serve_investorxdeckx07ma 
Proprietary and Confidential Investor Presentation 2026 Autonomous robotics at scale.→
2Proprietary and Confidential Forward-looking statements and disclaimers This presentation contains forward-looking statements within the meaning of Section 27A of the Securities Act of 1933, as amended, and Section 21E of the Securities Exchange Act of 1934, as amended. These statements are not historical facts or guarantees of future performance. Forward-looking statements include statements regarding the Company's future revenue generation, business and investment strategy, timing of robot manufacturing and deployment, ability to expand to additional markets, capabilities of the Company's robots, outcomes of planned and completed acquisitions, partnerships with multiple delivery platforms, and timing and ability to scale to commercial production. Actual results may differ materially from what is expressed or forecast. These statements are subject to risks and uncertainties described in our SEC filings, including our Annual Report on Form 10K for the year ended December 31, 2025. The Company disclaims any obligation to update forward-looking statements except as required by law. Market data in this presentation is obtained from third-party sources. Although we believe these sources are reliable, we have not independently verified the information. Third-party trademarks are the property of their respective owners.
3Proprietary and Confidential Every day, millions of small packages travel short distances in oversized vehicles. We built the robots that make that obsolete. Serve is a leading autonomous robotics company operating scaled commercial Level 4 fleets across 44 U.S. cities.
4Proprietary and Confidential Category creator The largest autonomous robot fleets operating commercially in the U.S. Multi platform Integration with Uber Eats and DoorDash for food delivery; Partnerships with 26 hospitals in the U.S. Triple Digit% Rev Growth Outlook for 2026 revenue: $26M (vs $2.7M in FY2025 Diversified Revenue ᐧ Last-mile Delivery ᐧ Healthcare ᐧ Branding & Advertising ᐧ Software Platform ᐧ Data Funded to scale Strong balance sheet with flexibility to invest opportunistically Serve Flywheel More Data → Better Models → Better Robots → Stronger Revenue → More Robots → More Data Serve Robotics is an industry leader in Physical AI.
5Proprietary and Confidential $450B Robotic & drone delivery opportunity by 2030 2.5 mi Median food delivery distance in the U.S. $810 Current per-delivery cost with human couriers Sources: ARK Big Ideas 2025, NHTSA, Company estimates | © 2026 Serve Robotics Inc. Short trips. Small packages. Massive inefficiency. The median U.S. food delivery is 2.5 miles – yet it travels in a 2-ton car Last mile delivery is broken. Robots fix it. The Opportunity
6Proprietary and Confidential 1.8M Sidewalk + Hospital deliveries To date Zero Major incidents Serious injury, fatality, or near miss 99.8% Completion rate Industry-leading reliability 2,000 100 Moxi Hospital robots deployed 44 U.S. cities 150+ neighborhoods 14 states We donʼt just build robots. We build delivery networks. Largest autonomous sidewalk fleet in the United States. Fleet grew 20x in one year. All figures based on internal operational data | © 2026 Serve Robotics Inc. Traction Serve Sidewalk robots deployed
7Proprietary and Confidential Commercial partner integrated since inception. Multi-year strategic partnership. Rollout underway in select U.S. cities. We don't compete with delivery platforms. We power them. Our robots are integrated into the two largest U.S. delivery apps - creating built-in demand from day one in every new market. Embedded in the leading platforms in U.S. food delivery The Ecosystem 40% Avg QoQ delivery growth Compounding since 2022
8Proprietary and Confidential A commercially scalable ecosystem backed by tier 1 partners across hardware, compute, and sensing. World-class hardware partners The Ecosystem Leading provider of high-resolution digital lidar sensors. Tier 1 automotive contract manufacturer. Gen3 production at scale. Jetson Orin compute platform. Powers Gen3 autonomy stack.
