Article Complete
AIM
Elon Musk Says He is ‘Optimistic’ About Releasing a GPT-6-Level Model
Elon Musk Says He is ‘Optimistic’ About Releasing a GPT-6-Level Model

Elon Musk Says He is ‘Optimistic’ About Releasing a GPT-6-Level Model

“We will keep accelerating,” he said.

Published Updated

Key takeaways

  • Elon Musk expresses cautious optimism about releasing a competitive AI model within three months.
  • SpaceXAI's AI development is relatively new, yet Musk believes they can achieve significant progress quickly.
  • Musk addresses performance criticisms, suggesting excessive intelligence in AI for certain tasks is unnecessary.
  • The company faces challenges with hardware and compute but has shown strong capabilities in scaling resources.
  • SpaceXAI's compute strategy focuses on expanding its supercomputer infrastructure significantly over time.
5 min

Elon Musk, CEO of SpaceXAI, shared on X that he is “cautiously optimistic’ that in the next three months, the company will release an AI model that will perform on par with Anthropic’s Fable or OpenAI’s GPT-6. 

“We will keep accelerating,” said Musk. The post comes shortly after he agreed with Anthropic CEO on slowing down the development of frontier AI models. 

Musk also pointed out that SpaceXAI’s AI efforts are only three years old, compared with how much longer Anthropic and OpenAI have been in the race. “If our second derivative remains strong, SpaceX will reach pole position in about 6 months.”

He also addressed certain criticism of the performance of the company’s Grok 4.7 models. He stated that once AI models exceed a certain calibre of intelligence for a class of tasks, additional intelligence is pointless. “You don’t need (and it would be cruel to put) Newton-level intelligence in your toaster,” he said. 

Currently, Grok 4.7 trails GPT-6 Astra and Anthropic’s Fable models on benchmark scores.

Musk also acknowledged challenges with hardware and compute, and said that SpaceXAI “has demonstrated exceptional ability” in bringing massive amounts of compute online, and will only get better. 

SpaceXAI’s compute strategy is centred on Colossus, its large-scale AI supercomputer in Memphis, which the company says has already reached 200,000 GPUs and is part of a longer-term roadmap to 1 million GPUs. 

The company built the original system in about four months and doubled its capacity to 200,000 GPUs in a further 92 days, according to SpaceXAI.

The scale of that infrastructure is also becoming a commercial asset. 

SpaceX’s IPO disclosures revealed that the company signed an agreement with Anthropic in May 2026 to provide access to compute capacity across Colossus and Colossus II for $1.25 billion per month through May 2029. 

The filing said SpaceXAI can use the infrastructure for its own models while allocating capacity to third parties.

Founded by Musk in 2023 as an independent AI startup, xAI was acquired by Musk’s SpaceX in February in an all-stock deal. 

Musk subsequently said xAI would be dissolved as a separate company and its AI products would operate under the SpaceXAI name. 

The rebrand was completed publicly in July, when xAI’s account changed to SpaceXAI.

Membership

Stay ahead with AI insights.

Unlimited access to premium articles, exclusive analysis, and the stories shaping India's AI economy.

Subscribe now

AIM Americas

1190 Miraloma Way Suite P, Sunnyvale, CA 94085

AIM India

1st Floor, Sakti Statesman, Bellandur, Bengaluru – 560103

© 2026 AIM. All rights reserved.