Home Blog Page 23

OpenAI Plans to Bring Dots AI Agents to Wider Consumer Base as It Tests $500 Premium Tier

By
Samuel Nwite
-
0

OpenAI plans to eventually make its new Dots AI agent system available across its consumer base, Chief Financial Officer Sarah Friar said, signaling that the company sees autonomous AI agents becoming a mainstream feature rather than remaining a premium product for heavy users and businesses.

Dots will initially be restricted to OpenAI’s higher-priced Pro, Business and Enterprise offerings, Friar said in a CNBC interview Tuesday. The company, however, intends to extend access to its broader consumer audience, which OpenAI says now numbers about 1.2 billion people globally.

“For the business, you’re going to see it show up inside our pro SKUs and then inside of our business enterprise offerings for now,” Friar said. “But absolutely our vision is to bring this to our whole consumer base as well, 1.2 billion people around the world.”

The comments more clearly indicate how OpenAI is approaching the economics of agentic AI. Rather than immediately putting an autonomous system into the hands of its entire user base, the company is starting with customers paying substantially higher subscription fees, where greater usage and computing costs can be absorbed more easily.

CEO Sam Altman made a similar point at OpenAI’s DevDay on Tuesday, describing Dots as a premium product initially but saying the company ultimately expects to offer a mass-market version.

“I think there are very good reasons for that, and I think it will help us understand and help people understand how expensive what AI can do,” Altman said.

“But you should, of course, expect us to do a mass market thing for billions of people,” he added.

Dots represents a shift in how OpenAI wants people to use ChatGPT. Instead of waiting for users to issue prompts, the system is designed to work across the applications people use and take action on their behalf.

Friar described it as “productivity” that can proactively remind users about tasks and interact with services such as email, calendars, and Slack.

“Dots is now productivity,” Friar said. “It’s the agent that can tap you on the shoulder and remind you that you forgot to do that LinkedIn post for tomorrow. So it can go to your email, it can go to your calendar, it can go to your Slack or wherever you work.”

The design is believed to align with OpenAI’s business model. A conventional chatbot largely consumes computing resources when users actively interact with it. An agent that continuously monitors information, reasons about tasks and takes actions can require substantially more compute even when the user is not directly interacting with it.

Starting Dots within premium subscriptions therefore gives OpenAI a controlled environment to measure usage, compute requirements and the willingness of customers to pay for the additional capability.

Neither Friar nor Altman provided a timetable for when Dots will reach the broader consumer market or what it would cost.

$500 Subscription Puts A Price On Intensive AI Use

OpenAI’s premium strategy is already expanding beyond its traditional $20-a-month Plus subscription. The company announced Pro 500, a $500 monthly ChatGPT plan that costs $6,000 annually and provides the highest usage allowance. The plan includes what OpenAI describes as ultrafast computing for GPT-6 Astra in ChatGPT Work and Codex, with speeds up to eight times faster.

OpenAI has also reopened its $200 monthly Pro subscription while reducing its computing allowance. Subscribers now receive 10 times the usage allowance of the $20-a-month Plus plan, compared with 20 times previously.

The pricing structure points to a differentiated ChatGPT market, with OpenAI charging users according to how intensively they consume its computing resources.

Friar argued that there is evidence customers are becoming more comfortable with higher-priced AI subscriptions.

“A year ago, two years ago, when we launched our $200 SKU, people thought we’d lost our minds. No one’s ever going to spend that per month. I heard that over and over again,” Friar said.

“And now, actually, what we see is not only are people willing to spend more, they’re actually moving more towards consumption in consumers.”

The challenge for OpenAI is that the economics of agents could be very different from those of conventional chatbot subscriptions. An agent that searches across multiple applications, continuously processes information, and executes multi-step tasks could generate substantially more inference demand per user.

That makes the current premium-first strategy commercially significant. OpenAI can use its highest-paying customers to establish how much people value autonomous assistance before deciding how much of the capability can be offered at lower prices.

The Mass-Market Ambition Comes With An Infrastructure Question

OpenAI’s stated goal of reaching billions of consumers creates a much larger economic challenge. The company says its consumer audience already stands at 1.2 billion people. Even a relatively small percentage of those users adopting autonomous agents could produce a significant increase in computing demand if each agent performs substantially more work than a conventional chatbot interaction.

The $500 Pro 500 subscription effectively establishes a high-end tier for customers willing to pay for that computing intensity. The company’s eventual mass-market offering will require OpenAI to determine which agent capabilities can be delivered economically at lower prices.