Compelling value proposition for merchants, consumers, and delivery platforms Economics With Serve Rising labor costs and regulations. Insurance overhead. High turnover. Limited operating hours. No driver wages. 14-hour operation. 65% lower hardware cost with Gen3. High utilization through platform integration. $810 per delivery status quo $1 expected delivery cost at scale Today →
10Proprietary and Confidential Every dollar of revenue funds more robots → more data → better models → more deployments → more revenue. 2,000 Serve robots on sidewalks. 100 Moxi robots in hospitals. Multi-domain, real- world data no one else can match. 1 Data End-to-end models train across our domains. What a robot learns in LA helps a robot in Dallas or with navigating a hospital corridor. 2 Models Scale autonomy onto live fleets. Uber Eats + DoorDash deliver demand. Every deployment generates new data. 3 Deploy Delivery, advertising, healthcare contracts, software, and platform revenue funding the next turn of the flywheel. 4 Monetize Real world data at scale fuels faster learning The Serve Flywheel Serve Physical AI Platform
11Proprietary and Confidential We donʼt just operate robots. We operate a platform that powers Physical AI. Our stack is end-to-end: perception, localization, planning, connectivity, and remote supervision. Itʼs been built and optimized across millions of miles of real-world operation. Indoor / Outdoor robot environments Applications Perception · Localization · Planning · Mobile manipulation Autonomy Low-latency assist operations Connectivity End-to-end models · Cross-domain training data Models and data Mission control · Safety · Software Deploy · Telemetry Fleet infrastructure The Serve autonomy stack One autonomy stack. Many environments. The Platform
Gen2 Gen3 Top Speed 7 mph 11 mph Weather 32104°F / Light rain 4113°F / Heavy rain Range 23 mi 10 hrs 48 mi 14 hrs Cargo 13 gal / 4 14" pizzas 15 gal / 4 16" pizzas Unit Cost Baseline 65% reduction Full-stack AV sensors Level 4 autonomy All-day battery Redundant connectivity NVIDIA Jetson Orin Third generation. Purpose-built for the sidewalk. The Serve robot
Purpose-built for the hospital. The Moxi robot Hardware & operating profile Compute NVIDIA RTX A2000 10× prior gen) Battery 70% capacity, 18h runtime Sensors 3D lidar, cameras Manipulation Mobile arm, drawer cameras Operating environment Multi-level hospitals, clinics, labs Diligent Robotics was acquired by Serve Robotics in early 2026. Moxi 2.0 Capabilities Retrieving and delivering lab specimens Inpatient pharmacy optimization Telemetry box distribution Meds-to-Beds programs Multi-floor navigation: open doors and manually operate elevators Cabinet & drawer ops with mobile arm
14Proprietary and Confidential First Gen3 robots roll off Magna assembly line Oct 2024 Fleet quadrupled. Miami, Dallas, Atlanta launched. H1 2025 Dec 2025 Fleet quadrupled again. 2,000th robot deployed. Largest sidewalk fleet in the U.S. 2026 Optimize utilization. Expand geographies. Invest to grow Serve + Moxi fleet. 100 2,000 Serve robots in twelve months Manufacturing Automotive-grade production with Magna International. Industrialized supply chain. Proven ability to execute against aggressive deployment targets.
● Los Angeles ● Miami ● Dallas ● Atlanta ● Chicago ● Ft. Lauderdale ● Alexandria Live: 2026 Growth drivers New metros Selected by density, order volume, and regulatory readiness Deeper penetration More neighborhoods within existing markets Platform pull Uber Eats & DoorDash demand drives expansion priority Fleet Expansion Capital efficient investment to scale the robot fleet City-by-city. Neighborhood-by-neighborhood. National network Our playbook Launch in high-density neighborhood, prove unit economics, then expand outward. Platform demand pulls us into each new market.
