RatingBuy CompanyTesla Inc. Company Update North AmericaUnited States ReutersTSLA.OQ ConsumerAutos & Auto Technology 1Q26 Preview Valuation & Risks Edison YuResearch Analyst+1-212-250-7263 Following 1Q deliveries, we provide our latest thoughts on the quarter, robotaxi,and Terafab. For the upcoming earnings, we expect key discussion topics to berobotaxi deployments, humanoid progress (Optimus V3), Cybercab production,and Terafab/capex. We trim our price target to $465 (vs. prior $480), reflecting someminor tweaks to our model/valuation framework; reiterate Buy. Winnie DongResearch Analyst+1-212-250-5121 Laura LiResearch Associate+1-212-250-2266 Soft quarter as expected Tesla’s 1Q26 volume performance reflects mostly expected headwinds andremains on track to achieve our full-year forecast. Deliveries of 358,023 vehicles fellslightly short of the company-compiled consensus (365,600). The deliberate wind-down of Model S+X and generally soft EV demand impacted deliveries globallyalthough European consumer sentiment appears to have bottomed out. While theenergy storage segment should continue to grow, 1Q deployment of energystorage productions was only 8.8 GWh, materially below expectations of 14.4GWh. This is likely due to lumpiness in fulfilling orders and should improvesequentially. Overall, we model 1Q revenue of $20.7bn, reflecting a sequentialdecline but up YoY. We expect auto GPM (ex-credit) of 14.7%, down from 17.9% in4Q25, driven by lower volume and the phasing out of pay-upfront option for FSD(effective as of 2/15). James Mulholland, CFAResearch Associate+1-212-250-6026 Robotaxi ride impression We took a ride in Austin last month and summarize our impression. nRoute: Downtown Austin to Austin BeerworksnDate: March 19nTime: 12:20 PM (solid traffic)nVehicle: Model Y (with Safety Monitor)nDuration: 40 minutes (Expected: 20 minutes)nFare: $17.35 Our ride included a safety monitor in the front passenger seat. The monitor (whoworks 8-hour shifts) didn't have to intervene once, but his presence was a reminderthat we are still in the "validation" phase of the robotaxi rollout despite reports onselectivemonitor out.From a technology perspective,the experience was 9 April 2026Autos & Auto TechnologyTesla Inc. impressive and seamless. In dense traffic, the vehicle was able to handle merges,initiate lane change signals, and navigate around a truck that changed lanes. Theonly frustration was perhaps the route logic. The car appeared to have opted for asignificant detour to avoid a certain route it isn’t yet authorized to use. While ahuman driver would have likely muscled through the traffic on the highway, thesystem prioritized more local routes, turning a 20-minute trip into nearly 40minutes. Ultimately, we think this issue will solve for itself once operatingparameters expand further. Looking to 1Q earnings, we expect Tesla to provide an update on expansions intoother cities (from Austin and SF currently) including those below. We will also belooking for updates on the production of CyberCab which should start productionthis month in Giga Texas. nTexas: Dallas and Houston (joining Austin).nFlorida: Miami, Orlando, and TampanArizona: Phoenix (positioning Tesla in direct competition with Waymo’shome turf)nNevada: Las Vegas Tesla also started to disclose active FSD subscriptions, exiting 4Q at 1.1m users(+38% YoY) globally with 70% being upfront purchases. Starting Feb 15, there willno longer be an option to pay upfront (i.e., only subscription at $99/month). Thismay put some pressure on margins near-term given upfront purchases are highlyaccretive but over time should level out. Terafab suggests more capex coming Long term, Tesla confirmed its intention to pursue Terafab, a very ambitiousendeavor for semiconductor vertical integration (logic, HBM4 memory, andpackaging) targeting 1 TW of capacity. Management believes current global chipproduction accounts for only about 2% of the future AI compute capacity needs.Through Terafab, CEO Elon Musk believes the JV can design, fabricate, and test newAI chips (like the upcoming AI5) at much faster speed. Applications for these chipswill span autonomous vehicles, Optimus, and space data centers (majority of futureallocation). The initial $20-25bn outlay will be spent on a phase 1 pilot facility at GigaTexas developing 2nm technology with target capacity of 100,000 wafer starts permonth. Also just this week, Intel announced it will be joining the Terafab effort,leveraging its expertise to “refactor” or redesign/optimize the silicon fabricationprocess. In practice, this seems to suggest Intel will be helping on the front end (forlogic), enabling Tesla to bring capabilities online faster compared to starting fromscratch. We think Tesla may next seek out a partner on the memory side. In terms of capex, Tesla already committed to +$20bn this year excluding Terafaband domestic solar cell production. While the economics of Terafab have not beenofficially d