That could result in a tiered system in which simple agent functions become widely available while more computationally intensive tasks remain subject to usage limits or higher subscription fees. It also creates a potentially important distinction between AI as software and AI as an ongoing service. Traditional software can be sold to millions of customers with relatively low marginal costs once it has been developed. Autonomous AI agents require continuing expenditure on computing every time they reason, retrieve information, or execute a task.

OpenAI’s pricing strategy is seen as an indication that it is increasingly trying to make that variable cost visible in the subscription structure. The company is believed to be pursuing two objectives simultaneously: expanding AI agents into everyday consumer workflows and ensuring that the economics of that expansion can support the computing required to operate them.

Dots’ initial restriction to premium and business customers gives OpenAI an opportunity to test that equation before taking the technology to its entire consumer base. The long-term ambition, however, is clear from Altman’s comments. OpenAI does not intend for autonomous agents to remain a niche feature for companies and high-paying users. It wants them to become part of the mainstream ChatGPT experience.

Nvidia-Branded Trailers Stolen With 20,000 Pounds of Sand Inside

By
Paul Ugbede Godwin
-
0

The theft of Nvidia-branded trailers in California has produced one of the strangest stories to emerge from the booming artificial intelligence industry. Thieves reportedly stole trailers displaying Nvidia branding, apparently expecting valuable technology or equipment.

Only to discover that the cargo inside was something considerably less glamorous: thousands of pounds of sand. The incident highlights both the growing value associated with Nvidia’s name and the unusual methods technology companies use to test the next generation of artificial intelligence and autonomous vehicles.

The trailers were associated with PlusAI, an autonomous-trucking company working on self-driving technology. Nvidia’s technology and branding are closely connected with the broader AI infrastructure ecosystem.

Where powerful processors, servers and computing systems have become some of the most valuable components in the technology supply chain. That association may have made the trailers appear particularly attractive to thieves.

Instead of expensive Nvidia GPUs or artificial intelligence servers, the trailers reportedly contained approximately 20,000 pounds of sand. The material was being used as simulated cargo for testing autonomous trucks.

By placing heavy loads inside a trailer, engineers can reproduce the weight and physical characteristics of a commercial shipment while evaluating how an autonomous vehicle responds to different driving conditions.

The stolen trailers were reportedly taken overnight from outside a PlusAI facility in Newark, California. Their Nvidia branding may have created the impression that valuable technology was being transported inside. But the apparent prize turned out to be an industrial testing material.

The trailers were eventually recovered after someone familiar with the company recognized them near another business location. The recovery meant that PlusAI avoided the potentially significant financial damage that could have resulted if expensive autonomous-driving equipment had actually been stolen.

The story points toward a broader security issue surrounding the artificial intelligence boom. As AI companies expand their data centers and autonomous-driving programs, the physical infrastructure supporting these technologies has become increasingly valuable.

Nvidia GPUs, networking equipment, servers and specialized computing systems can cost millions of dollars, creating new incentives for organized theft and supply-chain crime.

Nvidia has become one of the most recognizable companies in the AI economy. Its chips power a significant portion of the computing infrastructure used to train and operate advanced AI models.

As a result, the Nvidia name can itself become a signal of potential value, even when a shipment has little or nothing to do with high-end processors.

The California trailer incident therefore carries an ironic lesson for companies operating in the AI supply chain. Security cannot simply focus on the contents of a shipment. Branding, transportation routes, warehouse locations and publicly visible logistics can also reveal information that criminals may use to identify potential targets.

The incident ended with the trailers recovered and the sand apparently still intact. For the thieves, what may have looked like a valuable Nvidia shipment turned into a remarkably heavy haul of sand.

The episode is funny on the surface, but the underlying trend is serious. AI is creating a new ecosystem of valuable physical assets, from GPUs and data-center hardware to autonomous vehicles and specialized testing equipment.

As that ecosystem grows, companies will increasingly have to protect not only their digital systems but also the physical supply chains that make the AI economy possible.

The Nvidia trailer theft may have involved sand rather than semiconductors, but it reflects a larger reality: in the AI era, even an empty-looking trailer carrying a famous technology company’s logo can become a target.

Solana PAID Token Drops 38% as X Money Blocks Fee Payouts, While BitMine Surpasses 6M ETH Holdings

By
Paul Ugbede Godwin
-
0

Two developments highlight different sides of the crypto market: the fragility of new payment-linked token models on Solana and the growing scale of institutional Ethereum accumulation. They show how quickly liquidity, infrastructure and corporate treasury strategies can reshape digital-asset markets.

