AUSTIN, Texas — Tesla has announced the development of an upgraded self-driving computer called AI4 Plus, which will double the current system's RAM from 32 gigabytes to 64 gigabytes while delivering approximately 10 percent more compute and memory bandwidth. CEO Elon Musk revealed the plans during the Q1 2026 earnings call on April 22, framing the upgrade as a natural evolution of the company's AI4 hardware platform that will keep Tesla's fleet positioned for increasingly demanding FSD software.
The AI4 Plus — also described by Musk as "AI4.1" — is currently awaiting modifications from Samsung, which fabricates Tesla's AI chips on a 7-nanometer process. Production is targeted for 2027.
What Changes With AI4 Plus
The central improvement is memory capacity. Tesla's current AI4 computer uses two chips, each with 16 gigabytes of RAM, for a total of 32 gigabytes system-wide. AI4 Plus will bring each chip to 32 gigabytes, doubling total system memory to 64 gigabytes — matching the RAM available in NVIDIA's Orin platform used by competing autonomous vehicle programs.
Beyond raw capacity, AI4 Plus will provide roughly 10 percent improvements in both compute throughput and memory bandwidth. Memory bandwidth is the metric Musk has identified as the primary bottleneck for running large neural networks in real time — the same constraint he cited when confirming that HW3 cannot achieve unsupervised FSD.
Musk described the rationale clearly: "At some point the AI4 hardware is going to get so old that the only reason to keep the factory open is for AI4. We are planning an AI4 upgrade to use newer generation RAM."
Why AI5 Is Going to Optimus First
The AI4 Plus announcement came in the context of a broader explanation of Tesla's chip strategy. Tesla's next-generation AI5 chip — which will deliver dramatically higher compute — is being allocated first to Optimus robots and data centers rather than vehicles. The reason: Musk believes the current AI4 hardware is already sufficient for unsupervised Full Self-Driving in vehicles.
AI4 Plus therefore serves as a bridge. It extends the useful life of the AI4 platform with meaningful improvements while AI5 scales in Optimus production and data center applications. When AI5 eventually transitions into vehicles — potentially later in the decade — it will arrive on a platform already proven at scale.
A Platform Already Iterating
Notably, AI4 Plus will not be the first revision to Tesla's AI4 hardware. In January 2026, Tesla quietly began shipping a three-chip variant — internally called AI4.5 — in new Model Y units produced at its Fremont facility. That version was never formally announced but has been identified in vehicles delivered since then.
The progression is AI4 (original two-chip design) → AI4.5 (three-chip design, shipping since January 2026) → AI4 Plus (doubled RAM per chip, targeting 2027). Each iteration reflects the rapid pace at which Tesla is advancing its onboard AI infrastructure to stay ahead of growing software demands.
Implications for the Robotaxi Fleet
AI4 Plus matters most for Tesla's Robotaxi business. The Cybercab — Tesla's purpose-built two-passenger robotaxi — launched production at Giga Texas earlier this year and is currently operating in unsupervised commercial service across Austin, Houston, Dallas, and several additional markets. As Tesla scales that fleet and adds more cities, the software complexity of navigating dense urban environments will grow. More RAM and bandwidth directly translate into the ability to run larger, more capable neural networks without hardware constraints.
Tesla's plan to convert HW3 vehicles to AI4 through its urban micro-factory program will also benefit from the AI4 platform's continued improvement. Owners who receive the upgrade will be placed on a hardware roadmap with a clear path forward rather than a static endpoint.
Looking Ahead
With AI4.5 already in production, AI4 Plus on track for 2027, and AI5 in development for longer-term deployment, Tesla's hardware roadmap is more clearly defined than at any point in the company's autonomous vehicle history. The combination of expanding Robotaxi operations, the HW3 retrofit program, and a steady cadence of compute upgrades positions Tesla to maintain its lead as the largest deployer of autonomous miles in the world.
