PRODUCTS

6 AI trends that defined CES 2026

The Deep View was on the ground at CES 2026 to sort out the real AI advances from the marketing stunts and hardware slop. These were the ones that made the cut. 

1. AI got a speed boost and a price drop — Nvidia gave us the biggest surprise of CES by announcing (about six months ahead of schedule) the release of its next-gen Vera Rubin chip for AI training and inference. It can cut inference costs by up to 10x and train mixture-of-experts models with 1/4 as many GPUs as its current Blackwell chips. AMD announced that its AI chips have achieved a 1,000x performance boost over the past 4 years. —Jason Hiner

2. Robots are now 'Physical AI' — Tech firms are keen to give AI a physical form, and CES was all in on the trend. Nvidia was the star of the show, unveiling the Alpamayo self-driving platform, new Cosmos world models, and the GR00T humanoid models. Meanwhile, Franka Robotics displayed its AI-powered robotic arms; LEM Surgical unveiled a spine surgery robot; and AGIBOT showed off a dancing humanoid for its US debut. Don't get too excited about home robots yet, because enterprise is the physical AI cash cow. —Nat Rubio-Licht

3. Industrial AI enters its hype era  — What a tidal wave of industrial AI and enterprise AI announcements we saw at CES 2026. Siemens led the charge with its new digital twin technology that extends beyond product development. It can now use AI to simulate operations and become an engine for optimization and continuous improvement. From the new enterprise pavilion in North Hall came big AI announcements from Caterpillar, Oshkosh, Neural Concept, Lenovo, Hyundai, Hitachi, and others. This is clearly one of the fastest-expanding areas of CES. —Jason Hiner

4. Quantum wants to amp up AI — Quantum scored an entire pavilion called CES Foundry, dedicated to companies like D-Wave, Quantinuum and Quantum Computing Inc. (QCi), signaling that it's no longer a lab project. Mixed opinions abound as to when the tech will reach large-scale deployment, but experts agree that quantum will be a catalyst for AI. CES examples included QCi demoing its quantum machine learning technology for things like fraud detection, drug discovery and financial forecasting. —Nat Rubio-Licht

5. An unexpected winner in consumer AI — Audio emerged as the best consumer AI category of the show, with products often built on Gemini, ChatGPT and Claude models. Some that caught our eye include Elehear and Cearvol’s AI-powered hearing aids, Subtle Computing’s voice-isolating earbuds, and Gyges Labs’ note-taking voice ring and agent companion. The form factor presents a low barrier to entry, but the products will have to niche down to compete with Google, OpenAI, and other giants. — Nat Rubio-Licht

6. AI glasses spark copycat boom — AI glasses aimed at chipping away at Meta's 70% market share were everywhere at CES. Lenovo, XGIMI, Solos, Cellid, Vuzix, LLVision, Asus, and Rokid all introduced new products. Thankfully, none of them are pure knock-offs. Each has a unique focus, from live translation to industrial tasks to lightweight displays to gaming. The jury is still out on whether AI-on-your-face becomes a trend, but it's going to get plenty of chances in 2026. —Jason Hiner

TOGETHER WITH MIRO

This Free 3-Part AI Masterclass Is For You

The holidays may be over, but Miro is still doing their best Santa Claus impression and passing out gifts like, well, Santa Claus. The latest and greatest? This 3-part masterclass on how to transform your product development from the ground up with collaborative AI. 

Each of the 45-minute sessions dives into a new practical framework that will help enable your product team, along with live demonstrations to show these ideas at work – and it all takes place in Miro’s easy-to-master, AI-powered platform. 

Watch all 3 masterclass sessions from Miro right here.

MARKETS

China’s AI IPO wave tests a new growth model

China’s push to take its AI startups public is gaining momentum.

MiniMax Group, a Chinese developer of multimodal, large language models, raised $618.6 million USD in its Hong Kong IPO after pricing shares at the top of its range, Reuters reported. The company sold 29.2 million shares and plans to allocate most of the proceeds to research and development over the next five years. Trading is expected to begin this week.

The listing is one of the first major public debuts for a Chinese AI startup building frontier models. While U.S. players like OpenAI and Anthropic remain private, China is pushing its AI firms into public markets earlier. Hong Kong is becoming a key venue for Chinese AI companies due to its business-friendly environment, a shift unfolding amid U.S. export controls on Nvidia GPUs to train the models.

“The wave of IPO approvals does suggest a shift in accelerating AI startup development through capital market access,” Lian Jye Su, chief analyst at tech research firm Omdia, told Reuters in late December. “While the U.S. maintains a lead in frontier compute and model performance due to chip superiority, access to public funding helps China build a more self-sufficient AI ecosystem.”

The momentum started with chips. In January, Chinese chip firms Biren, OmniVision, and Shanghai Biren Technology all completed IPOs in Hong Kong and saw strong debuts, helping set a floor for investor appetite. Kunlunxin, Baidu’s AI chip unit spinoff, filed for a Hong Kong IPO that could raise up to $2 billion, while Montage Technology is preparing its own offering that could reach $1 billion.

