Tesla scraps Project Dojo: What it means for full self-driving and future of AI
Tesla has been focused on building driverless cars for almost a decade now. So why did Tesla end its attempts to produce chips in-house for its autonomous vehicles?
The decision came because Tesla CEO Elon Musk shifted to relying on Nvidia and AMD for compute chips, along with Samsung for manufacturing. According to Musk, Tesla was closing Project Dojo because it would be inefficient for the business to split its resources and scale two distinct AI chip designs. Dojo would not be included in Tesla's consumer goods because it was intended to train the company's autonomous driving programme.
Dojo was the automaker's effort to build custom supercomputers designed to train Tesla's neural network for driverless cars. A neural network functions like a simulated brain, supporting Tesla's goal of launching Robotaxi and Full Self-Driving (FSD) cars. This network is installed across a large fleet of Tesla cars today, enabling some automated driving functions that still require a human to be attentive behind the wheel.
Tesla's decision to shut down Dojo, which Musk had been talking about since 2019, marks a major strategic shift. Musk had previously described Dojo as the cornerstone of Tesla's AI ambitions and its push for full self-driving, thanks to its ability to 'process truly vast amounts of video data.' He even mentioned it briefly during the company's second-quarter earnings call.
The primary reason is Tesla's vision-only philosophy. For FSD to work, its neural networks must identify and categorise objects in the environment, then make driving decisions in real time. This requires training on enormous volumes of driving data so that, when activated, FSD can continuously gather and process visual data at speeds comparable to human depth and motion detection. In essence, Tesla wants to build a digital model of how the human brain and visual cortex function.
To achieve this, Tesla must store, process, and run millions of simulations on video data gathered from vehicles worldwide. Although Tesla depended on Nvidia to run its Dojo training computer, the aim was to reduce reliance on a single supplier and improve performance by boosting bandwidth and lowering latency.
Tesla's Dojo system was intended to serve as an AI training platform for FSD. A supercomputer consists of thousands of smaller computers, or nodes. Each node has a central processing unit (CPU) and a graphics processing unit (GPU). While the CPU manages the node overall, the GPU handles more complex tasks, such as splitting workloads into many parts and processing them in parallel. GPUs are critical for machine learning tasks like FSD training in simulation.
Musk's broader vision is for Tesla to become an AI company capable of making self-driving cars by mimicking human vision. Most other companies in the autonomous driving sector use a combination of sensors—lidar, radar, cameras—and high-definition maps. Tesla believes it can achieve full autonomy with cameras alone, paired with advanced neural networks to analyse data and make instant driving decisions.
In August 2024, Musk began promoting Cortex, Tesla's new AI training supercluster being built at its Austin headquarters to tackle real-world AI challenges. This shifted focus away from Dojo.
Bloomberg reports that Tesla will now rely more on Nvidia and other partners such as AMD for computing and Samsung for chip fabrication. On 28 July 2025, Tesla and Samsung signed a $16.5 billion agreement to produce AI6 inference processors for high-performance AI training, FSD, and Tesla's Optimus humanoid robots.
Meanwhile, Dojo's lead, Peter Bannon, is leaving the company, and the remaining team members will be reassigned to other data centre and compute projects within Tesla.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
40 minutes ago
- Time of India
Tesla shuts down its AI training supercomputer Dojo, loses 20 key execs to company founded by Ganesh Venkataramanan
Tesla is reportedly dissolving its Dojo supercomputer team, abandoning its in-house chip development for autonomous driving technology. According to a report by Bloomberg, - Tesla Inc. is disbanding its Dojo team and its leader will leave the company, according to people familiar with the matter, upending the automaker's effort to build an in-house supercomputer for developing driverless-vehicle technology. The Dojo shutdown coincides with Tesla's board offering Elon Musk a $29 billion compensation package to prioritize Tesla's AI goals over his other ventures, including xAI, a pure-play AI company. Meet DensityAI, AI startup founded by former Dojo head Ganesh Venkataramanan The move marks a significant pivot for the automaker, which has increasingly leaned on external partners like Nvidia, AMD, and Samsung for its AI and compute needs. Peter Bannon, the lead of the Dojo project, is leaving Tesla, and the remaining team members will reportedly be reassigned to other data center and compute initiatives within