Anthropic promotes its upcoming Mythos AI model as a game-changer, but Google is shifting the focus to cost and efficiency. Google claims its new Gemini 3.5 Flash model is just as powerful as other leading models but more budget-friendly. CEO Sundar Pichai recently remarked, “Companies are already blowing through their annual token budgets, and it’s only May.” He believes that combining Flash with other models could save companies significant money.
The rise in AI usage is making expenses a hot topic. As firms increasingly rely on token-heavy AI, they are also keeping a closer eye on their budgets. Smaller companies are raising prices, leading customers to rethink their spending.
This creates an opening for Google to compete on value rather than just power. Google has been refining its technology for over 25 years, giving it a distinct advantage.
For the first few years, the AI race was all about having the largest, smartest model. Now, as companies narrow the performance gap, the emphasis is shifting to infrastructure and how models are executed. Greg Brockman, President of OpenAI, noted, “The model alone is no longer the product.”
As AI becomes more complex, the costs to operate these systems soar. Monthly use of Google’s AI has jumped sevenfold, reaching 3.2 quadrillion tokens since last year. If top customers shifted to using Gemini 3.5 and other models, they could save over $1 billion annually, according to Pichai.
Businesses are feeling the pinch. Uber’s COO pointed out the challenge of justifying rising AI costs, while venture capitalist Chamath Palihapitiya mentioned that his firm, 8090, is moving away from a costly AI tool due to expenses. Analyst Dan Morgan noted that as AI agents grow more complex, organizations are hit with sticker shock.
Cost and return on investment (ROI) go hand in hand. For some companies, being at the cutting edge may no longer be necessary; sometimes “good enough” will do. Google stands well-positioned in this environment because it maintains tighter control over AI costs and performance. The company has its own TPU chips and relies directly on manufacturers, potentially saving it 50% to 75% on internal computing costs.
In contrast, OpenAI incurs costs from using cloud services that rely on other companies’ infrastructure, often leading to higher expenses.
Historically, Google has built its success on speed and cost-effectiveness. In the mid-2000s, Google Search took off not just for delivering good results but for also being faster and cheaper. The company opted for custom systems using affordable parts, creating a successful feedback loop that improved its offerings.
Now, Google is applying a similar strategy with its Gemini model. Their successful search advertising business can help fund AI ventures, unlike competitors racing to secure more funding and computing power.
If computing power shapes the future, Google is well-prepared thanks to years of investment in infrastructure. In a landscape where AI is becoming essential, its ability to provide efficient and cost-effective solutions could keep it at the forefront of the technology race.
Source link
google,company,rival,flash model,cost,year,speed,money,hand,late gemini,token,infrastructure,ai,product,google cloud