Chinese chip firms are also turning to their home turf. In December, Moore Threads, often described as China’s Nvidia, and MetaX Integrated Circuits both held IPOs on the Shanghai Stock Exchange, drawing strong interest from retail and institutional buyers.

China’s AI IPO wave reflects a country trying to reshape how AI models get financed. By leaning more on public markets, Beijing is betting that domestic investors can fuel AI growth even as geopolitical pressure mounts. But taking young AI firms public raises expectations on returns that may be difficult to meet. How these companies perform under that pressure will show whether China’s approach becomes a long-term advantage or a short-lived burst of enthusiasm.

TOGETHER WITH EMERGENT

Spend $5. Build Software That Pays You Back

Most people use AI to experiment. A few use it to make money.

Emergent is for builders who want revenue, not demos. With a simple chat, build production-ready apps people actually pay for, SaaS, paid tools, internal utilities, or niche apps you can launch fast.

Build, test demand, charge users, iterate.

That’s why 2.5M+ founders use Emergent to turn ideas into revenue, faster than traditional MVPs.

Plans start at $5. Build now

ENTERPRISE

96% distrust AI code — most deploy it anyway

Companies are widely using AI to trudge through mundane coding tasks. The problem is developers often don’t trust the outputs. 

A survey of 1,100 developers from code review firm Sonar found that, while AI accounts for 42% of all committed code, 96% of developers don’t fully trust AI to generate code that’s functionally correct. Still, despite the distrust, developers are pushing the code forward: Only 48% reported that they check their AI-assisted code before they commit it to projects, according to the report. 

AI-powered coding has emerged as a major point of focus for investors: 

• In December, Swedish AI coding startup Lovable raised $330 million, bumping its valuation to $6.6 billion. 

• In November, AI coding firm Cursor raised a $2.3 billion series D round, bringing its valuation to $29.3 billion. 

And it makes sense why investors are throwing money at it: These coding tools are massively popular. Lovable hit $200 million in annual recurring revenue in November, while Cursor claims to have surpassed $1 billion in revenue and has drawn in more than one million users. And while Anthropic doesn’t reveal user numbers, the company’s ever-popular Claude Code tool reportedly attracted 115,000 developers and processed 195 million lines of code in just one week.

These tools stand to make developers far more productive, with one internal study from Anthropic finding that developers who used Claude Code saw a 50% gain in productivity. However, like any AI tool, coding assistants can mess up. Productivity gains from these tools are only going to actualize into returns if we can trust the outputs, or have foolproof systems to ensure that bugs don’t slip through the cracks.

This survey is just one small indicator of a much larger problem: even if people don't actually trust AI to be correct, many might not bother to check their work. While in software development, the outcome is buggy code that needs to be fixed or patched, the implications of this tendency are broader when applied to other industries or the general public. For instance, what are the consequences of a hallucinating AI in healthcare, finance or manufacturing? What are the ramifications of an AI-powered search engine serving up incorrect information or repeating misinformation? For it to be safe, AI adoption in any form will require a healthy dose of skepticism, double-checking, and a human-in-the-loop where the stakes are the highest. 

LINKS

• Dell admits consumers don't want AI PCs and goes back to XPS brand 

• Ford announces hands-free, eyes-off self-driving is coming in 2028

• OpenAI vs. Elon Musk lawsuit will now go to jury in March

• AI training data startup Protege raises $30M from a16z to grow data network

• Snowflake will acquire Observe, a plug-in for monitoring its databases

• Gmail AI Inbox: Google is bringing Gemini to your inbox to help you track and summarize emails and streamline searchers.

• Copilot Checkout and Brand Agents: Microsoft is rolling out agentic experiences for shoppers to purchase directly in Copilot and for brands to implement AI chats. 

• Tiiny AI: An AI supercomputer the size of a power bank lets you run 120B parameter models locally. 

• Spec Coding by Capacity: An AI app builder that asks all the right questions to define your app before it starts building for you.

The Forward Deployed Engineer isn’t another requisition—it’s embedded execution capacity.

With Athyna, you get Forward Deployed Engineers who don’t just write code—they embed into your organization, understand your technical priorities, and deploy solutions that ship now. Here’s what you get:

• Senior technical depth — Engineers with 5-10+ years in Python, TypeScript, React, AWS, Kubernetes, Terraform, and AI/ML frameworks.

• AI-powered matching — Identifies engineers aligned to your stack, architecture, and delivery requirements—no guesswork, no mismatches.

• 4-day turnaround — From brief to interviews, so you maintain delivery momentum instead of waiting on hiring cycles.

• $80K+ annual savings per engineer — Enterprise-caliber talent without U.S. market rates.

Explore talent that ships →

GAMES

POLL RESULTS

Are AI factories and data centers one of the top AI trends you're following in 2026?

Yes (47%)
No (30%)
Other (23%)