Nvidia Opens Its Vera Rubin AI Designs to Partners as Chip Stocks Slide

Nvidia used one of the tech industry's biggest hardware gatherings to show where AI data centers are headed next.

At the Open Compute Project Global Summit in San Jose, which drew a record of about 15,000 attendees, Nvidia announced on Oct. 12, 2026 that it is opening up its next-generation Vera Rubin rack and compute-tray designs to industry partners through its MGX architecture, according to the company.

What Nvidia announced

• Vera Rubin designs: Nvidia will share the blueprints for how its next-generation AI systems fit into server racks, so other companies can build compatible equipment.
• Partners: Cisco, Google, Meta, Microsoft and Visionbay are joining the effort to standardize AI data-center infrastructure as facilities scale "from megawatts to gigawatts."
• Faster connections: Nvidia and Cisco unveiled a new "Ultra-Dense Small Form Factor Pluggable" optical module, built with TE Connectivity and Amphenol, that supports 400 gigabits per second per lane and up to 12.8 terabits per second in total.

Intel and AMD also made competing data-center announcements at the event, according to CRN.

A tough week for chip stocks

The announcements came during a rough stretch for semiconductor stocks. The Philadelphia Semiconductor Index fell for a sixth straight session on Oct. 14, its longest losing streak in nearly two years. Investors were worried about how the massive AI buildout is being financed and about hot inflation data that raised the odds of higher interest rates.

That creates a tension investors are trying to sort out: companies keep committing to huge AI investments, while markets question whether the spending, and the debt behind it, will pay off.

What the Open Compute Project is

The Open Compute Project began in 2011 when Facebook shared the designs of its own data center hardware publicly, aiming to make server and data-center equipment more efficient and interchangeable across companies. It has since grown into an industry-wide group where major technology companies and hardware makers agree on common specifications, so a server rack or cooling system built by one company can work with equipment built by another.

That matters for AI because modern AI computing doesn't happen on a single chip. It happens across racks packed with many chips, high-speed networking and specialized cooling, all of which need to fit together physically and electrically. When Nvidia shares its Vera Rubin rack designs through MGX, it's letting other manufacturers build compatible power supplies, cooling systems and server chassis, which can shrink the time it takes data center operators to deploy new equipment.

Why the connections matter too

The new optical networking module Nvidia unveiled with Cisco addresses a less visible but equally important bottleneck: moving data between thousands of chips fast enough that they aren't sitting idle waiting for information. As AI models get larger, the connections between chips increasingly determine how fast a data center can actually train or run them, which is why networking announcements get as much attention at events like this as the chips themselves.

For readers without a stake in chip stocks, the bigger takeaway is a familiar one in technology: individual companies compete fiercely, but much of the infrastructure underneath them, from shipping containers to Wi-Fi standards to these server designs, tends to get standardized over time because it's cheaper and faster for everyone when things are built to fit together.

What it means for investors

Nvidia is one of the largest companies in the stock market, so it's a big holding in most broad index funds. Its moves can swing the whole market.

If you own Nvidia or other chip stocks directly, a few things to keep in mind:

1. Chip stocks are cyclical and can fall sharply even when long-term demand looks strong.
2. Good product news doesn't always lift a stock if the broader market is worried.
3. Keeping a single stock to a modest share of your portfolio can help you ride out the swings.

The bottom line

Nvidia is pushing its next-generation Vera Rubin designs into the industry with help from Cisco, Google, Meta and Microsoft. But strong technology news ran into a market worried about AI spending and interest rates. For long-term investors, it's a reminder that great companies can still have volatile stocks.

Sources: https://blogs.nvidia.com/blog/ocp-global-summit-vera-rubin-scale-ai-factories/, https://nvidianews.nvidia.com/news/nvidia-and-cisco-introduce-new-pluggable-optics-form-factor-to-scale-ai-factories