Amazon is in early discussions to invest as much as $10 billion in OpenAI, highlighting how investors — including the world’s largest technology companies — continue to commit extraordinary sums to the sector, even as profitability remains elusive for most AI ventures.

According to CNBC, the talks could result in OpenAI relying more heavily on Amazon’s AI chips, deepening ties between the ChatGPT maker and the e-commerce and cloud computing giant. If completed, the deal would value OpenAI at more than $500 billion, Bloomberg reported, citing a person familiar with the matter.

The Information first reported that Amazon was exploring the investment.

The sheer scale of the potential valuation illustrates how strongly investors are betting on AI’s long-term promise, even as near-term financial returns remain uncertain. Generative AI products have yet to demonstrate a clear, durable path to profitability, largely because of the enormous costs associated with training and running large models. Data centers, specialized chips, energy consumption, and talent acquisition continue to consume vast amounts of capital, often far outpacing revenues.

Still, that has not slowed the flow of money.

Amazon’s interest in OpenAI comes as the company looks to broaden its exposure in the AI race. It has already invested about $8 billion in Anthropic, positioning itself as a major backer of OpenAI’s rival while using those ties to promote its cloud services and proprietary hardware. Earlier this month, Amazon unveiled the latest version of its Trainium AI chips and disclosed plans for the next generation, part of a broader effort to reduce reliance on Nvidia and strengthen Amazon Web Services as an end-to-end AI platform.

A deal with OpenAI would be strategically significant for Amazon. Securing OpenAI as a customer for its AI chips would validate years of internal chip development and potentially drive massive workloads onto AWS. It would also place Amazon more squarely at the center of the AI ecosystem, alongside Microsoft and Google, both of which have tightly integrated AI models into their cloud and consumer offerings.

However, the appeal of new capital is straightforward for OpenAI. The company recently completed its transition to a for-profit structure, a move that allows it to raise money more freely and reduce its dependence on Microsoft, which owns roughly 27% of the firm. That shift has come as OpenAI’s spending commitments have surged. Training increasingly capable models requires long-term contracts for compute and infrastructure that run into the tens of billions of dollars.

The broader AI landscape is now defined by what many investors describe as circular deals. Major cloud providers and chipmakers invest in AI startups, which then commit to using those same companies’ data centers and hardware.

In March, OpenAI invested $350 million in CoreWeave, which used the funds to buy Nvidia chips that ultimately power OpenAI’s own workloads. In October, OpenAI took a 10% stake in AMD and agreed to use its AI GPUs, while also signing a chip usage deal with Broadcom. In November, OpenAI struck a $38 billion cloud computing agreement with Amazon, even before any equity investment had been finalized.

These arrangements are seen as reflections of a shared calculation across the industry: securing access to compute capacity now is more important than worrying about margins later. For many backers, the risk of missing out on the next foundational AI platform outweighs concerns about current losses.

That mindset helps explain why capital continues to pour into AI despite growing unease in parts of the market. Some analysts have warned that spending on AI infrastructure is racing ahead of demand, raising fears of overcapacity. Others point to the lack of proven business models beyond enterprise subscriptions and experimental consumer products.

Yet valuations keep climbing, and funding rounds keep getting larger.

The logic is defensive as much as opportunistic for Big Tech. Companies like Amazon, Microsoft, Google, and Meta are under pressure to ensure that AI does not erode their core businesses. Investing heavily — even at uncertain returns — is seen as a way to stay relevant, shape standards, and lock in strategic partners.

The talks between Amazon and OpenAI remain preliminary and may not result in a deal. But they capture the moment the AI industry is in: a phase defined less by profits and more by scale, positioning, and the belief that whoever controls the infrastructure and relationships today will dominate the economics tomorrow, whenever they finally arrive.

A California administrative law judge has issued a landmark ruling against Tesla, finding that the electric vehicle giant engaged in deceptive marketing practices regarding its Autopilot and Full Self-Driving (FSD) features.

The decision, handed down by Judge Juliet Cox, marks a critical turning point in a long-standing legal battle initiated by the California Department of Motor Vehicles (DMV) in 2022. The court found that Tesla’s branding and descriptions created a “false impression” that its vehicles were capable of fully autonomous operation when, in reality, they remain Level 2 driver-assistance systems requiring constant human supervision.