the company, the report said citing anonymous sources. The decision follows the exit of approximately 20 Dojo team members who left to form DensityAI, a new AI startup founded by former Dojo head Ganesh Venkataramanan , alongside ex-Tesla employees Bill Chang and Ben Floering. DensityAI is preparing to exit stealth mode and is focused on developing chips, hardware, and software to power AI-driven data centers for robotics, AI agents, and automotive applications. The startup aims to address high-performance computing needs with an emphasis on scalable, energy-efficient solutions for next-generation AI workloads. As Tesla recalibrates, DensityAI's emergence could reshape the AI hardware landscape, leveraging the expertise of former Dojo engineers to challenge established players in the AI data center space. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like 75% off installation and replacement for LeafFilter LeafFilter Gutter Protection Get Rates Undo What makes shutting down of Dojo important The disbanding of Dojo comes at a pivotal moment for Tesla. CEO Elon Musk has been repositioning Tesla as an AI and robotics company, despite challenges with its limited robotaxi launch in Austin in June 2025. The launch, which involved Model Y vehicles with a human co-pilot, faced criticism after reports of erratic driving behavior. Dojo, first announced in 2019, was pitched as the backbone of Tesla's AI ambitions, particularly for achieving full self-driving (FSD) by processing vast amounts of video data. Musk highlighted Dojo as recently as Tesla's Q2 2025 earnings call, though his focus shifted to Cortex, a new AI training supercluster at Tesla's Austin headquarters. Dojo combined a supercomputer with Tesla's in-house D1 chip, unveiled at AI Day 2021. Presented by Venkataramanan, the D1 was designed to work alongside Nvidia GPUs to power Dojo, with a next-gen D2 chip planned to address data flow bottlenecks. In 2023, Morgan Stanley estimated Dojo could add $500 billion to Tesla's market value by enabling robotaxi services and software revenue. However, by August 2024, Musk's rhetoric pivoted to Cortex, signaling a strategic shift. Tesla is looking externally Tesla's reliance on external partners is growing. A $16.5 billion deal with Samsung, signed last month, will produce Tesla's AI6 inference chips, designed to scale from FSD and Optimus humanoid robots to high-performance AI training. Tesla is also deepening ties with Nvidia for compute power and AMD for additional chip solutions. During the Q2 earnings call, Musk hinted at streamlining efforts, suggesting convergence between Dojo 3 and the AI6 chip to avoid redundancies. AI Masterclass for Students. Upskill Young Ones Today!– Join Now
The Hindu
an hour ago
- The Hindu
The difficult path for Trump's ‘one big budget bet'
The most debated government downsizing reform that has been implemented in recent history has been the Department of Government Efficiency (DOGE) initiative under the Donald Trump administration in its second term. The initiative aims to reduce U.S. federal spending, deficit, debt and interest burden, streamline government operations, and enhance government efficiency. The initiative was much in the news with the appointment of Elon Musk, to advise and guide the government on the initiative, and his declaration that the overall goal is to achieve a leaner government by cutting the deficit and reducing the number of federal agencies from over 400 to 99. Root cause of deficit and debt The need for a smaller government was felt by the Trump administration due to enormous government spending and deficits. The size of the U.S. government, measured by public expenditure as a percentage of GDP, on average, was 36.49% from 2001 to 2024. Surprisingly, this was the smallest among the seven major advanced economies (MAE). France (56.53%) had the largest government, followed by Italy (49.81%), Germany (46.64%), Canada (41.24%), the United Kingdom (41.09%), and Japan (37.56%). The size of the government in the U.S. has remained the smallest among the MAE for most of the last 25 years. Yet, overall fiscal balance and government debt have been higher in the U.S. when compared to the MAE. In the triennium ending (TE) 2024, the U.S. had a fiscal balance and debt burden of (-)6.0% and 119.5% of GDP, respectively, compared to the figures of (-)4.1% and 108.6%, respectively, for the MAE. The underlying reason is the lower revenue collection in the U.S. The total revenue of the U.S. government as a percentage of GDP on an average was the lowest (30.55%) during the period 2001–2022 compared to France (51.48%), Italy (45.93%), Germany (44.93%), Canada (39.92%), and the U.K. (36.63%). Between 2001 and 2022, the U.S. has clocked the lowest tax-GDP ratio of 19.27% as against 28.59% for Italy, 28.12% for France, 27.99% for Canada, 26.89% for the U.K., 22.70% for Germany, and 32.74% for OECD countries. Initiatives and outcomes The several expenditure reforms the DOGE has implemented are the termination of unused