{"id":328268,"date":"2025-12-02T21:22:15","date_gmt":"2025-12-02T21:22:15","guid":{"rendered":"https:\/\/peraltafinancing.com\/uncategorized\/openai-needs-adult-supervision\/"},"modified":"2025-12-02T21:22:15","modified_gmt":"2025-12-02T21:22:15","slug":"openai-needs-adult-supervision","status":"publish","type":"post","link":"https:\/\/fivemor.com\/?p=328268","title":{"rendered":"OpenAI Needs Adult Supervision"},"content":{"rendered":"


\n<\/p>\n

\n
\n

At <\/span>first glance, OpenAI is on top of the world. The company recently announced that its revenues <\/span><\/span>have surged past $10B \/ year<\/span><\/u><\/span><\/a>, on track with <\/span><\/span>expectations to reach $125B by 2029<\/span><\/u><\/span><\/a>. That\u2019s more than Stellantis, Nestl\u00e9 or Bank of America make today. Though we\u2019ve become accustomed to \u201ccrazy\u201d numbers when it comes to AI companies, we do need to acknowledge that this type of growth is\u2026 extraordinary.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

There are, however, clouds on the horizon. Whether from ambition, arrogance or lack of discipline, OpenAI is in trouble today. The company must overcome significant internal and market issues if it is to survive the coming years.<\/span><\/span><\/span><\/p>\n<\/div>\n

Top-line: OpenAI’s monetization issue<\/span><\/span><\/h2>\n<\/p>\n
\n
\n
\n
\n
\n
\"Sam<\/wow-image><\/div>\n<\/div>\n<\/figure>\n<\/div>\n<\/div>\n<\/div>\n
\n

Even as the company struggles to monetize users amidst growing competition within the Large Language Model space, the paying customers still aren\u2019t being monetised properly. OpenAI either misunderstands A) the company\u2019s cost base or B) the value delivered by its products. Very unserious for a company that wishes to become a market leader.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n
\n

<\/p>\n

TLDR<\/span><\/span><\/strong>: OpenAI rules free chatbots (500M weekly users, 70% market share) but converts just 3% to paid, while its Pro tier hemorrhages cash; \u201cgreat free product\u201d + \u201cvibe-pricing\u201d will not bankroll AGI.<\/span><\/span><\/span><\/p>\n

<\/span><\/p><\/blockquote>\n<\/div>\n<\/div>\n

Bottom-line: OpenAI’s cost pressure<\/span><\/span><\/h2>\n<\/p>\n
\n

As frontier-model training costs continue to <\/span><\/span>increase towards $1B<\/span><\/u><\/span><\/a>, OpenAI\u2019s bottom line is being squeezed. The company\u2019s Operating Expenses (OPEX if you\u2019re dirty) rose to <\/span><\/span>$8.7B in 2024, leading to a total net loss of $5B<\/span><\/u><\/span><\/a> in the same year. And sure, the company is growing revenues fast at the same time, following the tried and true Silicon Valley model of breaking fast and moving things or whatever. Amazon and Uber famously took <\/span><\/span>years <\/span><\/span><\/em>to turn a profit.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

However, three large obstacles stand on the path to long-term profitability.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

Firstly, OpenAI\u2019s 2024 losses would have been more than double without Microsoft discounts on compute. <\/span><\/span>$12B is what has been estimated<\/span><\/u><\/span><\/a>. Because it lacks its own GPU infrastructure, OpenAI remains firmly tied to its Redmond benefactor. But, as <\/span><\/span>the relationship sours in real time<\/span><\/u><\/span><\/a>, Microsoft contracts appear increasingly under threat. Pursuing AGI independently and without shackles is a worthy endeavour, but without a backup plan, it will only add pressure on OpenAI\u2019s CFO (<\/span><\/span>who joined only last year<\/span><\/u><\/span><\/a>, likely to remedy an unsustainable situation).<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

Which brings us to the second issue. The two most successful companies of the past century (Apple, Nvidia), have focused on building railways (smartphones, GPUs) for people building and running trains (mobile apps, AI). Sam Altman knows this: his algorithms\u2019 success is entirely dependent on data centers and the chips that power them. Lack of direct access is an existential threat. And so, OpenAI, SoftBank and Oracle have announced the Stargate project, a $500B initiative to build a network of massive AI data centers across the US. That\u2019s a lot of money for a company <\/span><\/span>still some way away from profitability<\/span><\/u><\/span><\/a>. Especially so given <\/span><\/span>reports that the project is struggling to get off the ground<\/span><\/u><\/span><\/a>.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

And sure, at <\/span><\/span>a $300B valuation<\/span><\/u><\/span><\/a>, OpenAI benefits from an infinite money glitch, enabled by Wall Street goodwill\u2026 but the above checks will have to be written as OpenAI\u2019s competitive advantage versus other major players shrinks. This should worry investors, regardless of the company\u2019s shiny vision for the future.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n
\n

<\/p>\n

TLDR: With a $5B loss that would have been $12B without Microsoft GPU discounts, a struggling $500B data center moonshot, looming custom\u2011chip costs and $12B in CoreWeave obligations, the supposedly infinite money pile suddenly looks awfully finite as rivals narrow the gap.<\/span><\/span><\/span><\/p>\n

<\/span><\/p><\/blockquote>\n<\/div>\n<\/div>\n

External factors: increasing competition & shifting markets<\/span><\/span><\/h2>\n<\/p>\n
\n

The above wouldn\u2019t be much of a problem if OpenAI had a particularly unique competitive advantage. But 2022 was a while ago, and competitors are closing the gap on algorithm quality while cutting costs. And they\u2019re doing so <\/span><\/span>fast<\/span><\/span><\/em>.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

Every rival that matters has already hacked one or more structural levers to drag its dollar cost per million input tokens to enable a price well below OpenAI\u2019s <\/span><\/span>$2.50\/M tokens<\/span><\/span><\/strong>. Unless OpenAI matches or outflanks those levers, the company will bleed margin faster than it scales revenue.<\/span><\/span><\/span><\/p>\n<\/div>\n

\n

First, here\u2019s what the incumbent heavyweights are doing:<\/span><\/span><\/span><\/p>\n<\/div>\n

\n