Fast Track to 2028: LA's Clean Mobility Vision with AI

General
1

Attended the “Fast Track to 2028: Design Clean Mobility with AI” session organized by GACC West and German Consulate this morning at the Arts District. With the 2028 Olympics coming to LA, the pressure is on to solve our mobility nightmare.

The panel discussed using AI for:

  • Real-time traffic optimization across the entire metro area
  • Predictive modeling for public transit demand
  • EV charging station placement optimization
  • Autonomous shuttle networks for last-mile connectivity
  • Smart sensors for pedestrian flow and traffic signal coordination

LA Metro’s plans are ambitious: 2,700 zero-emission buses, 129,000 EV chargers by 2028, and a goal of 30% EVs on the road. The Olympics are meant to be a “transit-first” event across 70+ venues.

Here’s my hot take: LA’s car culture is so deeply embedded that even the best AI won’t fix it without serious policy changes. We need congestion pricing, dedicated bus lanes, and massive public transit investment FIRST. AI is a tool, not a silver bullet.

That said, the German approach to integrating AI into their mobility planning was fascinating. They’re treating transportation as a data problem, not just an infrastructure problem. The coordination between traffic signals, public transit, micromobility, and autonomous vehicles - all informed by real-time AI analytics - is genuinely impressive.

The equity question: One panelist noted that wealthier neighborhoods already have better transit, bike infrastructure, and will likely get EV chargers first. How do we ensure AI-powered mobility solutions don’t just optimize for affluent areas?

Anyone else at this session? What were your takeaways?

#CleanMobility #UrbanPlanning #LA2028 #ClimateAction #SmartCities

2

Marcus, you’re spot-on about policy being more important than technology here. I’ve worked on traffic optimization systems, and the tech is actually pretty mature - the bottleneck is always political will and implementation.

Technical Reality Check:

The AI/ML parts everyone gets excited about are mostly:

  1. Time series forecasting for transit demand (we’ve had this for years, just getting better)
  2. Reinforcement learning for traffic signal optimization (works in simulation, messy in reality)
  3. Computer vision for pedestrian/vehicle detection (reliable but expensive to deploy at scale)
  4. Route optimization for shuttles and micromobility rebalancing (this actually works well)

None of this is groundbreaking tech. The hard parts are:

  • Data integration - Getting transit, traffic, weather, event data all into one system
  • Real-time processing - Making decisions fast enough to matter
  • Edge cases - What happens when the system fails? Manual override?
  • Privacy - Tracking pedestrian flow without becoming surveillance

The 129,000 EV Charger Challenge

This is actually fascinating from an infrastructure standpoint. Optimal placement requires:

  • Current traffic patterns
  • Predicted EV adoption rates by neighborhood
  • Grid capacity and upgrade costs
  • Parking availability
  • Equity considerations (your point about wealthy areas)

AI can help with the optimization, but someone still has to dig the trenches, upgrade transformers, and negotiate with property owners. That’s where projects slow down.

What I’d Want to Know:
Did they discuss open data standards? If every city builds proprietary AI systems, we’re going to have a mess. We need interoperability.

Also, @product_david might have thoughts on the business model here - who pays for this infrastructure? Public-private partnerships always sound great until you hit the details.

3

Happy to weigh in on the business model question, @alex_dev!

The Public-Private Partnership Dilemma

LA’s approach is actually a case study in how NOT to do PPPs (in my opinion). Here’s the breakdown:

Public Sector Investment:

  • Metro is spending billions on zero-emission buses and infrastructure
  • Federal grants (Infrastructure Investment and Jobs Act) covering some costs
  • Local bonds and tax measures (Measure M, etc.)
  • Goal: Public good, environmental targets, Olympic legacy

Private Sector Play:

  • EV charging companies (ChargePoint, EVgo, Tesla) want the real estate
  • Mobility-as-a-Service platforms want the data and rider relationships
  • Autonomous vehicle companies want testing grounds
  • Goal: Profit, market share, data collection

The problem? Misaligned incentives.

