AI Models Get Faster in 2026 with New Hardware and Open Code

120 billion parameter AI models can now run on a single device, thanks to Microsoft's new Surface RTX Spark Dev Box, which packs Nvidia's Blackwell-architecture RTX Spark processor and 128 gigabytes of unified memory. This move directly challenges the per-token pricing model that has defined the AI industry's economics since ChatGPT launched three and a half years ago.

Meaningful Section Title

Microsoft's Surface RTX Spark Dev Box delivers **one petaflop of AI compute**, allowing developers to load, run, and interact with AI models exceeding **120 billion parameters** without sending a single API call to the cloud. In comparison, Apple's new Foundation Models announced at WWDC26 will include **five models**, some of which are local, some of which are cloud-based, and one of which lives in Google's servers running on Nvidia chips.

Subsection

• The global AI market is expected to reach **$190 billion by 2027**, growing at a CAGR of **34.6%** from 2020 to 2027, according to a report by **MarketsandMarkets**.
• **OpenAI** has also announced that its models and Codex can be accessed through Oracle Cloud, using existing commitments to build and deploy AI with enterprise security and governance.

"These class of devices, we think, will get to about **100 billion parameter model running," said Pavan Davuluri, Microsoft's executive vice president.

What the Sceptics Say

Some critics argue that the shift towards on-device AI could lead to **increased energy consumption** and **heat generation**, potentially negating the environmental benefits of cloud computing. Additionally, the high upfront cost of devices like the Surface RTX Spark Dev Box could be a barrier to adoption for many developers.

What This Means for the Industry

Companies like **Google**, **Amazon**, and **Facebook** will need to adapt to the changing landscape of AI computing, potentially shifting their focus towards on-device AI solutions. In the next **6-12 months**, we can expect to see more announcements from these companies regarding their AI strategies. For example, **Anthropic** has already apologized for the lack of transparency in its Claude Fable guardrails, highlighting the need for more open and explainable AI models.

Key Takeaways

1. Engineers: Start exploring on-device AI solutions like the Surface RTX Spark Dev Box to reduce cloud costs and improve model performance.
2. Investors: Look for companies that are investing in on-device AI solutions and open-source AI models, as these are likely to drive growth in the industry.
3. Business Leaders: Consider the potential benefits of on-device AI for your organization, including improved security, reduced latency, and increased model control.
4. Consumers: Expect to see more AI-powered devices and applications in the next year, with improved performance and reduced dependence on cloud connectivity.

Sources

• VentureBeat: Microsoft debuts Surface RTX Spark Dev Box to run large AI models without cloud costs
• OpenAI Blog: Access OpenAI models and Codex through your Oracle cloud commitment
• 9to5Mac: Apple’s new Foundation Models explained: on-device AI, cloud AI, and everything in between