OpenAI acknowledged the high cloud computing consumption of its Dots personal assistants, announcing the introduction of the economic GPT-6.1 Terra model alongside GPT-6 Astra to lower operational costs.
OpenAI’s acknowledgment of heavy compute consumption in Dots agents and the GrokBot issue
Introduction of the GPT-6.1 Terra model and automatic router for the $100 tier
New funding round challenges and target reduction from $30 billion to $20 billion
Reliance on Nvidia, development of the Jalapeno chip for 2027, and competition with Anthropic
Analytical reading of the strategic and economic dimensions of the event
Frequently asked questions
OpenAI’s acknowledgment of heavy compute consumption in Dots agents and the GrokBot issue
OpenAI acknowledged that its new personal assistants, known as “Dots,” consume a significant amount of computing power, prompting the company to begin routing part of these assistants’ work to a smaller, less expensive model. This explicit admission casts further doubt on the tech industry’s financial ability to sustain the operating costs of AI agents that run continuously. During a dinner with journalists in New York on Monday, November 2, 2026, Adam Fry, head of consumer for ChatGPT, stated that Dots agents fall “on the heavy side in terms of compute consumption,” explaining that each assistant runs its own cloud computer to execute tasks for users via their connected accounts.
Competitors face the exact same dilemma; Elon Musk stated this week that the cost of running “GrokBot,” a personal assistant launched by SpaceX in August, is “unsustainably high relative to revenue,” hinting at scaling back its allocated computing power so users hit their permitted usage limits faster. The industry’s rush toward these agents followed the massive success of Meta’s “Muse” assistant, according to Fortune magazine.
Introduction of the GPT-6.1 Terra model and automatic router for the $100 tier
OpenAI announced on Tuesday, November 3, the expansion of the family of models powering Dots agents to include the “GPT-6.1 Terra” model. Each “Dot” agent can now use different models for different parts of its work, as the Terra model is characterized by its smaller size and lower computing power requirements compared to the massive “GPT-6 Astra” model that originally powered every Dots task. The new automatic router aims to make the service less compute-intensive.
Currently, Dots agents are available exclusively to subscribers of the $100-per-month Pro tier. Fry explained that opening the service to a larger number of users depends on achieving a technical breakthrough in computing efficiency expected soon to lower costs, expressing his desire to turn Dots into a permanent presence through features like “background mode” which will require greater compute.
New funding round challenges and target reduction from $30 billion to $20 billion
Financial liquidity represents the second constraint for the company. The Wall Street Journal reported that OpenAI secured commitments of only about $11 billion in its latest funding round, falling short of its $30 billion goal, at a valuation of $1.25 trillion after previously hoping for a $1.4 trillion valuation.
The Financial Times reported that OpenAI trimmed its funding goal to $20 billion, with plans for a subsequent offering directed at wealthy individuals to bridge the funding gap, though the company has not publicly confirmed these figures.
Reliance on Nvidia, development of the Jalapeno chip for 2027, and competition with Anthropic
OpenAI relies for most of its operations on Nvidia processors and plans to scale up volume production of its own in-house chip named “Jalapeno” in 2027, according to Fortune magazine.
In contrast, its rival Anthropic, which uses 4 types of chips, has signed purchase agreements worth $895.4 billion “mostly tied to compute and infrastructure” according to its amended prospectus, yet lists the inability to secure more compute as a major risk. Fortune summarized the predicament by asking: “The more powerful and continuously working personal AI agents become, the more compute they consume; will OpenAI ever be able to catch up with itself?”
Analytical reading of the strategic and economic dimensions of the event
OpenAI’s decision to introduce the smaller, less expensive “GPT-6.1 Terra” model alongside the “GPT-6 Astra” model to power its “Dots” personal assistants via an automatic router reveals the core economic dilemma facing the new generation of persistent AI agents. When each personal assistant operates inside an independent cloud computer to continuously execute tasks across user accounts, inference computing costs multiply at rates that make offering the service outside the $100-per-month Pro tier extremely difficult—the same challenge that led Elon Musk to describe GrokBot’s costs as unsustainable relative to revenue.
This operational pressure on compute consumption intersects with a tighter funding environment; reports showed OpenAI raising only about $11 billion in commitments for its latest round at a $1.25 trillion valuation while scaling back its funding target from $30 billion to $20 billion, at a time when it plans to expand production of its proprietary “Jalapeno” chip in 2027 and rival Anthropic is committed to $895.4 billion in compute contracts. Therefore, the success of automatic routers and economic models in lowering the operating cost of persistent agents will be the primary prerequisite for transforming this technology into a profitable and sustainable mass product.
Frequently asked questions
Question: Why did OpenAI add the GPT-6.1 Terra model to Dots assistants?
Answer: Because each Dot assistant runs its own cloud computer and consumes a significant amount of compute, the new automatic router shifts simpler tasks to the smaller, cheaper Terra model instead of running everything on the GPT-6 Astra model.
Question: How much does the subscription cost to use Dots agents currently?
Answer: The Dots service is currently available exclusively to Pro tier subscribers for $100 per month, and expanding it to general users depends on lowering computing costs.
Question: What are the developments regarding OpenAI’s current funding round?
Answer: Reports indicated that the company raised commitments of about $11 billion at a $1.25 trillion valuation, and lowered its round target from $30 billion to $20 billion with a subsequent offering for the wealthy.
Question: When does OpenAI plan to expand production of its proprietary chip and what is its name?
Answer: The company plans to expand production of its in-house chip named Jalapeno in 2027 to reduce exclusive reliance on Nvidia processors.