Private companies will cherry-pick profitable locations (dense areas, wealthy neighborhoods, tourist zones). Public sector wants equitable coverage. AI optimization will recommend efficiency, not equity.

Market Opportunity

That said, there’s a HUGE opportunity here for companies that can solve the “last mile” problem:

  • Micromobility integration (bikes, scooters)
  • Demand-responsive shuttles
  • Unified payment/routing platforms

The winner will be whoever builds the “Uber/Lyft for public transit” - a single app that seamlessly combines Metro, buses, bikes, ride-shares, and walking directions with real-time AI-powered recommendations.

Companies like Remix (acquired by Via) and Moovit (acquired by Intel) saw this coming. LA needs to decide: build it themselves or partner?

The Olympics Deadline is Both Blessing and Curse

BLESSING: Creates urgency, unlocks funding, political will
CURSE: Shortcuts get taken, equity gets sacrificed for speed

My prediction: LA will deliver a flashy “transit-first” Olympics experience for venues and tourist areas. But the structural car-dependency won’t change. The real test is what happens in 2029.

@marcus_urbanist - curious your take on whether the Olympics deadline helps or hurts long-term sustainable planning?

4

As someone who builds ML systems, I’m fascinated by the data challenges underlying all of this. Let me break down what’s actually happening under the hood:

The Data Problem

To make AI-powered mobility work, you need:

  1. Historical data: Years of traffic patterns, transit ridership, weather correlations
  2. Real-time data: Live vehicle locations, pedestrian sensors, parking availability
  3. Predictive data: Event schedules, construction, special circumstances (Olympics!)
  4. External data: Ride-share patterns, micromobility usage, even airline arrivals

The challenge isn’t building the model - it’s getting clean, consistent, integrated data. Most cities have this spread across 15 different departments with incompatible formats.

Equity Through Data

@marcus_urbanist your equity question is critical. AI will optimize for what you measure. If you measure:

  • “Minimize average commute time” → Prioritizes dense areas and highways
  • “Maximize ridership” → Optimizes tourist/business districts
  • “Improve access for underserved communities” → Different solution entirely

The problem: that last one is WAY harder to quantify. How do you measure “access” fairly? Do you weight by:

  • Population density?
  • Income levels?
  • Current transit gaps?
  • Job access?

This isn’t a technical problem - it’s a values problem that requires political choices. The AI just amplifies whatever you choose to optimize.

The 30% EV Goal

I’d love to see the model behind this. Some questions:

  • What’s the baseline? (LA is currently ~7% EV adoption)
  • Is this achievable without massive subsidies?
  • What happens if charging infrastructure can’t keep up with demand?
  • How do you model the feedback loop? (More chargers → more EVs → more demand → need more chargers)

Prediction:

They’ll hit their numbers for the Olympics through aggressive short-term interventions. But the models will be overfit to Olympic conditions. Post-2028, when event traffic drops and political pressure relaxes, we’ll see regression to mean.

The real test: are they building systems that learn and adapt, or are they building systems optimized for a single deadline?

@alex_dev is right about open data standards. Without that, all these city-specific AI systems become expensive, fragile tech debt.

5

This thread perfectly illustrates why urban infrastructure projects are so complex - you need technical, business, data, and policy expertise all working together.

Infrastructure at Scale: The CTO Perspective

I’ve led large-scale infrastructure projects, and here’s what keeps me up at night about LA’s approach:

Technical Debt at City Scale

  • Legacy systems: LA’s traffic management runs on systems from the 1990s
  • Integration hell: Connecting new AI systems with 30-year-old infrastructure
  • Vendor lock-in: Once you pick a platform (SAP, Siemens, whoever), you’re married to them
  • Maintenance burden: Who maintains these AI models in 2030? 2040?

