The Software Industry’s Looming Ecommerce Problem
Software is on the cusp of a dramatic shift, one that will mirror the challenges and dynamics of the ecommerce world. Expect thinner margins, relentless competition, and a tougher path to profitability. This isn’t speculation; it’s a direct consequence of the rapid advancements being made by Anthropic, Google, and OpenAI.
From TripAdvisor to Zero: The Google Effect
Consider the evolution of travel booking. In 2010, searching for hotels in Novel York on Google largely directed users to platforms like TripAdvisor. By 2020, Google had integrated its own hotel booking system directly into search results, effectively cutting out the middleman. TripAdvisor’s stock price reflects this shift – a significant decline between 2015 and today.
Google’s success stemmed from capturing demand at the point of search, focusing on high-traffic categories like travel. However, Google was limited by its resources; it could only build dedicated products for the largest markets.
AI’s Infinite Capacity: A New Game
The same pattern is poised to unfold across the entire digital services landscape. But this time, the resource constraint that protected smaller software categories is gone. In 2025, asking ChatGPT for CRM recommendations might yield suggestions like Attio, Pipedrive, and Zoho. But by 2028, the expectation is that ChatGPT – or a similar frontier model – will build a CRM tailored to your needs, import your data, and run it at a fraction of the cost.
The key difference? Google needed human engineers to build each vertical product. OpenAI, Anthropic, and the current iteration of Google (with Gemini) won’t face the same limitations.
The Disappearing Moat
Currently, “vibe coding” – the process of generating software through AI prompts – is still imperfect. It requires human oversight, raises security concerns, and relies on external hosting. However, these hurdles are expected to diminish as frontier model companies develop their own infrastructure to host and maintain the software they generate.
Imagine Claude’s artifacts, but scaled up to encompass fully functional digital products – hosted, maintained, and updated by the AI that created them. The traditional software moat – code complexity and ecosystem lock-in – will erode when AI can rebuild a tool in seconds and seamlessly migrate data.
The Customer Grab
The impact isn’t just about increased competition; it’s about companies taking your customers before they even reach you. This shifts the software industry towards an ecommerce-like model, characterized by near-zero margins unless you control distribution and aren’t reliant on major platforms like Google or Meta.
The Race for Cost Efficiency
The competition is already visible in the race to deliver cheaper AI work. OpenAI’s GPT-5.4 mini is optimized for coding, computer use, and subagents. Anthropic’s Claude Haiku 4.5 is positioned as its fastest and most cost-efficient model, although Google’s Gemini 2.5 Flash-Lite is designed for at-scale usage. Pricing reflects this focus: Gemini 2.5 Flash-Lite currently offers the lowest rates on standard pricing, followed by GPT-5.4 nano and Claude Haiku 4.5.
The Pentagon Standoff: A Glimpse of the Future
The recent dispute between Anthropic and the U.S. Department of War highlights the ethical and control concerns surrounding AI. Hundreds of employees at Google and OpenAI have publicly supported Anthropic’s stance against unrestricted military access to its technology, particularly regarding mass surveillance and autonomous weaponry. This demonstrates a growing awareness within the industry of the potential risks and the necessitate for responsible AI development.
Pro Tip
Focus on building direct relationships with your customers. Owning your distribution channel will be critical in a world where AI can easily replicate functionality.
FAQ
Q: Will all software companies be affected?
A: Yes, any company offering digital services or software products is likely to feel the impact of this shift.
Q: What can software companies do to adapt?
A: Focus on building strong customer relationships, owning your distribution, and differentiating through unique value propositions beyond just the core functionality.
Q: Is this shift happening immediately?
A: The changes are expected to accelerate over the next few years, with significant impacts becoming visible by 2028.
Q: What is “vibe coding”?
A: Vibe coding refers to the process of generating software through AI prompts, often requiring a human in the loop for refinement and security.
Did you know? The cultures at Google and Anthropic are dramatically different, with Google being driven by engineering excellence and Anthropic prioritizing philosophical considerations.
Explore more articles on the future of AI and its impact on the tech industry. Subscribe to our newsletter for the latest insights and analysis.
Keep reading