“When people get very excited, as they are today about AI for example, **every experiment gets funded, every company gets funded — the good ideas and the bad ideas — and investors have a hard time, in the middle of this excitement, distinguishing between the good ideas and the bad ideas. But it doesn’t mean that anything that’s happening isn’t real. AI is real, it is going to change every industry.”
Jeff Bezos, October 2025 (at the Italian tech week)
This week OpenAI became the most valuable private company, crossing the $500 billion valuation mark and dethroning SpaceX. The valuations seem to be doubling and tripling in a short time, in the billions. One year old companies become unicorns and reach $100M ARR in record times. Top AI founders like Mira Murati, turn down $1 billion offers…. what’s going on?
The data confirms what we’re seeing on the ground: The AI boom is dominating venture capital in an unprecedented way. This year so far, venture capitalists have already poured a stunning $192.7 billion into AI startups this year, putting 2025 on track to be the first year that more than half of the total VC dollars went into this single industry globally. In the US, the trend is even more pronounced, with 62.7% of invested dollars dedicated to AI companies in the most recent quarter.
The obvious question, then, is inescapable: Are we in an AI bubble?
As investors, we must acknowledge the warning signs and the historical parallels while simultaneously recognizing the foundational shift AI represents. Prominent voices within the industry are already sounding the alarm: Sam Altman conceded that investors as a whole are “overexcited about AI”, and OpenAI Chairman Bret Taylor stated, “I think we’re also in a bubble, and a lot of people will lose a lot of money”.
The Scale of the Frenzy and the Investment Gap
The fears of a speculative bubble rivaling the late 1990s dot-com craze are growing. The spending levels are astronomical; the final bill for advanced chips and data centers may run into the trillions, with Altman himself anticipating OpenAI spending “trillions” on infrastructure.
Yet, we are pouring immense capital into a technology that, for all its potential, remains somewhat unproven as a profit-making business model. Here’s where the fundamentals raise eyebrows:
- The Revenue Shortfall: Bain & Co. predicted that by 2030, AI companies will require $2 trillion in combined annual revenue to fund the necessary computing power, but their revenue is likely to fall $800 billion short of that mark.
- Lack of Demonstrated ROI: Academic research is fuelling skepticism, with researchers at MIT finding that 95% of organisations saw zero return on their investment in AI initiatives. Furthermore, findings from Harvard and Stanford suggest that employees are creating “workslop”—AI-generated content that lacks substance—which could cost large organisations millions in lost productivity. While this is improving, AI penetration in the enterprise remains around 35%-40% still.
- Diminishing Returns: AI developers, betting on scaling laws, have begun experiencing diminishing returns from their costly efforts to build more advanced models over the past year. For example, it’s largely believed that LLMs are reaching some kind of plateau when it comes to improvement, especially given the massive investment in new models.
Like the dot-com era, we are seeing a massive infrastructure build-out, sky-high valuations, and sometimes questionable metrics for growth. However, like Amazon and Google emerged from the internet bubble to thrive, AI will likely transform the economy and create huge economic value in the future.
How the AI Bubble Changes Investor Evaluation
I recently attended an investor panel (Chatam house rules, so no attribution), where a leading investor in AI startups said something along the lines of “it doesn’t matter whether we are in a bubble or not. While many companies will fail, investors are overall better off deploying and getting exposure into the best AI companies”. The reality is, we’re in a bubble, but the size of the opportunity to disrupt every industry is real, and huge.
The market froth has led to a major shift in how capital is deployed. The overriding theme today is bifurcation
1. The AI/Non-AI Divide
As a PitchBook director of research Kyle Sandfords concisely put it: “You’re in AI, or you’re not getting funded easily”.
If you are a startup not focused on AI, the broader picture is bleak. The total number of companies securing venture funding globally is currently on track to be the lowest in years, as is the number of VC firms raising new funds. Due to the tight environment for public listings and acquisitions, many venture investors are simply unwilling to make new bets on unproven, non-AI focused companies.
Conversely, VCs and their backers are “being more deliberate” and focusing their money on AI. However, even within AI, capital is concentrated: most of the funding is flowing to established startups: OpenAI, Anthropic and xAI both raised billions in funding recently, while lesser-known upstarts might struggle as investors are weary of commoditisation and unit economics.
2. Evaluating Value Over Velocity
We are past the point where merely having “AI” in your pitch deck guarantees a check. Given the skepticism about profit-making models and the reported zero ROI for many corporate initiatives, we are now intensely focused on tangible, monetisable value.
AI developers are hoping that as models improve and handle more complex tasks, businesses will be willing to spend far more to access the technology, potentially justifying high subscription fees. Anthropic, for example, is sharpening Claude’s coding skills, recognising coding as the “most economically important and immediately tractable area”.
We are looking closely at how founders navigate the complexity of the enterprise market:
- Will the AI truly replace legacy enterprise applications like SAP or ServiceNow? Highly placed executives are skeptical, noting the high cost and time commitment of overhauling internal systems at large companies.
- Alternatively, is the solution in creating agents that interact with and automate the use of existing enterprise software? This approach of rethinking the user’s role, rather than replacing the software itself, might prove more viable.
Takeaways for Startups: How to Navigate a Bifurcated Market
For founders looking to survive and thrive in this climate, there are three critical takeaways:
1. Prove Your Economic Utility
If you are an AI startup, you must move beyond the “potential” and demonstrate that your solution generates measurable productivity gains and revenue, not “workslop”. Focus on use cases that are demonstrably “economically important and immediately tractable”. If your model can justify a substantial price point—like the reported $2,000 monthly subscription being discussed for some AI products—you will be attractive.
2. Understand the Global Playing Field
For European founders, the pull of the US is stronger than ever. The US offers a deeper capital pool, a more risk-tolerant culture, and significantly higher valuations—often double the price compared to London deals. American investors are better equipped to “understand the runway and how much money you need to burn” for heavy R&D and computing costs. Furthermore, American deals often close in weeks, compared to the months European founders endure at home.
Founders seeking fast growth and big checks are increasingly moving their businesses to the US, often requiring a “Delaware Flip”. If you are scaling an AI company, you may need to look at the US, where “it’s king in AI”.
3. Prepare for Computational Costs
AI requires heavy upfront spending for powerful computing infrastructure and specialist talent. While engineering talent may remain abundant and cheaper in places like Europe, the sheer capital needed for R&D and compute costs drives the necessity for large, fast funding rounds. This dynamic favors large firms capable of writing massive checks and creates unique arrangements, such as chipmaker Nvidia investing in customers like OpenAI.
Ultimately, while the AI bubble is undeniable, the underlying technology is transformative. Our job as investors is to be rigorous in due diligence, focus on the potential for long-term profit realisation rather than speculative frenzy, and back the resilient founders who are building the infrastructure and applications that will survive the inevitable shakeout. The next wave of tech giants will emerge from this bubble, and we intend to be alongside them.
At Remagine Ventures, we continue to invest in pre-seed startups leveraging AI to increase the speed of tasks, reduce costs and improve quality.
I'm a former general partner at google ventures, head of Google for Entrepreneurs in Europe and founding head of Campus London, Google's first physical hub for startups.
I'm also the founder of Techbikers, a non-profit bringing together the startup ecosystem on cycling challenges in support of Room to Read. Since inception in 2012 we've built 11 schools and 50 libraries in the developing world.
- Weekly Firgun Newsletter – February 27, 2026 - February 27, 2026
- The Rise of the Personal AI Agent - February 26, 2026
- Pre-Seed Is Stronger than ever, but less startups are getting funded - February 25, 2026
