Money was no object for the AI industry in early 2025. A vibe check crept in the second half of the year.\u00a0<\/p>\n
OpenAI raised $40 billion<\/a> at a $300 billion valuation. Safe Superintelligence<\/a> and Thinking Machine Labs<\/a> raised individual $2 billion seed rounds before shipping a single product. Even first-time founders were raising at a scale that once belonged only to Big Tech.\u00a0<\/p>\n Such astronomical investments were followed by equally incredible spends. Meta shelled out nearly $15 billion<\/a> to lock up Scale AI CEO Alexandr Wang and spent countless more millions to poach talent from other AI labs. Meanwhile, AI\u2019s biggest players promised close to $1.3 trillion in future infrastructure spending.\u00a0<\/p>\n The first half of 2025 matched the fervor, and investor interest, of the prior year. That mood has shifted in recent months to deliver a vibe check of sorts. Extreme optimism for AI, and the accompanying wild valuations, is still intact. But that rosy view is now being tempered with concerns over an AI bubble bursting, user safety, and the sustainability of technological progress at its current pace.\u00a0<\/p>\n The era of unabashed acceptance and celebration of AI is fading just a skosh at the edges. And with it, more scrutiny and questions. Can AI companies sustain their own velocity? Does scaling in the post-DeepSeek era require billions? Is there a business model that returns a sliver of the multi-billions of investment?\u00a0<\/p>\n We\u2019ve been there for every step. And our most popular stories of 2025 tell the real story: an industry hitting a reality check even as it promises to reshape reality itself.\u00a0<\/p>\n The biggest AI labs got bigger this year.\u00a0<\/p>\n Techcrunch event<\/p>\n \n\t\t\t\t\t\t\t\t\tSan Francisco<\/span> In 2025 alone, OpenAI raised a Softbank-led $40 billion round<\/a> at a $300 billion post-money valuation. The company also reportedly has investors like Amazon<\/a> orbiting with compute-tied circular deals, and is in talks to raise $100 billion at an $830 billion <\/strong>valuation<\/a>. That would bring OpenAI close to the $1 trillion valuation it is reportedly seeking in an IPO next year.\u00a0<\/p>\n OpenAI rival Anthropic also closed $16.5 billion this year across two rounds, its most recent raise pushed its valuation to $183 billion<\/a> with heavy hitters like Iconiq Capital, Fidelity, and the Qatar Investment Authority participating. (CEO Dario Amodei confessed to staff in a leaked memo<\/a> that he was \u201cnot thrilled\u201d about taking money from dictatorial Gulf states).\u00a0<\/p>\n Then there\u2019s Elon Musk\u2019s xAI, which raised at least $10 billion<\/a> this year after acquiring X<\/a>, the social media platform formerly known as Twitter that Musk also owns.<\/p>\n We\u2019ve also seen smaller, new startups get a hypey boost from froth-mouthed investors.\u00a0<\/p>\n Former OpenAI chief technologist Mira Murati\u2019s startup Thinking Machine Labs secured a $2 billion seed round at a $12 billion valuation<\/a> despite sharing almost no information about its product offering. Vibe-coding startup Lovable\u2019s $200 million Series A<\/a> earned it a unicorn horn just eight months after launching; this month, Lovable raised another $330 million<\/a> at a nearly $7 billion post-money valuation. And we can\u2019t leave out AI recruiting startup Mercor, which raised $450 million this year across two rounds, the latest bringing its valuation up to $10 billion<\/a>.\u00a0<\/p>\n These absurdly large valuations are still happening even against the backdrop of still-modest enterprise adoption figures and serious infrastructure constraints, heightening fears of an AI bubble.\u00a0<\/p>\n For the larger firms, those numbers aren\u2019t coming from nowhere. Justifying those valuations requires building vast amounts of infrastructure.\u00a0<\/p>\n The result has created a vicious cycle. Capital raised to fund compute is increasingly tied to deals where the same money flows back into chips, cloud contracts, and energy, as seen in OpenAI\u2019s infrastructure-linked funding with Nvidia<\/a>. In practice, it\u2019s blurring the line between investment and customer demand, stoking fears that the AI boom is being propped up by circular economics rather than sustainable usage.<\/p>\n Some of the biggest deals this year powering the infrastructure boom<\/a> were:\u00a0<\/p>\n But cracks are beginning to show. A private financing partner, Blue Owl Capital, recently pulled out<\/a> of a planned $10 billion Oracle data-center deal tied to OpenAI capacity, underscoring how fragile some of these capital stacks can be.\u00a0<\/p>\n Whether all that spending ultimately materializes is another question. Grid constraints, soaring construction and power costs, and growing pushback from residents and policymakers \u2013 including calls from figures like Sen. Bernie Sanders<\/a> to rein in data center expansion \u2013 are already slowing projects<\/a> in some regions.\u00a0<\/p>\n Even as AI investment remains enormous, the infrastructure reality is beginning to temper the hype.\u00a0<\/p>\n In 2023 and 2024, each major model release felt like a revelation, with new capabilities and fresh reasons to fall for the hype. This year, the magic faded, and nothing captured that shift better than OpenAI\u2019s GPT-5 rollout<\/a>.\u00a0<\/p>\n While it was meaningful on paper, it didn\u2019t land with the same punch<\/a> as <\/em>earlier releases like GPT-4 and 4o. Similar patterns emerged across the industry as improvements from LLM providers were less transformative and more incremental or domain-specific.