CITIES LIST City by city. Neighborhood by neighborhood. ● New York ● San Jose ● Vancouver ● Sydney 2026: ● Seattle ● Philadelphia ● Boston ● Melbourne ● Perth ● Adelaide ● Tokyo ● Toronto ● Manchester ● London 2027+: ● Osaka ● Taipei ● Madrid Cities listed represent potential expansion markets and constitute forward looking statements subject to change based on business, regulatory, and operational considerations. Across the globe. Expansion
17Proprietary and Confidential From proof of concept to revenue inflection. Financial path 2,000-robot fleet creates the foundation for utilization, monetization, and platform expansion. FY 2024 FY 2025 $1.8M Early fleet. R&D phase. $2.7M Guided $2.5M. Fleet → 2,000. FY 2026 Guidance $26M 10x growth. Utilization ramp. At scale Physical AI Platform Multi domain robots Shared autonomy stack Continuous model improvement with real world data
18Proprietary and Confidential 18 Diversification across geography, environment and customer Fleet Revenue Growth Engine Diversified revenue streams across Serve Revenue Diversification Food delivery: Partnerships with UberEats and DoorDash Healthcare: Long-term fixed monthly + usage contracts at hospitals Branding: Triple-digit growth percentage QoQ in Q1 FY2026 Kitchen: Vebu acquisition closed in Q1 High margin and increasingly recurring and durable Software Revenue High Margin SaaS Over 45% recurring driven by migration to subscription based contracts Autonomy platform: Licensing model that will continue to scale in 2026 As the fleet scales, the mix is shifting toward durable, recurring, and high-margin revenue.
19Proprietary and Confidential Built by people who've shipped robots, platforms, and products. Team Ali Kashani Co-founder, CEO VP at Postmates. Ph.D. Robotics UBC. 15 patents. Touraj Parang President & COO VP Corp Dev at GoDaddy. Yale Law & Stanford. MJ Burk Chun Co-founder & VP Product + Design Director, Postmates. 17+ yrs in robotics & marketplaces. Brian Read Chief Financial Officer Controller, Apptronik. Public finance at REE & Coherent. CPA. Dmitry Demeshchuk Co-founder & VP of Software Staff engineer & founding engineer at Postmates X. Anthony Armenta Chief Software & Data Officer CTO at BrightDrop GM, VP at Postmates (acq. Uber), Anki, Dell, Wyse (acq. Dell), BS in CS & Math UC Davis) Rajesh Radhakrishnan VP of Autonomy Director at Ghost Autonomy; Head of ML at John Deere. Founding engineer at Blue River (acq. John Deere) MS in Computer Science UT Arlington) Euan Abraham Chief Hardware & Manufacturing Officer SVP Hardware at Latch. VP Hardware at GoPro. Lead engineer at Apple. BS in Engineering U of Sheffield)
20Proprietary and Confidential A category- defining 2025 → Largest scaled commercial fleet on sidewalks and hospitals in the U.S. → 20X robot fleet size to 2,000 deployed robots → Cover 80% of U.S. food delivery demand with UberEats and DoorDash → Scaled to 20 cities and 6 metro areas from LA to D.C. corridor → Maintained 99.8% delivery completion rate and exemplary safety record → Expansion to 4500+ merchants and coverage for 3.75M consumers → Business expansion into healthcare, software, and data monetization
21Proprietary and Confidential Q1 FY2026 Financial Highlights $3.0M Q1 revenue exceeds plan 578% $197M cash & marketable securities as of 31 Mar 2026 Results Outlook YOY revenue growth $26M 2026 revenue guidance $25M 2026 capex $160$170MNon-GAAP operating expense reflecting continued investment in autonomy development, fleet scale, and platform capabilities across both delivery and healthcare robotics. Business expansion into healthcare, software, and data monetization
Slide 16 Header: The Long-Term Vision Unbundling the car for dense cities Specialized autonomous robots for everyday urban tasks. ● Safer streets ● Lower emissions ● More efficient local commerce Design direction: Strong closing image. Minimal text. One idea only. The Operating Layer of Physical AI Nasdaq: SERV | serverobotics.com | investor.relations@serverobotics.com The Long-Term Vision
23Proprietary and Confidential Appendix Projected 2026 Operating Expenses (USD$ millions) GAAP Operating Expenses $ 197 212 Stock Based Compensation 37 42 Non-GAAP Operating Expenses $ 160 170 Reconciliation of GAAP Operating Expenses to Non-GAAP Operating Expenses