Last week, more than 3,000 participants, myself included, flocked to Las Vegas for the highly anticipated HumanX AI conference. Over three days, one theme echoed across every panel, fireside chat, and product launch — trust in AI. In a world where artificial intelligence continues to evolve rapidly, industry leaders grappled with a pressing question: how can we build trustworthy outcomes from a technology rooted in probabilities?
Ironically, this conversation unfolded just as Apple delayed the release of its latest AI features, citing accuracy concerns. At the same time, one glaring statistic from an AWS report lit up the conference screens — only 6% of AI projects actually make it to production. It was a blunt but necessary reminder of how experimental — and challenging — the AI race has become.
Still, the momentum is undeniable. According to a joint report from HumanX and Crunchbase unveiled during the conference, the AI sector attracted more than $100 billion in investments in 2024, marking a staggering 80% jump from 2023. This wave of capital signals just how fiercely competitive the landscape has grown.
The expo hall buzzed with energy, offering a mix of product demos, Q&A sessions, and networking lounges. A dedicated app made it easy to connect, creating countless opportunities to meet founders, investors, and engineers reshaping the future of AI.
Yet, what truly stood out were the candid insights from leaders at OpenAI, Anthropic, and Mistral — companies shaping the very models driving this AI revolution, just two years after ChatGPT’s breakthrough moment.
OpenAI Reflects on Its Edge in a Crowded Market
Despite OpenAI’s jaw-dropping $157 billion valuation, even they couldn’t escape the tough questions. CNBC’s Kate Rooney asked the one thing everyone wanted to know: “What’s your moat?”
Kevin Weil, OpenAI’s chief product officer, didn’t sugarcoat it. He admitted the days of having a 12-month advantage were gone. Now, any lead is reduced to a tight three to six months — but even that, he argued, is still immensely valuable in today’s fast-evolving AI landscape.
Weil captured the sentiment of the moment perfectly. “Every two months, some new model emerges doing things computers have never done before,” he said. Gone are the days of simple database upgrades — AI development now moves at breakneck speed.
Still, OpenAI’s scale is staggering. Weil revealed that over 3 million developers currently build on their API, while 400 million people use ChatGPT weekly. Even more impressive, over 2 million businesses have integrated their enterprise products, highlighting the company’s growing foothold.
Anthropic’s Claude Code and the Future of Human-AI Collaboration
One of the conference’s most engaging conversations featured Mike Krieger, Anthropic’s chief product officer, in discussion with The Verge’s Alex Heath. Krieger pulled back the curtain on Claude Code, Anthropic’s newly launched coding assistant, which hit 100,000 users in just a week.
Krieger didn’t shy away from the tension brewing between API providers and third-party developers. Before launching Claude Code, he reached out to major partners — including Anysphere (Cursor), Codeium’s Windsurf, and GitHub’s Copilot — since Anthropic was now competing directly with them.
But for Krieger, shipping a first-party product wasn’t just about competition. It was about learning. “You can’t gather the feedback you need if you’re only an API provider,” he explained. Every user interaction helps refine the model, making it better for everyone.
On a deeper note, Krieger shared his personal motivation for joining Anthropic — shaping the future of human-AI interaction. “If a year from now all we’ve built are chat boxes and chatbots, we’ve failed,” he added. His vision clearly extends beyond simple conversational AI.
Mistral’s Open-Source Bet and the Rise of Smaller Models
While OpenAI and Anthropic focus on scale, Mistral AI is betting on openness and decentralization. The France-based startup champions open-source AI models to prevent a handful of tech giants from dominating the space. According to co-founder and CEO Arthur Mensch, there’s massive demand from enterprises and governments seeking better control over their data.
Beyond open source, Mistral’s strength lies in its smaller, nimble models — perfect for real-world robotics. “Vision-to-action models running on specific hardware will be crucial in the coming years,” Mensch noted. Mistral is already proving that point through partnerships with European defense startup Helsing for drone systems and multiple robotics companies in the Bay Area.
Though Mistral started as an enterprise-first player, Mensch said the line between enterprise APIs and consumer products is thinner than ever. Case in point: Mistral’s consumer chatbot, Le Chat, went live last month.
HumanX 2025 Moves to San Francisco — Where Deals Are Brewing
With so much money and momentum flowing into AI, it’s no surprise that the next HumanX AI conference is heading to San Francisco — the heart of the AI gold rush. The shift feels strategic, especially as insiders predict that 30% of companies presenting this year could be snapped up by acquirers within the next 12 months.
By this time next year, the AI landscape — and the HumanX stage — could look very different. But one thing is certain: as generative AI matures, the race to build trust, deliver real-world applications, and define the future of human-AI collaboration is only just heating up.