UsePaid, a Solana-based protocol that routes creator fees from token launches to X accounts through X Money, has been forced to change its payout system after a sharp increase in activity overwhelmed its payment process.

The platform reported $1.54 million in fees claimed over 24 hours, roughly 26 times the amount processed the previous day. It initially imposed a $750 daily payout cap before suspending X Money payments.

The disruption exposed a key weakness in the model: although the underlying fees are generated on-chain, the original payout mechanism depended on an external payment rail.

UsePaid subsequently introduced a web-based claims portal allowing eligible recipients to withdraw qualifying fees directly to Solana wallets, generally in SOL. Historical X Money balances are being handled separately.

The market reaction was immediate. The PAID token fell sharply, with reports putting the decline at about 38% around the initial disruption, while subsequent market data showed an even larger 47% 24-hour decline at one point.

The move demonstrated how closely the token had become associated with the protocol’s ability to process creator-fee flows. The episode also illustrates the risks of building financial infrastructure across multiple layers.

Solana can settle the underlying transactions rapidly, but converting those revenues into conventional payments introduces another dependency. When transaction activity accelerates faster than the payout infrastructure can absorb it, an on-chain business can still experience an off-chain bottleneck.

At the same time, BitMine Immersion Technologies is pursuing the opposite strategy: concentrating increasingly large amounts of capital into Ethereum. The company purchased another 17,362 ETH during the week ending September 27, taking its holdings to 6,001,302 ETH.

That represents approximately 4.9% of Ethereum’s estimated 122.1 million circulating supply. BitMine’s position was valued at roughly $16.2 billion using its stated ETH reference price of $2,698.

Across crypto, cash, marketable securities and other investments, the company reported approximately $17.2 billion. Its portfolio also included 213 BTC, $672 million in cash and marketable securities, and equity positions in Beast Industries and Eightco Holdings.

More importantly, about 5.07 million of BitMine’s ETH was staked, representing roughly 84% of its Ethereum holdings. The company estimated annualized staking revenue at approximately $358 million, although actual returns can vary with network conditions, validator performance and other factors.

BitMine began its Ethereum treasury strategy in June 2025 and says it has purchased ETH every week since then. Its stated objective is to reach ownership equivalent to 5% of Ethereum’s supply.

Putting the company close to a milestone that would give one public corporation an unusually large economic position in a major blockchain network. The two stories capture an important tension in crypto. Smaller protocols are experimenting with new bridges between social platforms, token launches and payments.

Where infrastructure failures can rapidly translate into token volatility. Meanwhile, larger companies are treating established networks such as Ethereum as strategic treasury assets, accumulating billions of dollars in exposure.

The contrast is about scale and infrastructure. One model is testing how crypto payments can become embedded into internet identities; the other is testing how far corporate balance sheets can integrate blockchain assets.

Both demonstrate that in crypto, technological design and financial structure increasingly move markets together.

Instinct’s $1 Billion Bet Signals a New Phase for Consumer AI Agents

By
Paul Ugbede Godwin
-
0

The artificial intelligence industry is entering a phase in which the most valuable systems may not simply answer questions but act on behalf of their users.

Instinct, a young AI startup developing a personal agent capable of performing everyday tasks autonomously, has underscored that shift by raising $1 billion in Series C funding at a $10 billion valuation.

The new valuation is four times the $2.5 billion level assigned to the company only about a month earlier. The financing, backed by Sequoia Capital, Benchmark and Coatue, is striking not only because of its size but also because of the speed at which investors have repriced the company.

Instinct launched its invite-only service in August, yet it has already attracted enormous attention from venture capital investors searching for the next major consumer AI platform. At the center of the company’s proposition is a simple change in how people interact with software.

Instead of asking an AI chatbot for information and then completing the work themselves, users can ask Instinct to execute the task. The agent can plan trips, order groceries, make reservations, purchase tickets, pay bills and cancel subscriptions.

It can communicate through text or phone calls while using its own computer and phone systems to complete actions. That distinction is important for the emerging agentic-AI economy. Traditional generative AI largely transformed search, writing, coding and information retrieval.

AI agents attempt to transform execution. The economic opportunity therefore extends beyond software subscriptions toward transactions that currently pass through travel companies, retailers, restaurants, service providers and other intermediaries.

Instinct has introduced a concierge capability designed to handle tasks requiring phone conversations, including appointments with businesses that lack online booking systems. Its trusted-person network allows different Instinct agents to coordinate with one another.

