Good morning. Lean and mean today, like Nathan Bateman’s dance moves.
Today’s tech news below. Have a wonderful weekend. —Andrew Nusca
Want to send thoughts or suggestions to Fortune Tech? Drop a line here.
China's DeepSeek quietly releases a rival to GPT-5
Chinese AI startup DeepSeek shocked the world in January with an AI model, called R1, that rivaled OpenAI’s and Anthropic’s top large language models at a fraction of the cost.
Now, just two weeks after OpenAI debuted its latest model, GPT-5, DeepSeek is back with an update to its flagship V3 model that experts say matches GPT-5 on some benchmarks—and is strategically priced to undercut it.
The debut of DeepSeek’s new V3.1 model touches several of today’s biggest AI narratives at once.
DeepSeek is a core part of China’s broader push to develop, deploy, and control advanced AI systems without relying on foreign technology. While U.S. companies have been hesitant to embrace DeepSeek’s models, they’ve been widely adopted in China and increasingly in other parts of the world.
DeepSeek’s new release, coming just after OpenAI’s GPT-5—a rollout that fell short of industry watchers’ high expectations—underscores Beijing’s determination to keep pace with, or even leapfrog, top U.S. labs.
What makes the new model notable is how it was built, with a few advances that would be invisible to consumers. But for developers, these innovations make V3.1 cheaper to run and more versatile than many closed and more expensive rival models.
For instance, its “mixture-of-experts” design means only a fraction of the (otherwise huge) model activates when answering any query, keeping computing costs lower for developers.
The new model also combines both fast answers and reasoning in one hybrid system—something that few open-weight models have been able to do thus far. —Sharon Goldman
Google claims Gemini AI now 33x more energy-efficient
Google has released a detailed methodology to measure the energy, water, and carbon footprint of its AI models, addressing a critical gap in understanding AI’s environmental impact.
The company claims to have achieved major efficiency gains for its own AI. Over 12 months, the median energy use per Gemini Apps text prompt fell 33x, and carbon emissions dropped 44x, while still improving output quality, says Google.
To put it in perspective, each prompt now uses less energy than nine seconds of TV.
With a 12% reduction in data centre emissions despite rising electricity demand, Google claims it is a "milestone" in sustainable AI development and energy transparency.
Google says it measures improvements by taking various factors into account, including total energy footprint, CPU and RAM usage, power usage effectiveness (PUE) of the data center, and data center water consumption.
Google has been focused on improving data centre efficiency for a long time.
In 2024, for example, it reduced data centre energy emissions by 12% even as electricity consumption grew by 27% year-over-year on the expansion of its business and services.
The company made its latest achievements thanks to improvements at every layer of AI, including more efficient model architectures, efficient algorithms and quantisation, inference and serving, custom-built hardware and ultra-efficient data centres.
The search engine giant says the aim of its study is to drive industry-wide progress toward more efficient AI. —Manoj Sharma
How Palantir became the S&P 500’s best and worst stock of 2025
Palantir Technologies has created one of the most dramatic stories on Wall Street this year, defying conventional investment narratives.
In 2025, it became the top-performing stock in the S&P 500, surging over 106% and at points climbing 144% from the start of the year—outpacing even AI heavyweights like Nvidia.
This explosive growth was fueled by its robust financial performance, notching its first billion-dollar quarter with momentum from government and commercial AI contracts.
However, Palantir’s meteoric rise has been followed by a brutal reversal.
Over the past six trading sessions, Palantir shares plunged more than 17%, wiping out $73 billion in market capitalization and marking the largest drop since April.
In recent days, Palantir has also been the worst performer in the S&P 500.
Palantir’s dramatic stock moves followed fresh fire from short-sellers, particularly Citron Research, led by Andrew Left. In a scathing report, Citron argued that Palantir’s stock was detached from business fundamentals and sound analysis.
Citron’s thesis is that OpenAI, widely recognized as the leader in AI, is about to receive a $500 billion valuation, with projected revenue of $29.6 billion in 2026, resulting in a price-to-sales ratio of nearly 17.
By contrast, Palantir is forecasted to deliver $5.6 billion in revenue in 2026. Applying OpenAI’s valuation multiple to Palantir would yield a stock price of just $40.
Short-sellers like Left insist that Palantir’s business isn’t as scalable or as subscription-based as Wall Street prefers, in stark contrast to OpenAI. Palantir’s dependence on government deals introduces uncertainty and volatility.
It all leads Citron to claim that the stock is unjustifiably expensive even after recent losses—a study, if you will, in market euphoria versus valuation reality. —Nick Lichtenberg
More tech
—Apple HR drama. Nine current and former staffers accuse VP of Fitness Tech Jay Blahnik of inappropriate and retaliatory behavior.
—Target CEO Brian Cornell steps down. Intern-turned-COO Michael Fiddelke takes the retailer’s reins on Feb. 1.
—Trump taps Airbnb cofounder. Joe Gebbia will be the nation’s first chief design officer.
—Russia forces chat app installations. The state-backed messenger app MAX must be preinstalled on new phones and tablets from next month.
—Meta moves to Google Cloud. A reported $10 billion deal for servers, networking, storage, and more.
—Zoom shares jump 6% thanks to a sunnier-than-expected 2026 outlook.
—“Bell to bell” bans catch on. Seventeen U.S. states institute policies to curb students’ smartphone usage.
—Rent the Runway hands over the keys. Exchanges a controlling stake for the elimination of $240 million in debt.
—ByteDance releases open source AI model. After all, why not? Why shouldn’t I?
Endstop triggered
