Big AI labs eat the original playbook – here’s where innovators can still compete
REDMOND, Wash. – Startup founders used to worry that tech giants would make their product obsolete. Now it’s even harder: AI labs don’t just enter the startup space, they also provide tools for customers to try DIY solutions themselves.
The question of where startups can find a profitable niche was a recurring theme this week at the Technology Alliance Seattle Investor Summit+Showcase, held at Microsoft headquarters.
“Anthropic and OpenAI are different little animals,” Anthos Capital’s Bryan Hale said during an afternoon investor panel. “I don’t think people appreciate the speed at which you have to go to overtake the big monsters that eat you.”
Hale recalled a time when founders feared Andy Jassy would take over Amazon Web Services: Rename the division and announce 100 new AWS services, one or two of which inevitably sounded like their product. That went very well. Anxiety was overwhelming.
This time, he said, it’s different. Large companies are increasingly using coding tools like OpenAI’s Codex and Anthropic’s Claude Code to build what they often buy as bundled software. The result is that many of the startup applications that were once sold as subscriptions – sales automation, lead scoring, demand planning, accounting, etc. – now they are built in the blink of an eye.
Hale, who was an early executive at Seattle startup Chef and then the AI2 Incubator, said the pace at which you need to survive has completely changed. “If I could take my clock speed as a good but not good worker from 10 or 15 years ago, I would be smoking today,” he admitted.
AI2 Incubator’s Yifan Zhang, a two-time co-founder of startups GymPact and Loftium, said the toughest problems still reward patience, deep expertise, and a willingness to grind.
If the code is cheap, he said, “to stand out, to be distributed, to attract attention, is more difficult than before.” But he cited portfolio companies in mining, shipping containers, and immigration — domains where founders have a reputation for selling to specialty industries.
Mia Lewin of TheFounderVC said the formula starts with choosing the right target — a sword small enough that AI labs won’t bother chasing it. Since then, innovators need in-depth industry knowledge that allows them to create better products quickly without common mistakes, and a data pipeline that includes personalization or reinforcement learning.
“If you create a flywheel of rapid growth, it’s hard for them to catch you,” he said.
Madrona’s Tim Porter said the concern is real but oversold. He cited legal tech startups Harvey and Legora, which have continued to grow rapidly even after Anthropic launched its AI for the legal profession. Lawyers care about accuracy, workflow integration, and avoiding blind spots.
“They’ve crushed it, they keep growing faster and faster,” Porter said, explaining that the importance of understanding a particular domain doesn’t stop even with advanced tools.
Earlier in the day, OpenAI’s Vijaye Raji offered a view from the other side of the field. Raji, CTO of applications maker ChatGPT and Codex, told the same audience that founders have faced this fear throughout the platform era and won.
He agreed that the key is domain technology. Big platforms chase the biggest area they can reach. “There’s a lot of innovation that can be done at the edge of your domain with AI,” he said.
In fact, Raji highlighted one possible startup. Within OpenAI, the proliferation of AI-generated code has outpaced the company’s testing and deployment plans. “The problem is no longer generating codes,” he said. “The problem now is all the test tools, all the build tools.”
When he and his former Facebook colleagues started Statsig in the Seattle area in 2021, investors warned him that AWS would inevitably develop a similar software tool for AB testing. “If every startup decided to think that,” he said, “no SaaS company would ever exist.”
OpenAI acquired Statsig in September 2025 for $1.1 billion.
