Skip to main content

DeepSeek Founder on Pricing: Slashed Model Price by 75%, Profit Isn't the Goal

DeepSeek's founder Liang Wenfeng recently sat down with investors to talk pricing, and his message was clear: the company isn't in it for the money—at least not in the way you'd expect.

During a highly anticipated investor meeting, Liang laid out DeepSeek's strategy, which he says is built on earning "reasonable profits" rather than squeezing every last dollar. The core idea? Use extreme cost efficiency to unlock more powerful models, then let the technology speak for itself.

Image

One example Liang shared was a model that initially carried a high price tag—mainly because the team worried demand might overwhelm their servers. But then they did something bold: they slashed the price to just one-fourth of the original. "Given our limited computing resources, computational efficiency is what really determines how big a model we can train," Liang explained. Instead of throwing more hardware at the problem like many big companies do, DeepSeek chose to focus on efficiency first.

That efficiency extends to their entire operation. DeepSeek's team is surprisingly lean—no dedicated sales force, no customer service department. The product is designed to attract users on its own. "Our operational mechanism is extremely lightweight," Liang said. "We only need a small number of people to keep things running."

This approach puts serious pressure on competitors in the API market. By significantly lowering prices, DeepSeek is forcing others to rethink their own cost structures. But Liang insists this isn't about starting a price war for the sake of it. It's about making AI accessible.

When asked about commercialization, Liang was candid: DeepSeek has been pushing for commercial implementation, but it's not the endgame. "The time point for a complete shift to commercialization is still far off," he admitted. For now, the priority remains technological advancement.

DeepSeek's model is turning heads because it flips the traditional playbook. While many AI companies race to scale up computing power and chase revenue, DeepSeek is doubling down on efficiency and lean operations. This has already lowered the barrier to entry for AI, and it's pushing the entire generative AI industry away from blind resource accumulation toward a smarter focus on algorithm and architecture efficiency.

In a world where bigger often seems better, DeepSeek is proving that sometimes, less really is more.

Key Points

  • DeepSeek cut one model's price to a quarter of its original, prioritizing cost efficiency.
  • Founder Liang Wenfeng says the company aims for reasonable profits, not maximization.
  • The company operates with a minimal team, no sales or customer service staff.
  • DeepSeek's strategy is driving the AI industry toward efficiency over resource accumulation.