Suggesting a future in which software agents could negotiate and organize activities between people without requiring every participant to interact directly. The enormous valuation comes while the company remains in early access.

Instinct has not publicly disclosed comprehensive user or revenue figures, making the $10 billion valuation less a reflection of established financial performance than an expression of investor expectations about the potential size of personal AI.

Reuters reported that the company is expanding its technology while emphasizing privacy protections such as isolated sandboxes and short-lived credentials. Privacy is particularly important because an agent capable of acting autonomously requires considerably more access than a conventional chatbot.

To book a flight, manage subscriptions or make purchases, an AI may need access to communications, accounts, payment information and personal preferences. Instinct has faced questions around the amount of information users must provide, highlighting the tension between convenience and control.

Competition is also intensifying. Meta’s Muse is pursuing a similar consumer-agent opportunity while benefiting from integration across Meta’s enormous ecosystem. That creates a difficult strategic environment for startups: they must build superior agents while competing against technology companies with vast computing resources, distribution networks and existing consumer relationships.

Instinct’s $1 billion raise therefore represents more than another spectacular AI funding round. It reflects investor conviction that autonomous software could become a new layer of consumer computing. Whether that conviction ultimately translates into sustainable revenue will depend on reliability, privacy, transaction economics and user trust.

For now, Instinct’s rapid rise demonstrates how aggressively capital is moving toward an AI future where software does not merely respond to people—it acts for them.

PwC Survey Reveals 4 Types of Workers as OpenAI Scraps GPT-6.1 Astra Over Safety Concerns

By
Paul Ugbede Godwin
-
0

Artificial intelligence is no longer simply changing how people work; it is beginning to separate workers into distinct economic categories. PwC’s 2026 Global Workforce Hopes and Fears Survey.

Based on responses from 49,364 workers across 48 countries and regions, describes a workforce increasingly divided by access to AI, the value of employees’ skills and their ability to adapt to technological change.

At almost the same moment, OpenAI’s decision to scrap the planned launch of GPT-6.1 Astra after safety tests raised concerns offers a reminder that the AI revolution is advancing faster than the systems designed to control it.

PwC identifies four groups. The largest, representing 56% of workers, is the “engine room”: employees whose skills are less scarce and who have not yet gained significant advantage from AI.

They perform much of the everyday work that keeps organisations operating, yet only about two in five say they have access to the learning and development resources they need. Another 18% are “AI insurgents,” workers who may not possess scarce skills but are using AI to increase what they can accomplish.

Then come the “front-runners,” representing 14% of workers. They combine scarce, highly demanded skills with strong AI capabilities. PwC says three-quarters of this group are optimistic about their organisation’s future.

While 85% say they are ready to adapt to new ways of working. Finally, 11% are “indispensables”: workers with valuable, scarce expertise who have not progressed as far along the AI learning curve.

The significance is not merely technological. It is economic. AI adoption can increase the productivity and bargaining power of workers who know how to use it, while employees without access to training risk being trapped in roles where automation gradually reduces the value of their existing skills.

PwC’s broader 2026 AI Jobs Barometer reinforces this shift: job postings requiring AI skills are growing substantially faster than the overall jobs market, while workers with AI skills command a significant wage premium.

But the other headline complicates the narrative that faster AI development is automatically better. OpenAI has decided not to release GPT-6.1 Astra, originally expected in October, after internal evaluations found problems involving safety and alignment.

The company said the model did not consistently remain within user-authorised boundaries or accurately communicate what it had done. The concerns are particularly important because Astra had already reached what OpenAI describes as a “Critical” level of cybersecurity capability.

The company’s own earlier safety assessment said the model could, with appropriate tools and access, identify previously unknown vulnerabilities and develop exploitation methods without step-by-step human guidance.

The two developments reveal the central tension of the AI economy. Inside workplaces, companies are trying to move quickly enough to capture productivity gains and prevent employees from falling behind.

Inside AI laboratories, developers are discovering that greater capability can also create new forms of risk that cannot simply be solved by releasing the technology faster.

The future workplace may therefore depend on two kinds of adaptation: workers learning to collaborate effectively with increasingly capable systems, and AI developers learning when a model is not ready to be deployed.

The dividing line may be less about humans versus machines than about who can adapt, who receives the tools and training to adapt, and whether increasingly powerful systems can remain within the boundaries humans set for them.

1...222324...9,437Page 23 of 9,437