The Penalty: A ‘Stayed’ 30-Day License Suspension

The judge endorsed the DMV’s request for a 30-day suspension of Tesla’s sales and manufacturing licenses in California as a penalty for misleading consumers. However, in a move to balance regulatory enforcement with economic stability, the DMV has stayed the order, granting Tesla a 60-day window (with some sources citing up to 90 days for certain appeals) to bring its marketing into compliance.

Under the terms of the ruling, Tesla must either rebrand its software—specifically the “Autopilot” and “Full Self-Driving” names—or demonstrate that its vehicles have achieved the technical level of autonomy the names imply. While the manufacturing license suspension at the Fremont factory has been indefinitely stayed to prevent massive disruption to the regional economy, the threat to Tesla’s dealer license remains active. If Tesla fails to remove the allegedly deceptive language or rename the features within the compliance window, the 30-day sales halt will be enforced.

Tesla has responded with public defiance, asserting on social media that “sales in California will continue uninterrupted.” The company’s primary defense rests on the claim that the order is a “consumer protection” overreach, noting that the DMV failed to produce a single customer who testified to being personally deceived or harmed by the branding. Tesla’s legal team argued that the terms are protected commercial speech and that the company provides sufficient disclaimers informing drivers they must remain attentive.

Judge Cox rejected this argument, stating that the DMV’s authority to regulate advertising is preventative and does not require evidence of a specific victim. The ruling emphasized that names like “Full Self-Driving” are “unambiguously false” in a legal and technical context because they imply the driver’s undivided attention is not required. The judge noted that without the threat of suspension, there was no reason to believe Tesla would voluntarily alter its misleading representations.

Economic Stakes and the California Market

A suspension of sales in California would be a devastating blow to Tesla’s domestic operations. California represents roughly one-third of Tesla’s total U.S. sales, with nearly 135,500 vehicles registered in the state during the first nine months of 2025 alone. Furthermore, the Fremont factory is not merely a regional hub; it is the sole production site for the Model S and Model X, and it remains responsible for all Model 3 sedans destined for the North American market.

The timing of this ruling is particularly sensitive as Tesla faces a “widening gap” between its marketing hype and actual deployment. While the company has recently added the qualifier “(Supervised)” to its FSD suite to appease regulators, the court found this insufficient to rectify the broader perception created by the “Autopilot” name.

This California ruling adds to a growing mountain of legal pressure on Tesla. The company remains under investigation by the Department of Justice (DOJ), the Securities and Exchange Commission (SEC), and the California Attorney General over similar allegations of misleading autonomy claims. These federal probes are examining whether Tesla’s marketing constituted wire or securities fraud.

Paradoxically, as California tightens the reins on consumer vehicle marketing, Tesla is pushing forward with its unsupervised Robotaxi trials in Austin, Texas. On December 15, 2025, Tesla officially removed human safety monitors from its small test fleet in Austin, allowing cars to operate entirely empty. However, CEO Elon Musk has clarified that these Robotaxis utilize a fundamentally different version of driving software than the FSD system sold to everyday consumers, further highlighting the distinction between Tesla’s future autonomous goals and its current commercial products.

Google on Wednesday unveiled Gemini 3 Flash, a new model designed to be faster, cheaper, and more widely deployed than its predecessors, as the company moves aggressively to counter OpenAI and tighten its grip on the fast-moving artificial intelligence market.

The launch underscores how the contest for AI leadership is increasingly narrowing into a direct rivalry between Google and OpenAI, with the outcome likely to shape not only the future of AI tools but also productivity, competition, and economic power across industries.

Gemini 3 Flash brings the reasoning capabilities of Gemini 3 Pro into a lighter, more efficient model that Google says can be run at lower cost and higher speed. The strategy is: rather than reserving its most capable models for premium users or narrow enterprise use cases, Google is pushing advanced AI deeper into its mass-market products.

“This is about bringing the strength and the foundation of Gemini 3 to everyone,” Tulsee Doshi, senior director and Gemini product lead, said in an interview with Axios, framing the release as an access play as much as a technical upgrade.

That philosophy is reflected in how quickly Google is embedding the model across its ecosystem. As of Wednesday, Gemini 3 Flash becomes the default model in the Gemini app, replacing Gemini 2.5 Flash for everyday use. It is also now the default model powering AI Mode in Google Search, instantly exposing hundreds of millions of users worldwide to the new system.

Major enterprise software firms, including Salesforce, Workday, and Figma, are already using Gemini 3 Flash, signaling early traction with business customers.