Organizational Readiness

The technology is the easy part. The hard part is:

  1. Cross-departmental coordination - Transit, traffic, parking, events, police, fire all need to share data and coordinate
  2. Change management - Training thousands of city employees on new systems
  3. Governance - Who decides when the AI’s recommendation gets overridden?
  4. Accountability - When something goes wrong, who’s responsible? The algorithm? The city planner who trusted it?

The Build vs Buy Decision

@product_david mentioned this - LA faces a critical choice:

Build in-house:

  • PRO: Customized to LA’s unique needs, public ownership
  • CON: Expensive, slow, requires hiring/retaining top AI talent (competing with tech companies)

Buy commercial:

  • PRO: Faster deployment, vendor support
  • CON: Expensive licensing, vendor dependence, may not fit LA’s needs

Partner/Open Source:

  • PRO: Shared costs, interoperability, community support
  • CON: Coordination complexity, slower decision-making

My recommendation: Hybrid. Build the integration layer and policy engine in-house. Use commercial/open-source for the AI models.

The Real Question

@data_rachel nailed it - are they building systems that learn and adapt, or systems optimized for a single event?

If it’s the latter, LA will spend billions on infrastructure that becomes obsolete the day after the closing ceremony.

The Olympics should be a forcing function for long-term sustainable transportation, not a vanity project that looks good on TV for two weeks.

@marcus_urbanist - what’s your read on whether the city leadership actually gets this?

6

This is exactly the kind of multi-disciplinary discussion we need! Let me respond to the excellent points raised:

@alex_dev on open data standards:
YES. The German presenters actually emphasized this - they use the Mobility Data Specification (MDS) standard across cities. LA is… not there yet. Each department has its own data silos. LADOT, Metro, and the city’s transportation department barely talk to each other, let alone share standardized data.

The session mentioned SAP as a partner, which worries me about vendor lock-in exactly as @cto_michelle described.

@product_david on the Olympics deadline:
You nailed it. The deadline is both blessing and curse. Here’s my take:

BLESSING:

  • Unlocked Measure M funds that were sitting unused
  • Created political cover for congestion pricing pilots (maybe)
  • Forced Metro to finally coordinate with micromobility companies

CURSE:

  • “Transit-first” will mean building showcase routes from airports to venues
  • Underserved communities get lip service, not actual investment
  • Post-2028, maintenance budgets will be slashed and we’ll have shiny new buses sitting unused

Historical precedent: Look at Athens 2004, Beijing 2008, Rio 2016. Massive transit investments became white elephants.

@data_rachel on equity metrics:
This is THE question. At the panel, I asked specifically: “What equity metrics are you optimizing for?”

The answer was vague bureaucratic speak about “community input” and “environmental justice.” No concrete metrics. Which tells me they haven’t actually figured this out.

Your point about AI optimizing what you measure is critical. If the metric is “minimize Olympic visitor wait times,” guess which neighborhoods get prioritized?

@cto_michelle on city leadership:
Honestly? Mixed bag.

  • Metro leadership: Actually gets it. They’ve hired some smart people who understand both tech and policy.
  • City Council: Half don’t understand the tech, half don’t care about equity, very little overlap in the Venn diagram.
  • Mayor’s office: Wants the flashy Olympics win. Long-term sustainability is someone else’s problem.

The most telling moment at the session: A German transit official asked about car-free zones and congestion pricing. The LA panelist literally laughed and said “we’re not Copenhagen.”

That attitude - that LA’s car culture is immutable - is the real barrier. No amount of AI will fix that.

My Actual Prediction:
2028: Impressive transit showcase for Olympic venues and tourist areas
2029: Ridership drops, maintenance budgets cut, some bus routes eliminated
2030: Conference papers written about “lessons learned” from LA’s Olympic transit investment
2032: Rinse and repeat for the next city

I hope I’m wrong. But I’ve seen this movie before.