Opinion
The moves by AI giants Google and OpenAI challenge startups - and offer a massive opportunity
"Stop trying to build the next ChatGPT; that race is over. Instead, use open-source models as a base and focus on building dedicated, regulation-compliant infrastructure in fields such as law, finance, and healthcare," writes Lee Ben-Gal, an Investor at lool ventures.
In the world of technology investing, we are accustomed to experiencing seismic shifts. But the most significant tectonic change today isn't driven by a new technological development, but by a series of quiet, strategic moves that are changing the rules of the game. We are witnessing the forced maturation of the AI market: the "Wild West" era, which was open to all, is coming to an end. The giant companies that dominate the market, Google and OpenAI, are beginning to build their walled gardens, thereby setting clear boundaries.
It started with what looked like a small adjustment but marked the beginning of a major shift. Google, by quietly removing the &num parameter from its search URLs, eliminated the long-standing ability to request hundreds of search results at once. This is not an innocent technical update; it constitutes the closing of a key oxygen pipeline. For years, this capability served startups and AI researchers in collecting large-scale data to train artificial intelligence models. Now, what once required a single request now requires dozens, making data collection more expensive, slower, and riskier.
By doing so, Google is leveraging its search monopoly to assert its dominance in AI. By cutting off easy access to data collection, it has positioned itself as the sole gatekeeper of information. This means Google isn’t just shaping what we find; it’s also shaping what future AI systems are able to learn. In effect, it now controls the most crucial resource in the AI economy: data.
This trend, which began with Google building a wall around its data, has continued with OpenAI, now starting to build a wall around the data's output—the applications. Recently, OpenAI tightened its rules and prohibited ChatGPT from providing personalized advice in sensitive fields, such as health, law, and finance. Ask for an analysis of an X-ray or a legal opinion, and the model will refer you to a professional. The new policy explicitly prohibits providing customized advice that requires a license.
On the surface, this is bad news for entrepreneurship. The two most prominent players are blocking access— one cutting off data, the other restricting access to the most valuable applications.
It is easy to see this as the end of the innovation era, but as an investor, I see the complete opposite. We’re not watching the market close in on startups; we’re watching it reshape itself. These moves are a clear indication that the tech giants are engaging in strategic risk management. They understand that they are too large to operate in gray areas that require strict regulation and professional liability. They are afraid, and justifiably so, of regulators and lawsuits.
In their retreat, OpenAI and Google are effectively drawing new boundaries and creating what is known as an enormous, well-defined "white space" for startups. They are telling us: we will take the general, mass-market, and safe territories. You, the entrepreneurs, take the vertical, complex, and profitable markets. Here lies the next big opportunity in Vertical AI—dedicated artificial intelligence models built to operate within strict regulatory frameworks.
While ChatGPT retreats from giving medical advice, a startup that builds a dedicated, HIPAA-compliant model, complete with human-in-the-loop mechanisms and partnerships with licensed professionals, can conquer the HealthTech market. Regulators will never trust a giant, general-purpose "black box," but they will trust a small, transparent, and auditable system built for one specific purpose.
As investors, we will need to re-evaluate the risk of data access, as well as the added value startups gain by embedding accountability and compliance into their products.
The message to entrepreneurs is clear: Stop trying to build the next ChatGPT; that race is over. Instead, use open-source models as a base and focus on building dedicated, regulation-compliant infrastructure in fields such as law, finance, and healthcare. While the tech giants flee from liability, the startups that embrace it will win. The new boundaries that have been set are a map to the next territories for them to conquer.
Lee Ben-Gal is an Investor at lool ventures, specializing in early-stage startups.
