🐋 OpenAI vs. DeepSeek

It’s Thursday!

Sam Altman is set to unveil new AI tech to D.C. leaders this week, aligning OpenAI with Trump’s administration, while SoftBank is reportedly in talks to invest up to $25 billion in OpenAI.

1: OpenAI accuses DeepSeek of using its models as AI battle escalates

The AI race just got messier. OpenAI is investigating whether Chinese startup DeepSeek improperly used its models’ outputs to train its own system—raising concerns over IP theft and the future of AI competition.

Why it matters: DeepSeek has shocked the industry by developing a low-cost, high-performance AI model rivaling OpenAI’s o1. If it relied on OpenAI’s outputs, it could signal a major shift where AI companies struggle to protect their tech.

What OpenAI is saying

• It has seen signs of “distillation”—where smaller models learn from the outputs of more advanced ones.

• “We take aggressive countermeasures to protect our technology,” OpenAI said, pledging to work with the U.S. government.

• Trump’s AI czar hinted at a crackdown on “copycat models.”

Altman: ‘DeepSeek is impressive, but we’ll deliver better models’

Despite the accusations, OpenAI CEO Sam Altman called DeepSeek’s model “impressive” but insisted OpenAI will “deliver much better models” soon.

Meanwhile, DeepSeek’s app surged to No. 1 on the App Store, though it’s been criticized for censoring politically sensitive topics.

U.S. national security concerns

The White House’s National Security Council is reviewing DeepSeek’s rapid rise, calling it a “wake-up call” for the U.S. AI industry.

• The administration may impose tighter restrictions on AI exports and collaboration.

• DeepSeek’s efficiency challenges the U.S. strategy of limiting chip access to slow China’s AI progress.

What’s next? DeepSeek’s breakthrough has already triggered a $600 billion sell-off in Nvidia stock and rattled the AI industry. With IP battles, regulation debates, and geopolitical tensions heating up, the AI war is only beginning.

Presented by Dell & Nvidia

Everything Edge AI — a live session by Dell and NVIDIA

Join Jags Kandasamy, CEO and Co-Founder of LatentAI, and Logan Lawler of Dell Precision AI for a live (and lively) discussion on February 6.

On the agenda: Edge AI solutions that enable efficient AI processing directly on devices, reducing the need for constant cloud connectivity, with optimized performance and minimal energy consumption.

This event is episode #6 of the Powering AI with Precision and NVIDIA webinar series, exploring how Dell Precision workstations equipped with NVIDIA Ada Generation GPUs are transforming local AI development.

Attend live for full experience

2: Trump’s cash freeze puts AI and tech funding in limbo

Trump’s sudden halt on federal grants has thrown key AI and tech projects into uncertainty. While a judge temporarily blocked the freeze, the damage is already done.

Why it matters

Tech firms rely on government-backed funding for AI research, semiconductor production, and cybersecurity. The pause could slow innovation and shake confidence in U.S. tech leadership.

What’s at risk

• AI research: The National AI Research Resource (NAIRR) faces uncertainty, with projects on AI efficiency and fairness now in limbo.

• Semiconductors: Efforts to boost U.S. chip production could be delayed.

• Telecom security: Programs replacing Chinese gear in rural networks are stalled.

The business impact

Startups and tech firms betting on government-backed AI infrastructure now face a tougher funding landscape. With grants becoming political, securing long-term investment just got more complicated.

Dig deeper

3: Alibaba claims its AI model beats DeepSeek

Chinese tech giant Alibaba just entered the AI arms race with Qwen 2.5-Max, a model it claims outperforms DeepSeek-V3, OpenAI’s GPT-4o, and Meta’s Llama-3.1-405B.

Why it matters: DeepSeek’s low-cost, high-performance AI shook Silicon Valley and sent tech stocks tumbling. Now, even China’s biggest tech firms are scrambling to keep up. Alibaba’s unusual decision to release Qwen 2.5-Max on Lunar New Year’s Day underscores the urgency.

The AI showdown

• DeepSeek’s meteoric rise forced domestic rivals into rapid upgrades.

• ByteDance (TikTok’s parent) recently updated its AI, claiming it outperforms OpenAI’s o1.

• Alibaba now claims Qwen 2.5-Max beats them all.

The bigger picture: DeepSeek’s shockingly low development costs have made investors question whether U.S. tech giants are overspending on AI. Meanwhile, competition within China’s AI scene is heating up, suggesting the country’s biggest players won’t let DeepSeek dominate without a fight.

What’s next? With AI advancements accelerating faster than expected, the real question is: Who will win the race—and at what cost?

Dig deeper

More news

• Google says hackers from China, Iran, and North Korea are using Gemini to boost productivity.

• Mark Zuckerberg said Wednesday it was too soon to say how advancements by DeepSeek, a Chinese startup, would impact Meta's heavy investments in AI.

• Snowflake in talks to acquire analytics startup Redpanda.

Editors picks ☕️

The Verge: This year’s Super Bowl will be full of AI ads.

The Information: Meta and Microsoft Show AI Bets Still in Play.