federal government office space leases, cancellation of wasteful contracts, recovery of misallocated funds, federal workforce optimisation involving initiatives such as hiring restrictions, workforce reduction, and offering voluntary buyouts, deregulatory measures aimed at ensuring lawful governance and reducing red tape. There is also Artificial Intelligence-based monitoring of federal employee activity, productivity assessment, and detection of inefficiencies, identifying and consolidating overlapping functions across departments, and cutting down overseas humanitarian and development spending. DOGE follows a transparent communication strategy by publicly sharing updates (on X or DOGE's portal) on the outcomes. The total estimated expenditure savings from contract/lease cancellations and renegotiations, grant cancellations, fraud and improper payment detection and elimination, asset sales, employee reductions and regulatory changes have been $190 billion, which amounts to $1,180 per taxpayer. DOGE has made the distribution of government grants transparent through a payments portal, which allows citizens to see recipient-wise payments of federal grants. DOGE's Workforce Portal provides information on the agency-wise size of the U.S. government civilian workforce. To streamline layoffs, it has implemented the 'Workforce Reshaping Tool', a modernised version of the Pentagon's Automated Reduction in Force (AutoRIF) software, which assists in terminating federal employees based on criteria such as seniority and performance. So far, approximately 2,60,000 federal employees have been laid off, retired early, or accepted buyouts. Several federal regulations were repealed and modified, with a cost savings of $30.1 billion due to reduced regulatory compliance; also, 1.8 million words have been deleted from the rules of several Federal regulations. DOGE publishes a unique unconstitutionality index (UI) measuring the extent of bureaucracy's role in shaping federal policy. As per UI, for every law passed by Congress in 2024, there were about 19 rules created by the bureaucracy. The path ahead However, DOGE's journey has seen an unexpected turn after Elon Musk publicly disagreed with U.S. President Donald Trump over the provision incorporated in the much-debated One Big Beautiful Bill (OBBB) for removing tax credits offered to purchase electric vehicles. The OBBB is a legislative extension of the DOGE reform, as it aims to embed multiple fiscal priorities and DOGE-style government efficiency reforms into a single federal law. Ironically, by opposing OBBB, Mr. Musk contradicted his initiatives as an adviser to DOGE. However, the bigger challenge facing the DOGE initiative is that its extended form, the OBBB, is unlikely to resolve the U.S.'s deficit and debt problem, as the tax cuts proposed in the OBBB far exceed the spending cuts, leading to an addition of $3.2 trillion to the U.S. national debt over the next decade. Currently, the U.S. has the lowest and below-OECD-average corporate tax rates among the MAE, lower effective tax rates on the rich, and tighter secrecy laws that enable tax evasion by wealthy individuals and corporations. Therefore, unless efforts are made to boost government revenues, Mr. Trump's 'One Big Budget Bet' of reducing America's deficit and debt through DOGE-style expenditure reforms is unlikely to pay off. Sthanu R Nair teaches Economics and Public Policy at the Indian Institute of Management Kozhikode, Kerala. Pratik Sinha is a consultant for a multinational firm. He was an Assistant Director, Nehru Yuva Kendra Sangathan, Ministry of Youth Affairs and Sports, Government of India
&w=3840&q=100)
Business Standard
2 hours ago
- Business Standard
China urges US to ease AI chip export controls for trade deal: Report
China wants the United States to ease export controls on chips critical for artificial intelligence as part of a trade deal before a possible summit between Presidents Donald Trump and Xi Jinping, the Financial Times reported on Sunday. Chinese officials have told experts in Washington that Beijing wants the Trump administration to relax export restrictions on high-bandwidth memory chips, the newspaper reported, citing unnamed people familiar with the matter. The White House, State Department and China's foreign ministry did not immediately respond to requests for comment on the report. HBM chips, which help perform data-intensive AI tasks quickly, are closely watched by investors due to their use alongside AI graphic processors, particularly Nvidia's. The FT said China is concerned because the U.S. HBM controls hamper the ability of Chinese companies such as Huawei to develop their own AI chips. Successive U.S. administrations have curbed exports of advanced chips to China, looking to stymie Beijing's AI and defence development. While this has impacted U.S. firms' ability to fully address booming demand from China, one of the world's largest semiconductor markets, it still remains an important revenue driver for American chipmakers.
