AI: The New Email – Why Artificial Intelligence Might Be a Loss Leader for Tech Giants
Imagine this: It’s the early 2000s, and you’re signing up for a brand-new email account. It’s free. Totally, completely free. Fast forward a few decades, and email remains free. Yet, companies like Google and Microsoft are making billions. How? By using email as a loss leader—a freebie that keeps you inside their ecosystem, consuming their other paid services.
Now, swap out “email” for “AI,” and we are probably looking at the next big loss leader—an incredibly disruptive loss leader, of course, but a loss leader nonetheless.
Loss Leaders: The Business Trick Happening Now
A loss leader is a product or service that companies practically give away or sell at a loss to hook customers into buying something else that makes real money.
Think of your local grocery store offering dirt-cheap tins of beans. They know you’ll walk in for a can of beans and walk out with $100 worth of snacks, drinks, and fancy cheeses you never planned to buy.
Tech companies have mastered this. Microsoft Word? Once a one-time purchase, now part of a monthly subscription. Amazon Prime? You think you’re paying for free shipping, but you’re actually spending more because you feel like you have to use the service to justify the membership fee.
AI: The Next Great Loss Leader?
Right now, companies are throwing billions into AI development. ChatGPT, Google Gemini, Claude, Alibaba, and even the new DeepSeek, among others, are being rolled out with generous or completely free tiers.
Even enterprise AI pricing is surprisingly affordable, given the insane compute costs. But here’s the back-office truth—tech companies are realising that they are not expecting to make their biggest profits by selling AI access directly. They’re using AI to supercharge their real money-makers.
- Microsoft is stuffing AI into Office 365 (because an AI-powered Excel makes subscriptions stickier).
- Google is integrating AI into search and ads (to make sure businesses keep paying for visibility).
- Amazon is offering AI tools through AWS (because businesses need cloud computing to run AI-powered features).
This playbook is straight from the Web Email Era. Gmail wasn’t about email—it was about getting you into Google’s ecosystem, where you’d search more, click more ads, and store your files in Google Drive. Likewise, AI isn’t the main meal—it’s the enticing appetiser that makes you stay for the full buffet and the expensive wine.
AI is likely going to be that rare breed of technological innovation that acts as both an ecosystem anchor and an enabling technology, depending on how it is deployed and used.
Commercial models for both are horrifically difficult to master as a newbie trying to make a commercially viable business.
Bad News for AI-Only Startups
If AI becomes the equivalent of free email, what happens to the companies betting everything on selling AI itself? Not good things.
Imagine launching a new bottled water brand… right as Coca-Cola decides to give away unlimited free water dispensers in every home. The market dynamics get brutal—fast.
AI-only companies face:
- Falling Prices: If companies are giving away open-source AI or bundling free access and AI features with their products, AI startups charging standalone fees will struggle to compete and are likely to be consolidated into the big players.
- Customer Retention Problems: If a business is already paying for AI-powered Microsoft Word, why would they shell out extra for a separate AI writing tool?
- Funding Challenges: Investors may hesitate to back companies trying to charge for something the tech giants are giving away for free.
Now before you call BS on this, we can agree that history is a great teacher, and one of the benefits of being a little older is the benefit of foresight influenced by hindsight.
I remember using all of these tech products with awe for years before they inevitably fell into obsolescence or were acquired for dimes on the dollar.
- Netscape was the first widely used web browser and made money by selling browser licenses. Then, Microsoft bundled Internet Explorer for free with Windows, instantly killing Netscape’s paid model.
- RealNetworks pioneered online streaming with RealPlayer, charging for premium features. But then Apple made iTunes free, and YouTube came along with free streaming, wiping out RealPlayer’s business.
- BlackBerry dominated business communication with its secure messaging and QWERTY keyboards. Then Apple and Android made touchscreen smartphones mainstream. A late pivot to cybersecurity saved its bacon, but it never regained its former glory.
- MapQuest was the go-to online mapping service. Then Google Maps launched for free and started offering turn-by-turn navigation, instantly making MapQuest irrelevant.
