The Cheap AI Era Ends: DeepSeek Raises Prices Sharply as Apple Builds a China-Only Model With Alibaba

The era of ultra-cheap artificial intelligence appears to be drawing to a close, as the Chinese labs that spent the past year competing to offer the lowest prices are now busy raising them, a reversal driven by the rising cost of running these systems at scale.

At the center of the shift is DeepSeek, the lab that became famous precisely for its low prices, which has sharply revised its rates with a new model and pushed the rest of the industry to confront a very different economic reality.

Prices climb across the board

With the release of its V4-Pro model, DeepSeek raised some of its programming interface prices by as much as one thousand one hundred percent, with the new rates going live around August 16 and forcing developers who built on cheap access to rethink their budgets.

The company also introduced a split between peak and off-peak pricing, a design meant to push work into the quieter hours of the day, while the cost of processing during busy periods climbed steeply above the levels customers had grown used to.

For builders who had structured their businesses around rock-bottom prices, the increases are a direct hit to their margins, and many are now weighing whether to absorb the cost, pass it to customers, or redesign their products around cheaper approaches.

Not just one lab

The trend reaches beyond a single company, because other Chinese firms including Alibaba and Moonshot AI are also raising prices or introducing paid tiers, as surging infrastructure costs make low-price strategies increasingly difficult to sustain over time.

The broader picture is an industry shifting from a price war toward differentiation, after a stretch of aggressive low-cost competition that had made access to capable models nearly free for many users and startups across the region.

That transition mirrors what tends to happen in many technology markets, where the early land-grab phase eventually gives way to a period of sustainability, with companies forced to charge what it actually costs to keep their systems running.

Apple bets on Alibaba

Against this backdrop, Apple has trained a large language model built specifically for China with help from Alibaba, a notable move that marks a break from simply leaning on partners to power its artificial intelligence features in that market.

Taken together, the price increases and Apple's partnership point to a maturing market, one where the cost of intelligence is becoming visible again and where success depends less on being cheapest and more on being distinct, reliable and fast.