenance operation from scratch. And it has to get regulatory approval across dozens of jurisdictions, each with their own rules and political dynamics.

Any one of those is a multi-year engineering challenge. Together? This might be the most ambitious product roadmap in the history of the auto industry. I'm not exaggerating.

Meanwhile, Waymo is actually running a commercial robotaxi service right now. Alphabet's autonomous driving unit is completing over 150,000 paid rides per week across San Francisco, Phoenix, and Los Angeles, based on data the company shared in late 2024. That's not a prototype. That's a business. And it took them over 15 years and billions of dollars to get there, using vehicles that cost significantly more than a standard car.

Toni Sacconaghi at Bernstein has repeatedly questioned whether Tesla's FSD capabilities are on a trajectory to achieve true Level 4 autonomy on Musk's timeline, pointing out that the gap between supervised FSD and unsupervised robotaxi-grade driving remains massive. That's not bearish posturing. That's engineering realism.

The Bench Is Getting Dangerously Thin

Baglino wasn't the only senior leader to leave. Rohan Patel, Tesla's VP of Public Policy and Business Development, departed at the same time. The layoffs cut deep into engineering teams across the board.

I've been building software and leading engineering teams for over 14 years, and I've seen firsthand what happens when organizations lose senior technical people during a strategic pivot. The remaining team doesn't just absorb the workload. They lose the person who could say "that won't work because of X" in a design review. The person who remembered why a particular architecture decision was made three years ago and why the alternative would have been a disaster. That kind of knowledge doesn't transfer through documentation. It walks out the door with the person.

This is especially dangerous at Tesla, where Musk's management style already concentrates decision-making at the top. Fewer senior technical voices in the room means a higher risk of optimistic timelines going unchallenged. And Tesla's track record on timelines is — let's be honest — terrible. The Cybertruck was years late. The Semi has been perpetually "next year." FSD has been "almost ready" since 2016.

I've seen this same dynamic play out in my own career building complex systems. Companies making bold strategic pivots almost always underestimate how much their execution depends on specific people, not just the strategy itself. Microsoft learned this with Xbox. Tesla may be learning it now.

Waymo's Lead Is Measured in Years, Not Months

While Tesla is still working toward unsupervised autonomy, Waymo has been quietly building the operational infrastructure that actually makes robotaxis work: remote assistance systems, fleet maintenance protocols, insurance frameworks, rider trust.

I've shipped enough production systems to know that the hardest 20% of any product is the operational reality. The edge cases. The maintenance schedules. The regulatory compliance that's different in every city. It's the boring stuff that determines whether a service actually works or just demos well.

Waymo has years of lead time here. They got there with dedicated hardware — lidar, radar, cameras — on vehicles purpose-built for autonomous operation. Tesla's approach, vision-only sensing on consumer vehicles, is more ambitious and potentially more scalable. But "potentially" is doing a lot of heavy lifting in that sentence.

The question isn't whether Tesla can build a robotaxi. It's whether they can build one fast enough to justify a stock price that already assumes they've done it.

For engineers and investors watching this space, the real signal isn't in Musk's announcements. It's in who's still at the company to execute on them. The engineering bench is shrinking while the ambition is growing. Something has to give.

The Boring Answer Is the Right One

Tesla's robotaxi event and subsequent product announcements will test whether the company can back its timeline with actual engineering substance. The rise of autonomous AI agents across the software industry shows that autonomous systems are maturing fast. But software agents operating in digital environments and physical vehicles navigating the real world are fundamentally different problems. The stakes — literally life and death — aren't comparable.

I think Tesla will eventually ship some version of a robotaxi. Musk's ability to attract engineering talent and capital is unmatched, and the underlying FSD technology is genuinely improving. But "eventually" probably means 2028 or 2030, not the timelines baked into the stock price today. Baglino's departure makes the later end of that range more likely.

If you're an engineer evaluating Tesla's technical claims, look past the demos and the AI Day presentations. Ask who's actually building the systems. Ask how deep the engineering bench goes after the layoffs. Ask whether a company can execute on its most ambitious roadmap ever with fewer of the people who understood the hardest problems.

Building a robotaxi network is a decades-long engineering effort, not a software update. Baglino knew that. And now he's gone.

---

Originally published on kunalganglani.com