\u00a0<\/p>\n Even Gemini 3<\/a>, which is topping several benchmarks, was only a breakthrough insofar as it brought Google back up to equal footing with OpenAI \u2013 which sparked Sam Altman\u2019s infamous \u2018code red\u2019 memo and OpenAI\u2019s fight to maintain dominance.<\/p>\n There was also a reset this year in terms of where we expect frontier models to come from. DeepSeek\u2019s launch of R1, its \u201creasoning\u201d model<\/a> that competed with OpenAI\u2019s o1 on key benchmarks, proved that new labs can ship credible models fast and at a fraction of the cost.\u00a0<\/p>\n As the size of each leap between new models shrinks, investors are focused less on raw model capacity and more on what\u2019s wrapped around it. The question is: who can turn AI into a product that people rely on, pay for, and integrate into their daily workflows?\u00a0<\/p>\n That shift is manifesting in several ways as companies see what works, and what customers will let fly. AI search startup Perplexity, for example, briefly floated the idea of tracking users\u2019 online movements to sell them hyper-personalized ads<\/a>. Meanwhile, OpenAI was reportedly considering charging up to $20,000 per month for specialized AI<\/a>, a sign of how aggressively companies tested the waters of what customers might be willing to pay.<\/p>\n More than anything, though, the fight has moved to distribution. Perplexity is trying to stay relevant by launching its own Comet browser<\/a> with agentic capabilities and paying Snap $400 million<\/a> to power search inside Snapchat, effectively buying its way into existing user funnels.\u00a0<\/p>\n OpenAI is pursuing a parallel strategy, expanding ChatGPT beyond a chatbot and into a platform. OpenAI has launched its own Atlas browser <\/a>and other consumer-facing features like Pulse<\/a>, while also courting enterprises<\/a> and developers<\/a> by launching apps inside ChatGPT<\/a> itself.\u00a0<\/p>\n Google, for its part, is leaning on incumbency<\/a>. On the consumer side, Gemini is being integrated directly into products like Google Calendar<\/a>, while on the enterprise side, the company is hosting MCP connectors<\/a> to make its ecosystem harder to dislodge.\u00a0<\/p>\n In a market where it\u2019s getting tougher to differentiate by dropping a new model, owning the customer and the business model is the real moat.\u00a0<\/p>\n AI companies received unprecedented scrutiny in 2025. More than 50 copyright lawsuits wound through the courts, while reports of \u201cAI psychosis\u201d \u2013 the result of chatbots reinforcing delusions and allegedly contributing to multiple suicides and other life-threatening episodes \u2013 sparked calls for trust and safety reforms.\u00a0<\/p>\n While some copyright battles met their end \u2013 like Anthropic\u2019s $1.5 billion settlement to authors<\/a> \u2013 most are still unresolved. Though the conversation appears to be shifting from resistance against using copyrighted content for training, to demands for compensation (See: New York Times sues Perplexity for copyright infringement<\/a>).<\/p>\n Meanwhile, mental health concerns around AI chatbot interactions \u2013 and their sycophantic responses<\/a> \u2013 emerged as a serious public health issue following multiple deaths by suicide<\/a> and life-threatening delusions in teens<\/a> and adults after prolonged chatbot usage. The result has been lawsuits, widespread concern among mental health professionals, and swift policy responses like California\u2019s SB 243 regulating AI companion bots<\/a>.<\/p>\n Perhaps most telling: the calls for restraints are not coming from the usual anti-tech suspects.\u00a0<\/p>\n Industry leaders have warned against chatbots \u201cjuicing engagement<\/a>,\u201d and even Sam Altman has cautioned against emotional over-reliance on ChatGPT.\u00a0<\/p>\n Even the labs themselves started sounding alarms. Anthropic\u2019s May safety report documented Claude Opus 4 attempting to blackmail engineers<\/a> to prevent its own shutdown. The subtext? Scaling without understanding what you\u2019ve built is no longer a viable strategy.<\/p>\n If 2025 was the year AI started to grow up and face hard questions, 2026 will be the year it has to answer them. The hype cycle is starting to fizzle out, and now AI companies will be forced to prove their business models and demonstrate real economic value.<\/p>\n The era of \u2018trust us, the returns will come\u2019 is nearing its end. What comes next will either be a vindication or a reckoning that makes the dot-com bust look like a bad day of trading for Nvidia. Time to place your bets.\u00a0<\/p>\n<\/div>\nHow the year started<\/strong><\/h2>\n
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/08\/GettyImages-2194585046.jpg?w=680\")
\n\t\t\t\t\t\t\t\t\t\t\t\t\t|<\/span>
\n\t\t\t\t\t\t\t\t\t\t\t\t\tOctober 13-15, 2026<\/span>\n\t\t\t\t\t\t\t<\/p>\n<\/p><\/div>\n<\/p><\/div>\n<\/div>\nBuild, baby, build<\/strong><\/h2>\n
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/07\/GettyImages-2224533087.jpg?w=680\")
\n
The expectation reset<\/strong><\/h2>\n
![\"In](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/01\/GettyImages-2196335614.jpg?w=680\")
From model breakthroughs to business models<\/strong><\/h2>\n
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2024\/12\/GettyImages-2151467531.jpg?w=680\")
The trust and safety vibe check<\/strong><\/h2>\n
![\"Character.AI](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/10\/U18-Product-Image.jpg?w=680\")
Looking ahead<\/strong><\/h2>\n