The timing of the release is no accident. Google’s move comes less than a week after OpenAI launched GPT-5.2 and just a day after OpenAI rolled out ChatGPT Images, highlighting the increasingly compressed release cycles at the top of the AI market. Each major update now appears calibrated not only around technical readiness but also around competitive signaling.

More efficient models like Gemini 3 Flash matter because they lower the barrier to using advanced machine learning. Faster inference and lower compute costs make it easier for small businesses, developers, and everyday consumers to rely on AI tools for practical tasks, from planning trips and summarizing information to learning complex concepts. Google says Gemini 3 Flash is particularly strong in such everyday reasoning scenarios, while also supporting multimodal inputs, allowing users to combine text, images, video, and audio in a single workflow.

One notable claim from Google is that Gemini 3 Flash outperforms Gemini 3 Pro on SWE-bench Verified, a widely watched benchmark for evaluating coding agents. If that performance holds up in real-world use, it strengthens Google’s hand with business clients, a segment where Anthropic’s Claude has gained momentum and where OpenAI has been racing to deepen its enterprise appeal.

The broader context is a rivalry that has become increasingly binary. Google and OpenAI now dominate mindshare, investment, and usage at the cutting edge of large language models. OpenAI retains the advantage of being first to capture public attention with ChatGPT, but Google’s strength lies in distribution. By embedding Gemini directly into Search and core productivity tools, Google can scale adoption in ways few rivals can match.

That ubiquity appears to be translating into growth. According to data cited by The Information, Gemini’s share of weekly mobile app downloads, monthly active users, and global website visits has been rising faster than ChatGPT’s in recent periods. The figures suggest Google is beginning to close the gap by leveraging its existing user base rather than relying solely on standalone AI products.

Still, the race remains unforgiving. The rapid cadence of releases reflects how quickly leadership can shift at the frontier of AI development. Neither Google nor OpenAI is far enough ahead to relax, and both continue to face pressure from rivals such as Anthropic, Meta, xAI, DeepSeek, and a growing field of well-funded startups.

The next test is a scale for Google as pushing Gemini 3 Flash across search and consumer apps raises questions about consistency, reliability, and accuracy when used by vast and diverse audiences. Maintaining performance while operating at that scale will be critical if Google wants to convert distribution into lasting dominance.

However, the launch of Gemini 3 Flash signals a strategic bet that in an AI market increasingly defined by two giants, speed, cost efficiency, and reach may matter just as much as raw model power.

As 2025 draws to a close and the promises of 2026 come into view, a critical question confronts every enterprise: how must we redesign our businesses to win? At Tekedia Institute, we have long emphasized that business models are supreme. If the underlying logic through which a firm captures value, for fixing frictions in the market, is broken, no amount of effort, talent, or technology will change the outcome.

A robust business model is paramount for a company’s success, even more so than factors like strong leadership or execution alone. The business model, encompassing how a company creates, delivers, and captures value, is considered “supreme” because it dictates the fundamental logic and operations of the business.

Essentially, even with the same products or services, the business model adopted can drastically impact a company’s performance. Freemium or subscription business model on the same products? Whatever you decide will re-align how factors of production within that firm will be used.

In this AI era, the central issue is no longer whether you are using AI operationally in your firm, but how AI is reshaping your business model. Yes, besides running with AI, is AI transforming the enterprise, moving from artificial intelligence to enterprise intelligence?

Many business models that worked well in the past are now stale. Companies that once hired thousands of young people in Asia, Africa, and Latin America to provide entry-level engineering services to firms in Europe and America are now under severe pressure. Many have folded; others are shadows of their former selves. The reason is simple and uncomfortable: AI has disintermediated the entry-level roles that sustained those business playbooks.

When AI arrived, we paid attention to the signals. At Tekedia Institute, we recognized early that change was not incremental; it was structural. As supply of courses became abundant, value shifted away from merely offering courses toward improving learner outcomes. In a world where AI can generate unlimited content, the advantage no longer lies in static online materials. It lies in helping learners make sense of abundance, think into knowledge, and acquire actional insights.

That realization forced a redesign. We pivoted toward more live, interactive programs, focusing on guidance, interpretation, and execution rather than content alone. The business model evolved because the environment changed. That is why Tekedia AI Lab program is a live program.

The question now is this: what signals are you seeing in your own business? As 2026 arrives, are you rethinking the logic of how you create and capture value or are you hoping yesterday’s model will survive tomorrow’s realities? Rethink your business model for 2026.