- Palm made PDAs (Personal Digital Assistants) that were hugely popular for productivity. Then smartphones integrated all PDA features for free.
- AOL made billions charging for dial-up internet access. Then broadband became widely available, and people no longer needed AOL’s paywalls. It now exists as part of Yahoo.
- Sun Microsystems was a leader in enterprise computing and created Java. But then Linux, open-source software, and cloud computing disrupted the business. Oracle acquired Sun in 2010, but its brand disappeared.
- Skype pioneered internet video calling. But then Zoom, FaceTime, WhatsApp, and Teams made video calls free, easy, and integrated into existing platforms.
All were the OpenAI’s of my time, all brilliant latest tech that became somewhat irrelevant as the market changed around them.
Lesson for AI Startups?
If history tells us anything, once a paid product becomes free or bundled into an ecosystem, standalone businesses collapse fast. AI companies charging for basic AI tools will face the same fate if giants keep integrating AI into their free and paid services (which they have to do to stay relevant).
Where AI Startups Can Still Win
So, should AI-only companies pack up and go home? Not necessarily. If history tells us anything, loss leader markets don’t destroy all competition—they just force smarter business models.
- Go Niche: General AI tools may become free, but specialized AI with deep expertise (e.g., legal AI, medical AI, cybersecurity AI) can still command a premium if built for industry-specific needs.
- Avoid Competing on Generic AI Features – The more AI is bundled into existing platforms, the harder it is to charge for it. AI startups need to go beyond generic features.
- Sell the Picks and Shovels: The companies that made the most money during the California Gold Rush weren’t miners—they were the ones selling shovels, jeans, and supplies. AI infrastructure, fine-tuning models, and AI-enhanced API services could be the “picks and shovels” of the AI gold rush. Hugging Face is my bet for this—the best shovel vendor out there.
- Community & Customisation: OpenAI, Google, and Microsoft are playing a volume game. Smaller AI firms can focus on customisable, tailored solutions where customers pay for personalised experiences.
- Frontier Tech: Utilise QML and other advanced tech to try and outplay the giants—it’s hard, almost impossible, but feasible.
- Go Enterprise, Not Consumer – AI companies should target B2B and enterprise contracts, where companies are still willing to pay for customised AI solutions.
- Develop IP and Defensible Moats – Proprietary datasets, unique models, and exclusive partnerships can help differentiate AI businesses from open-source alternatives - even then its a hard slog to compete and differentiate on a general stage.
Conclusion: AI’s Future is Hidden in Plain Sight
Is this going to happen? YES.
Ask anyone who watched the web bubble of 2000 fold in on itself like a giant soufflé, wiping out billions in a few months.
What does this mean for OpenAI?
OpenAI shouldn’t fold - its a pretty big name, but its leadership is going to need to get super creative, acquire, and roll out chargeable add-ons—or they will become irrelevant quickly. I would even say its going to take a streak of commercial business genius to keep it as a stand alone business and from being absorbed into Msoft or A.N Other company - but we will see.
When you speak to four people in 24 hours who canceled a paid OpenAI account, (as I have) it doesn’t take a rocket scientist to read the writing on the wall.
Too much hype, too much late market silly money and too much competition has led to some gob-smackingly irrelevant investment portfolios in AI with huge valuations against low or non existent revenues.
I envisage a fast approaching point where there will be a scramble to sell, lots of consolidation and lots of write offs in AI investments that fail to gain traction as the ability to compete becomes harder and harder and harder…
I believe just like the early 2000’s AI is approaching the start of the end of its hype cycle, and to borrow from Churchill - Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning.(and perhaps the end of the silly valuations placed on a tech that is set to be an enabling loss leader in its general form).
Since 2022, 74 AI companies have closed funding rounds with post-money valuations of $100 million or more, with estimates that >80% of these companies generate less than $10 million in ARR.
Am I wrong? We will see, but certainly the events of the past few days have been played out and I have a distinct feeling I have seen this movie before, it will be interesting to see if the ending is the same.
This will not be the death of AI. Far from it.
But only the best will survive as the viciousness of this great rug pull increases in its momentum.
