Sam Altman has publicly addressed his fractured relationship with Elon Musk for the first time in a major interview, revealing deep disappointment with how his former collaborator has evolved. In an exclusive conversation with The Times' tech podcast ahead of the Paris AI Action Summit in February 2025, the OpenAI CEO also explained why the company abandoned its ambitious AI-native browser project, offering rare insight into the economics of building standalone AI products.

What Changed Between Sam Altman and Elon Musk?

The relationship between Altman and Musk has deteriorated significantly since OpenAI's founding. When asked directly about his feelings toward Musk, Altman's response was candid and restrained. "No. I think he has really changed. I regret it," Altman said when asked if he misses his former friend. The two co-founders no longer stay in touch, and Musk is currently suing OpenAI in what Altman describes as a lawsuit with multiple versions.

Despite the rift, Altman maintained a measure of respect for Musk's accomplishments. "I still see him as a person who has accomplished incredible feats, a great motivator," he acknowledged. However, the warmth that once defined their partnership has clearly evaporated, replaced by professional distance and personal disappointment.

Altman

Why Did OpenAI Shut Down Its ChatGPT Atlas Browser?

On August 9, 2026, OpenAI officially decommissioned ChatGPT Atlas, its AI-native browser, less than 10 months after launch. The product, which CEO Sam Altman had described as a "once-in-a-decade opportunity" when it debuted in October 2025, never materialized into a sustainable business. The macOS version was the only version ever released; a long-rumored Windows version never launched.

The economics of building a standalone AI browser proved unsustainable. Building an AI-native browser requires significantly higher investment than simply adding AI features to an existing browser. According to technical experts cited in the reporting, maintaining a Chromium-based browser with a complete native interface, Agent runtime, memory system, background services, and security infrastructure requires engineering investment that is 3 to 10 times higher than creating an AI plugin for Chrome.

Atlas faced additional challenges beyond engineering costs. The Agent Mode feature, which allowed AI to directly operate users' web pages and accounts, created substantial security risks. Hackers could embed malicious text in web pages that the AI might interpret as user instructions, making defense "very difficult." In real-world testing, Atlas struggled with practical performance: it took approximately 10 minutes to add three items to an Amazon shopping cart, slower than manual operation.

How to Understand OpenAI's Product Strategy Shift

• Feature Integration: Rather than abandoning the technology entirely, OpenAI is integrating Atlas's core capabilities directly into ChatGPT, including cross-site multi-step task execution, browsing memory, and in-depth operations on logged-in websites.
• Revenue Model Reality: Atlas required a ChatGPT Plus or Pro subscription to unlock Agent Mode features, meaning the browser generated no independent revenue. Its value was measured only by traffic diverted to ChatGPT and user data collected.
• Market Consolidation: The shutdown signals a broader industry trend where mature browsers like Google Chrome and Microsoft Edge are embedding AI directly, while independent AI-native browsers shrink to enterprise-only scenarios.
• Cost-Benefit Analysis: OpenAI determined that the incremental gross profit from each Atlas user, minus browser maintenance and security costs, did not justify continued standalone development.

This decision mirrors OpenAI's earlier shutdown of Sora, its AI video application, within a year of launch. Both products demonstrated that even well-resourced companies must make difficult choices about which standalone consumer products can sustain the investment required to compete.

Altman's Vision for AI's Future and Global Competition

Beyond the browser shutdown and personal rifts, Altman outlined an ambitious vision for artificial intelligence's trajectory. He predicted that within ten years, "everyone on Earth will be able to accomplish more things than the most influential people today". This optimism is grounded in observable progress: OpenAI's "Deep Research" feature, launched just a week before his interview, enables users to complete work that previously took days or weeks in approximately 20 minutes.

Altman

Regarding the rise of Chinese AI competitor DeepSeek, which shocked markets earlier in 2026 with a powerful, low-cost model, Altman struck a measured tone. "These things never surprise me at an abstract level," he said, "we know that sooner or later more powerful competitors will emerge. But you never know which morning you will wake up and find that it has happened". He acknowledged DeepSeek's achievements in chain-of-thought reasoning and free-tier availability, but characterized the advancement as "not a major update for us" at the research level.

Altman emphasized a macro trend in AI economics: after reaching a certain level of intelligence, prices drop by approximately 10 times annually, totaling more than 100-fold reductions over recent years. This pace of cost reduction far exceeds Moore's Law, the historical benchmark for semiconductor advancement.

On geopolitics, Altman expressed hope for international cooperation rather than an "arms race" framework. "I would be very happy if the West and China could reach some consensus on AI standards," he stated, including "common safety standards, common benefit frameworks". However, he acknowledged fundamental disagreements about whether AI should move toward more authoritarian or democratic systems.

What Does Altman Say About Working With the Trump Administration?

Altman has become more publicly aligned with the Trump administration, attending the announcement of Project Stargate, a $500 billion large-scale training and inference system led by SoftBank. He explained that the Biden administration made infrastructure development "impossible," whereas President Trump has "completely different views on construction, permitting, power supply, and American manufacturing".

Altman denied that Silicon Valley companies are currying favor through political calculation. Instead, he framed the shift as a genuine opportunity: "People will say, 'Hey, maybe we can build a lot of things in the US now.' Not just software, maybe we can get semiconductor manufacturing, robot factories, and new energy back on track". He emphasized that this represents an "atmosphere shift" rather than opportunistic alignment.

Regarding European AI regulation, Altman acknowledged the trade-offs inherent in strict oversight. "We will abide by the rules, it is up to the European people and leaders to decide what they want. But whatever decision is made, there will be consequences," he explained. If Europe requires three to five months of testing before new model deployment while OpenAI launches new models every three months, the company will "never catch up, never be able to deploy, and always have outdated products."

Altman

The broader picture emerging from Altman's statements is one of pragmatic realism: AI development requires massive infrastructure investment, competitive pressure from global players is accelerating, and the companies building these systems must navigate complex political and regulatory landscapes while making difficult product decisions based on economic sustainability.