DomainHuntdomainhuntAdvertise

SpaceX Super Intelligence Cuts About 60 Grok Imagine Jobs as Musk Shifts AI Priorities

SpaceX Super Intelligence has reportedly eliminated roughly 60 positions in Grok Imagine's video-generation unit, signaling a shift toward coding tools and autonomous agents as expensive visual AI models face mounting computing costs.

CompanyBigGo Finance ·

SpaceX Super Intelligence has cut approximately 60 positions from its Grok Imagine video-generation operation, exposing a strategic shift toward coding software and autonomous AI agents. The reduction leaves a substantially smaller group maintaining the visual-generation product, according to reporting by The Information cited by BigGo Finance.

The decision offers a significant indication of where Elon Musk's combined aerospace and artificial intelligence business intends to concentrate its resources. Video generation demands substantial computing capacity, while coding assistants and autonomous agents offer a potentially more economical route to commercial AI applications. Grok Imagine has not disappeared, but the team supporting it has been dramatically reduced.

Musk changes AI priorities

The restructuring follows the consolidation of Musk's artificial intelligence businesses under the SpaceX Super Intelligence identity. SpaceX absorbed his AI company, xAI, through a $250 billion all-stock transaction in February. The resulting organization has since begun reorganizing its technical teams and directing investment toward different product categories.

One particularly visible priority is software development. SpaceX reportedly spent $60 billion in August acquiring Cursor, an artificial intelligence startup specializing in coding assistance. The company has also elevated Grok Bot, an autonomous agent product, within its development strategy. Such agents are designed to perform tasks with greater independence than conventional question-and-answer chatbots.

The contrast with Grok Imagine is significant. Generating video requires models to process and produce sequences of visual information rather than individual text responses. That makes the process considerably more demanding in terms of computing resources. Training models and generating videos for customers both require expensive infrastructure, creating pressure to justify continued investment through commercial returns.

Video models face pressure

SpaceX's retreat follows a similar decision at OpenAI, which discontinued its Sora video-generation product earlier this year after determining that its computing requirements had become too expensive. OpenAI subsequently redirected resources toward enterprise software and coding applications.

Musk had previously presented OpenAI's departure from video generation as an opportunity for his own business. In March, he used X to promote an upcoming Grok Imagine release, describing it as something that would be "epic" and indicating that the company intended to increase its commitment to the category.

The latest staffing reductions suggest that the economics of the business have become more influential than those earlier ambitions. The situation also changes the competitive landscape in the United States. With SpaceX stepping back, Google and Meta Platforms are identified in the report as the remaining major American technology companies training their own proprietary video-generation models.

Performance presents another difficulty. Independent evaluations referenced in the reporting placed Grok Imagine behind competing systems. That weakness makes it harder to justify maintaining a large engineering organization and substantial computing expenditure for a product that has yet to establish a stronger competitive position.

Inside xAI's wider rebuild

The cuts are part of a much broader transformation following the February acquisition. The original xAI founding group experienced a rapid breakdown after the transaction. By mid-May, every co-founder except Musk had reportedly left the company, according to The Information.

More than 50 researchers and engineers have departed since the acquisition, although that total includes different circumstances, from voluntary resignations to dismissals and layoffs. The departures have extended beyond the Imagine organization. Leaders associated with coding, world-model development and Grok Voice have also exited, while the central pretraining group, responsible for developing the foundational AI models, has reportedly undergone substantial reductions.

Musk has characterized the restructuring as a fundamental reconstruction rather than a limited management adjustment. In a March statement on X, he acknowledged that the original organization had been established incorrectly and compared the rebuilding effort with previous organizational changes at Tesla.

That approach reflects a management pattern visible across his businesses. However, the scale and consequences of each restructuring have varied, and the latest developments suggest that SpaceX Super Intelligence is applying a similar willingness to change staffing and organizational priorities to its AI operations.

