Why Overrated AI Startups Are Getting Their Funding Pulled

Vinod Pandey


In January of 2023, Microsoft announced that they would be investing $10 billion into OpenAI, the parent company of ChatGPT. It’s been over a year since that day and OpenAI is now reporting that they’ve only received a fraction of the promised $10 billion. 

This isn’t to say that Microsoft won't follow through with their full investment. After all, $10 billion is just a drop in the bucket for a $2.5 trillion giant like Microsoft. But, it’s also clear that Microsoft isn’t exactly jumping up and down to pour more funding into OpenAI. 

Overrated AI Startups Are Getting Their Funding Pulled

In fact, from the very beginning, Microsoft structured the investment in tranches that would need to be unlocked over a long period of time. Reading between the lines, it seems that tensions between OpenAI and Microsoft are more than just financial. 

In November, for example, Microsoft was forced to take a non-voting position on the OpenAI board despite owning 49% of the company. And with the whole Sam Altman fiasco, Microsoft may even have a legitimate reason to completely back out of the deal, and that’s just the most notable AI drama. 

It turns out that these sorts of fallouts are happening across the board throughout the AI world as VCs pull back on funding AI projects with ridiculous valuations. I mean, just think about this. OpenAI was looking to raise at a valuation of $90 billion and the company barely pulls in $1 billion in revenue per year. 

Not to mention, Sam Altman himself has admitted that OpenAI will need $100 billion not in valuation, but in just funding to fully develop generative intelligence. All of this has been during a time in which ChatGPT has been slowly bleeding users month after month. 

None of this is to say that AI is not the future because it obviously is, but it does seem like the AI industry very much got ahead of itself, and VCs are starting to pull the plug. Just recently, for example, a $4 billion AI health startup named Olive went under. 

And Olive is simply one of a growing list of unicorns turned zombies. So, here’s the rise and fall of the AI hype and what’s next for the AI industry. 


All burgeoning industries go through boom and bust cycles. That’s just how new innovations and tech are created. You have these massive bursts of enthusiasm and excitement followed by a harsh reality check before true mainstream adoption begins. That’s exactly what happened with all the dot-com companies in the early 2000s. 

After years of unlimited funding, the vast majority of dotcom companies ended up going under never to be heard of again. It was only after this cleansing that internet companies like Amazon, Google, and Facebook were really able to cement themselves for the long term. 

AI has a long history of being next big thing graph

In most industries, this sort of cycle takes maybe 20 to 30 years, but the same cannot be said about the AI industry. In fact, AI has been repeating this cycle over and over again since the 1950s which brings us back to the father of modern computing: Alan Turing. 

In 1950, Turing published a landmark research paper called “Computing Machinery and Intelligence”. In the paper, he explores the idea of whether computers can think for which he offers a very simple answer. 

If a computer can engage in conversation with a human without the human noticing that they’re talking to a computer, then, the machine can indeed think. He dubbed this test the imitation game and this is what would eventually become the famous Turing Test. 

Inspired by these ideas, hundreds of scientists would begin working on their own AI projects hoping to pass the Turing Test. This included more simple applications like a checkers bot in 1952 but it also included generative AI. 

In fact, generative AI itself is by no means a new invention. As far back as the 1960s, we’ve seen scientists grapple with the concept of language models and chatbots. The ELIZA chatbot for example attempted to simulate the work of a psychotherapist way back in 1961. 

Scientists were even working on the groundwork for self-driving vehicles. In the 1970s, for example, the Stanford cart was able to navigate itself through a room filled with obstacles. It seemed like AI was truly the next big thing? But, do you know what happened after these breakthroughs? The first AI winter. 

Funding for AI startups was wiped out as investors realized that it was a lot more profitable to invest in chip and computer companies with strong revenue potential as opposed to these theoretical projects. And that’s exactly what they did. 

After they made a crap ton of money with companies like Sun Microsystems, Intel, Microsoft, Sony, and Panasonic they would come back to the speculation table. Japan, for example, would commit $850 million to what they called the fifth-generation computer project. And IBM would begin work on next-generation AI called Deep Blue. 

The goal was to create a chess bot that was better than the world had ever seen. This was by no means an easy task but in 1997, IBM would do it. Deep Blue would decisively beat the chess world champion at the time: Garry Kasparov. 

Once again, it seemed like AI was truly the next big thing. Not only was AI able to outsmart an average Joe but a world champion. This was definitive proof that AI could do great things. But, after that, we would enter yet another AI winter. 

With investors losing money hand over fist during the dotcom bust, the last thing they wanted to do was speculate on AI projects with no revenue potential. That was until they got rich again thanks to web companies like Google, Facebook, Amazon, Uber, AirBnB, and so on bringing us into the current cycle. 

As you can see, for 70 years now, AI has consistently been this sort of elusive next-big thing that’s almost within grasp but not quite. So, will this time be any different? Well, at some point, it does have to be different. 

That’s kind of how AI eventually becomes mainstream. But, judging by history, there’s no guarantee that this cycle is the cycle. That’s what people have been thinking for the past 70 years only to be proven wrong over and over again which brings us to the question of why? 

Also Read:

What Happened To Browser Toolbars?

