Venture capitalist and renowned solo investor Elad Gil, a figure whose reputation in the startup ecosystem has reached near-legendary status, took the stage at TechCrunch Disrupt to share his reflections on the turbulent and transformative evolution of artificial intelligence. According to Gil, the extraordinary acceleration of AI represents one of the least predictable surges in technological progress he has ever witnessed. Given his extensive track record—his name appears on the cap tables of nearly every major success story of the past decade, including many of the AI powerhouses now defining the industry—his perspective carries particular weight.
Despite his deep involvement across the sector, Gil emphasized that the past year has brought striking consolidation within certain branches of the AI landscape. In his view, a few core markets now appear largely dominated by clear frontrunners, leaving limited room for newcomers to meaningfully disrupt them. Yet, beyond those tightly controlled territories, an immense frontier of opportunities remains unclaimed, representing open ground for ambitious innovators with new ideas and bold execution.
Reflecting on his early entry into the space, Gil recalled that he began investing in generative AI technologies back in 2021—a period when only a small subset of the tech community recognized the transformative potential of large-scale language models. His interest was sparked by what he described as a staggering leap in capability between two key milestones: the release of GPT‑2 in 2019 and GPT‑3 in 2021. The improvement between these iterations, he observed, was so pronounced that any extrapolation of the underlying scaling trends pointed to one unavoidable conclusion—AI was on the verge of becoming central to nearly every digital domain. This conviction led him to focus his capital and insight on early-stage startups that harnessed large language models (LLMs) for novel applications.
Gil’s portfolio quickly came to encompass both the builders of foundational models—organizations such as OpenAI and Mistral—and a diverse set of application-layer innovators, including Perplexity, Harvey, Character.ai, Decagon, and Abridge. Over the following years, particularly throughout 2024 and into 2025, AI development accelerated at an almost bewildering pace. Each successive model release seemed to rewrite the boundaries of what machines could achieve, forcing investors and entrepreneurs alike to recalibrate their understanding of the field every few months.
As he reflected, Gil admitted that AI was unusual even by the standards of disruptive industries. “It was the one market,” he said, “where the more knowledge I gained, the less certain I felt.” In most areas of technology, experience yields predictability: deeper understanding leads to better forecasting and more confidence in strategic bets. But AI defies that logic. Its rapid evolution, combined with unpredictable emergent properties, creates persistent opacity—a landscape where clear foresight is exceptionally rare. Gil noted that even today, several of the most dynamic AI segments remain enveloped in that same atmosphere of uncertainty.
At the same time, he acknowledged that some domains are now showing unmistakable signs of stabilization around dominant players. The clearest case, in his view, concerns the foundational models themselves. Although hundreds of organizations worldwide have developed or are attempting to develop their own large-scale models—with nations such as South Korea investing heavily in sovereign alternatives—a small number of global leaders have already emerged as the prevailing forces. Gil predicted that companies such as Google, Anthropic, OpenAI, xAI, Meta, and potentially Mistral will likely remain at the forefront of this race, cementing their positions as the architects of core AI infrastructure.
Another market where leadership appears increasingly entrenched is AI-driven software development. Gil observed that the rapid rise of AI-assisted coding has given certain players an early and possibly enduring advantage. Foundational model creators themselves, including Anthropic with its Claude Code system and OpenAI with Codex, have entered the space alongside startup innovators like Anysphere’s Cursor and Cognition’s Devin—both of which, he argued, have established formidable lead positions. Even among new entrants with strong funding, such as Magic—an entity Gil described as a potential outlier—and Poolside, catching up with these leaders remains a daunting challenge.
The healthcare technology sector also shows signs of consolidation, particularly within medical transcription and clinical documentation. Companies like Abridge have risen as front-runners, while others, including Ambiance, hold important yet smaller positions in the same niche. Similarly, customer engagement and support—an early proving ground for both traditional automation tools and modern AI agents—has become another battleground defined by a few major players. Gil pointed to his own portfolio company, Decagon, which recently secured $131 million in funding at a $1.5 billion valuation, as evidence of sustained investor confidence. Competing ventures, such as Sierra (founded by OpenAI chairman Bret Taylor), and entrenched incumbents including Salesforce and HubSpot, continue to enhance their AI capabilities in pursuit of greater efficiency and customer satisfaction.
When asked where untapped potential still lies, Gil referred to several promising areas that remain relatively open-ended. He mentioned financial technology and accounting resources powered by AI, the critical and still-emerging field of AI security, and a broader range of adjacent markets characterized by their inherent business attractiveness but lingering uncertainty regarding which team or product will eventually dominate them.
Interestingly, Gil cautioned that in today’s climate, rapid revenue growth no longer guarantees long-term success in the way it once did. The current moment, he explained, is marked by a collective urgency across the corporate world: CEOs of major enterprises are issuing top-down mandates for their teams to formulate robust AI strategies. This widespread directive has led to an environment where enormous companies are experimenting with tools and platforms that they would have ignored only a few years ago, purely because the transformative promise of AI demands attention. As a result, early-stage startups entering new AI markets can now attract significant enterprise revenue almost immediately—yet such early enthusiasm may prove ephemeral.
According to Gil, these dramatic influxes of customer interest must be interpreted carefully. True endurance can only be measured after the initial testing period has passed, once customers decide whether a product genuinely integrates into their long-term workflow or fades as a temporary curiosity. “There’s false signal,” he warned, “and then there’s the rare case where something truly works.” One company he cited as clearly belonging to the latter category is Harvey, the legal AI platform that he identified as a standout example of authentic market traction. In 2025 alone, Harvey completed three major fundraising rounds in rapid succession, escalating its valuation from $3 billion to $5 billion, and then to an impressive $8 billion within only a few months—a testament to both investor conviction and tangible user adoption.
In summarizing his outlook, Gil left the audience with a portrait of the AI ecosystem as both exhilarating and humbling. Certain territories, he argued, are already spoken for, dominated by powerful incumbents whose advantages will be difficult to dislodge. Yet surrounding those strongholds is an immense expanse of opportunity, a dynamic environment still open to explorers willing to navigate its risks and embrace its uncertainty in pursuit of the next defining breakthrough.
Sourse: https://techcrunch.com/2025/11/03/elad-gil-on-which-ai-markets-have-winners-and-which-are-still-wide-open/