Conflicting reports on cuts

The exact number of affected employees remains uncertain. The Information's account, as reported by BigGo Finance, puts the reduction in the Grok Imagine operation at approximately 60 positions. A separate report from Chinese technology publication 36kr described a different outcome, saying the multimodal organization had approximately 70 employees and that the entire unit had been eliminated.

The two accounts have not been reconciled. The difference matters because a reduction that leaves a small maintenance team would represent a different organizational decision from the complete dissolution of a department.

There is also evidence that the company has not entirely abandoned multimodal development. xAI has advertised a position for a Model Engineer, Multimodal Understanding in Palo Alto, with compensation listed between $160,000 and $250,000, plus equity.

The advertised position covers systems that can understand and generate images, video, audio and text. Its appearance alongside the reported cuts complicates the interpretation of the restructuring. Rather than demonstrating a complete withdrawal from visual AI, the changes could indicate a narrower hiring strategy focused on different technical capabilities.

Musk's familiar restructuring

The latest changes also resemble earlier decisions at Musk's other companies. In 2024, Tesla dismissed approximately 500 employees from its Supercharger organization following disagreements over additional staffing reductions. Some essential members subsequently returned through arrangements involving partners and suppliers, after the initial cuts had disrupted the organization.

Tesla also dismantled its Dojo supercomputer team in 2025, redirecting computing investment toward its AI6 chip and a more general-purpose architecture. Musk described the second generation of the Dojo project as an evolutionary dead end. The decision illustrated his willingness to abandon an established technical direction when he believes another approach deserves the resources.

His acquisition of Twitter provides another example. After completing the purchase, Musk reduced the company's workforce by approximately half within a month. Across these businesses, large organizational changes have repeatedly accompanied shifts in his strategic priorities.

At SpaceX Super Intelligence, the reorganization has also changed the reporting structure. Technical leaders received broader responsibilities earlier in the year, while managers recruited from SpaceX and Tesla were placed in direct reporting relationships with Musk.

Four competing AI tracks

The company is now organized around four principal development areas. These cover Grok's foundational models and voice capabilities, coding models and machine-learning infrastructure, Imagine's multimodal generation work, and Macrohard, an initiative focused on automating enterprise processes.

Each area received new leadership assignments earlier in the year. Some technical executives gained wider authority, giving the company a structure intended to coordinate different AI products and underlying technologies under stronger central management.

The cuts to Imagine raise questions about how those four tracks will compete for computing resources, engineering talent and management attention. Coding and enterprise automation appear to have stronger immediate backing, while the visual-generation operation faces the challenge of demonstrating that its costs can be justified.

Why it matters

For investors and competitors, the restructuring points to a more selective approach to artificial intelligence development. SpaceX Super Intelligence appears increasingly interested in applications where computing expenditure can be supported by identifiable commercial demand. Coding assistants and autonomous agents offer a different economic proposition from systems that generate lengthy visual sequences.

The shift could create opportunities for Chinese developers that continue treating video generation as a central business priority. ByteDance, Kuaishou and MiniMax are among the companies identified in the reporting as maintaining strong ambitions in visual AI. ByteDance's Seedance 2.5 is described as an industry benchmark, while MiniMax's open-source H3 model has attracted broad adoption, including use as the underlying architecture for American startups.

That contrast suggests the competitive center of gravity in generative video may become less American if domestic technology leaders continue reducing investment. However, the latest developments do not establish that SpaceX has permanently abandoned the category.

Grok Imagine remains available, and the advertised multimodal engineering position indicates continued interest in capabilities spanning several media formats. The decisive question is whether a smaller team can improve the product enough to justify further investment.

The next test for Musk's AI organization will be whether it can turn its emphasis on coding and autonomous agents into stronger products without surrendering the visual-generation market to competitors. The company's future hiring decisions, the distribution of computing resources across its four development tracks and any further changes to Grok Imagine will provide the clearest evidence of how permanent this strategic shift becomes.

Read the original at BigGo Finance

More news

Written by DomainHunt from reporting by BigGo Finance. Read the original article for the full story. Not financial advice.