Exploring Walmart's Flower Selection: Types, Prices, and Options

What Happened To The Once-massive Telecom Company Sprint

Why LinkedIn may be the most powerful social media platform

How Did Europe Fall So Far Behind In Technology


One of the biggest themes surrounding each and every one of these AI winters is funding. The reality is that AI is an extremely expensive game, and it’s only getting more and more expensive each cycle. OpenAI for example is having to pay $1 million per year to get access to the right engineers. 

Not to mention the massive data centers and loads of GPUs they need to make the entire operation possible. This is why Sam Altman thinks they need $100 billion. But, all of that is only half the story because AI is a double-edged sword. 

Not only does it take a stupid amount of money to bring to market but it usually has very little revenue potential at least in its original state. It often has be evolve substantially before being able to generate serious revenue. 

As such, investors are only willing to take this gamble when times are good. When their other investments are doing well when the tech industry as a whole is booming, and when they don’t know where else to invest their money. 

This was exactly what was happening throughout the early 2020s. Investors had hit it big with web companies and unicorns, and with 0% interest rates, they simply did not know where to put all those profits, so they put it into AI. 

But when the music stops playing as was the case in the 1970s, the early 2000s, and right now, they’re quick to cut losses and move back to safety. But, even putting the whole financing of AI aside, it seems that the industry is always stopped by a consistent internal factor themselves. For example, limited computing power. 

Even in the modern day where we have enough computing power for basically any task imaginable, somehow, we don’t have enough computing power for AI. And the main reason for this is that our needs for AI are also constantly evolving. Before we were just trying to create AIs that could hold conversation or excel at a board game. 

Today, we’re trying to create AIs that can not only hold conversations but scour the entire web to answer our questions. Similarly, we’re no longer trying to create carts that can navigate a room within 5 hours but cars that can navigate any road live. 

Our demands for AI are rising exponentially and as such, the resources needed to make these projects a reality are also rising exponentially. But even regardless of how many resources we have, it’s usually only a matter of time until we run into Moravec’s Paradox which describes the odd disparity between human and computer intelligence. 

It turns out that doing things that are super hard for humans like mass data analysis and calculations is super easy for computers. However, doing things that are easy for humans like sensing depth, fine motor skills, and innately leveraging feedback control systems is nearly impossible. 

And if all of those challenges were not hard enough to overcome themselves, well, we have yet to discuss the biggest challenge which is the rise of AI phonies. 


With any hot trend, it’s only a matter of The AI Phonies' time until the opportunists step onto the scene. In crypto for example, for every bitcoin and Ethereum, there are 100s of trash coins, rug pulls, NFT projects, and straight-up scams. 

It’s the same thing with the AI space, just in a more sophisticated manner. For every OpenAI or Tesla FSD-type effort, you’ve got dozens of garbage startups. Most of them have nothing to do with AI whatsoever or they barely leverage AI, but they’ll put it all over their marketing and brand themselves as an AI company. 

In reality, they’re just a construction or consulting firm. But all of a sudden, they're an AI-based construction firm or an AI-based consulting firm. From an outside perspective, these are obviously trash companies that way overstating their use of AI. 

But for desperate VCs who missed out on investing in OpenAI, this truly seems like the future. Sure, they might’ve missed out on bringing AI to search or cars, but they can still be the ones to bring it to the sewer industry or the energy industry even though the application is far more limited. 

Fortunately, this BS doesn’t go on forever, but the fall of these zombie companies is by no means graceful. Usually, what happens is investors pull back on all kinds of AI investing and only the good companies end up surviving the funding cut which is exactly what’s happening right now. 

graph showing Global VC funding has plummeted from $170 billion in the 4th quarter of 2021 to $70 billion in the 3rd quarter of 2023

Global VC funding has plummeted from $170 billion in the 4th quarter of 2021 to $70 billion in the 3rd quarter of 2023. AI was one of the few sectors that was holding up through this funding collapse, but we’re starting to see cracks within the AI world as well. 

In Q3 of 2023, for example, the number of AI VC deals shrunk by 27%. Technically, the total value of these deals increased by 39% to $6.1 billion, but $4 billion of that was just due to one investment from Amazon. And you could say the same thing about all of 2023 actually. 

The main reason that the AI industry has held up as well as it did was simply due to a handful of massive deals from companies like Microsoft, Nvidia, Google, Salesforce, and Amazon. When we take these flagship deals out of the picture though, the outlook for the rest of the AI industry becomes quite bleak.

What happens next? 

Well, most likely another AI winter filled with bankruptcies. The flagship companies like OpenAI will most likely survive but you probably won't hear nearly as much about them on the news as much as they enter a new phase of research and development away from the public light. 

That isn’t to say that this cycle was useless though. Another intrinsically cyclical sector is the chip industry. They have booms and busts basically every decade or even multiple times per decade. But, just because the industry is volatile doesn't mean that they’re not making forward progress. 

Chips today are not even comparable to chips from the 1950s, and it’s the same thing with AI. Historically, it’s simply been a highly volatile industry, and we’re currently going from a high to low, but eventually, the industry will see the other side once again. Do you think AI is overhyped? Comment that down below.

Post a Comment


Post a Comment (0)