SpaceXAI cuts 60 staff to pivot from image generation to coding agents
The company is scaling back its Grok Imagine team to focus on high-margin coding tools and AI agents.
The short version
- SpaceXAI cut roughly 60 people from the Grok Imagine team over the past week.
- Management is shifting resources toward AI coding tools and an agent product built on the Grok chatbot.
- Grok Imagine features remain live but lack a clear roadmap or timeline for future updates.
- The move follows a trend of reducing generalist staff to focus on high-payoff areas like specialized tutors.
Why it matters
This shift signals a move toward B2B software where companies pay per seat, rather than consumer-facing media generation which is costly to run. For users, it means Grok Imagine may enter a state of 'quiet decay' with fewer updates.
A pivot toward profitability
SpaceXAI has reportedly laid off about 60 employees who worked on Grok Imagine, the company's image and video generation tool. According to The Information, these cuts occurred over the past week and leave only a small team to maintain the existing product. This move marks a significant shift in how the company allocates its human and technical resources.
The decision appears to be driven by the high costs associated with training and serving visual media models. Management is reportedly moving away from these expensive features to focus on AI coding tools and an agent product. By prioritizing these areas, the company aims to target markets with higher profit margins and more consistent revenue streams.
While the Grok Imagine features are still available to users, there is currently no updated roadmap or timeline provided for the product. The reduction in staff suggests that the team is being moved into a maintenance phase rather than an active development cycle. This allows the company to redirect its focus toward its new primary objectives.
The economics of AI compute
Training and serving video generation models requires a massive amount of GPU time, which is a significant expense for any AI lab. The people paying for these features are often consumers on free or low-cost tiers, making the business model difficult to sustain. In contrast, coding tools and agents typically sell to companies that pay per seat or per token, providing a more stable financial return.
SpaceXAI is navigating a landscape where infrastructure is becoming a primary business focus. The company has massive compute resources, but recent deals with Anthropic and Cursor suggest that infrastructure may be a more viable path than building frontier models. By moving staff to coding tools, the company is swapping one type of competition for another with a bigger paying market.
The shift also reflects a broader trend of moving away from generalist roles toward specialized tasks. In September 2025, xAI let go of roughly 500 data annotators on the Grok team to hire specialist tutors instead. This pattern shows a deliberate strategy to spend money only where the leadership believes the payoff will be highest.
Losing the race in video
The company may also be choosing to exit the video generation race to avoid losing it publicly. Chinese companies like ByteDance have been setting the pace with models like Seedance, while Google and OpenAI continue to advance their own offerings. Grok Imagine was attempting to catch up with a smaller team and less patience than its primary rivals.
By scaling back the team, SpaceXAI can avoid the public scrutiny of falling behind in a highly competitive field. It is often easier for a company to step back from a race than to struggle and lose against well-funded competitors. This strategic retreat allows the company to consolidate its efforts on products where it can maintain a more distinct advantage.
The history of the Grok Imagine team suggests it has already faced significant challenges. More than 50 researchers and engineers have left the group since SpaceX took over xAI in February 2026. The departure of key leaders like Haotian Liu and Guodong Zhang further indicates that the original vision for the image tool is being phased out.
What this means for users
For people who currently use Grok Imagine, the most immediate risk is what could be described as quiet decay. This means the product may see slower bug fixes, stale models, and a lack of new features over time. Since no new roadmap has been shared, the product may simply exist in its current state without significant evolution.
nFor investors and observers, the move highlights the difficulty of building frontier models with finite resources. SpaceXAI is attempting to find a balance between high-profile innovation and sustainable business growth. By focusing on coding agents, they are entering a crowded field occupied by established players like Anthropic and OpenAI.
The success of this pivot will depend on how quickly the company can deploy effective tools to engineering teams. The next few product releases will serve as the first real evidence of whether this strategy works. For now, the company is betting that the enterprise market for coding will be more lucrative than the consumer market for video.
What to watch
- How quickly SpaceXAI can launch its new agent product compared to Anthropic's Claude Code.
- Whether Grok Imagine receives any major updates or remains in a maintenance state.
Try it
Look at the current features of Grok Imagine to see if the tool still meets your needs before the team moves into a maintenance phase.
Sources and links
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by Startup Fortune · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
