--- title: "Chamath Palihapitiya Says Companies Are Overspending On AI As Cheaper Models Rapidly Close The Gap: 'Burning Through Massive Budgets'" type: "News" locale: "en" url: "https://longbridge.com/en/news/288955591.md" description: "Venture capitalist Chamath Palihapitiya warns that companies are overspending on premium AI models despite cheaper alternatives rapidly closing the performance gap. He highlights a massive pricing disparity, citing costs ranging from $105,000 for GPT-5.5 Pro to $2,740 for DeepSeek R1 for similar token volumes. Palihapitiya criticizes firms for 'burning through massive budgets' without governance and suggests adopting mixed-model strategies to optimize costs while reserving expensive systems for high-value tasks." datetime: "2026-06-07T07:23:05.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/288955591.md) - [en](https://longbridge.com/en/news/288955591.md) - [zh-HK](https://longbridge.com/zh-HK/news/288955591.md) --- # Chamath Palihapitiya Says Companies Are Overspending On AI As Cheaper Models Rapidly Close The Gap: 'Burning Through Massive Budgets' On Saturday, venture capitalist **Chamath Palihapitiya** argues that many companies are overspending on premium artificial intelligence (AI) models even as lower-cost alternatives rapidly close the performance gap. ## **AI Cost Gap Remains Wide** Palihapitiya highlighted what he described as one of the biggest surprises of 2026: the narrowing capability gap between leading open-weight AI models and top proprietary systems. In a post on **X**, he wrote, "The capability gap between the best open-weight/source models and the best closed models has narrowed much faster than the pricing gap." He added that while performance differences have become relatively small, "the pricing gap remains enormous." To illustrate the disparity, Palihapitiya compared estimated monthly costs for a company processing 1 billion input tokens and 1 billion output tokens. He estimated costs of roughly $105,000 for GPT-5.5 Pro, $30,000 for Claude Opus 4.8, $5,220 for DeepSeek V4 Pro and $2,740 for DeepSeek R1. Palihapitiya also shared a response generated by ChatGPT suggesting that businesses should adopt a mixed-model strategy, using lower-cost models for high-volume tasks and reserving premium systems for workloads where their additional capabilities generate measurable value. He argued that many corporate teams instead default to the most expensive models, "burning through massive budgets with zero governance, audit ability and control." > Your margin is my opportunity: AI version… > > The biggest surprise of 2026 is that the capability gap between the best open-weight/source models and the best closed models has narrowed much faster than the pricing gap. The pricing gap remains enormous while the capability gap is… https://t.co/LjLgpR82S9 > > — Chamath Palihapitiya (@chamath) June 6, 2026 **Read Also: Oracle's AI Spending Bill Keeps Growing: Capex Could Hit $100 Billion As Stargate Expands** ## **AI Boom Faces Questions On Financing, Hiring And Returns** Earlier, Investor **Michael Burry** questioned whether complex financing structures may be obscuring how **Nvidia Corp.** (NASDAQ:NVDA) GPUs and other AI hardware were funded, suggesting that SPVs and related vehicles may have moved assets off balance sheets. He also described the broader setup as "all Fugazi," pointing to ties between insurers, private credit and retirement funds. **Uber Technologies Inc.** (NYSE:UBER) CEO **Dara Khosrowshahi** said AI tools had boosted productivity so significantly that the company was forced to slow hiring after quickly exhausting its annual AI budget. He said engineers had become more efficient and Uber was "metering" headcount growth while expanding AI adoption. **Futurum Equities** CEO **Daniel Newman** said investors should not expect quick returns from the AI boom, calling short-term expectations "idiotic." He argued that the technology remains early-stage and that its most valuable use cases have yet to emerge, recommending a longer-term, five-year outlook. **_Disclaimer_:** _This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors._ **Read Also: Bill Gurley Says Anthropic Isn't Building Software—It's 'Midwifing A Deity'** _Photo courtesy: CarlaVanWagoner / Shutterstock.com_ ### Related Stocks - [ORCL.US](https://longbridge.com/en/quote/ORCL.US.md) - [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md) - [UBER.US](https://longbridge.com/en/quote/UBER.US.md) - [SSTK.US](https://longbridge.com/en/quote/SSTK.US.md) - [ORCL-D.US](https://longbridge.com/en/quote/ORCL-D.US.md) - [NVD.DE](https://longbridge.com/en/quote/NVD.DE.md) ## Related News & Research - [How the 'Nostradamus of AI' failed to spot a market crash](https://longbridge.com/en/news/294580418.md) - [Meet the Gen Z AI whiz at the center of a hedge fund meltdown](https://longbridge.com/en/news/294425537.md) - [Here are my 3 top artificial intelligence (AI) stocks to buy in August](https://longbridge.com/en/news/294601562.md) - [Wall Street Is Rewarding AI Results, Not Just AI Spending](https://longbridge.com/en/news/294613820.md) - [Consultants have found an unlikely new career path: policing AI](https://longbridge.com/en/news/294353497.md)