Breaking

Meta and xAI Lead on Data Centers While Lagging in AI Models

July 7, 2026 at 06:05 EDT

Infra & Chips
Foundation Models

Bloomberg Opinion reported on July 7, 2026, that Meta's Mark Zuckerberg and xAI's Elon Musk trail OpenAI and Anthropic in AI model development while leading rivals in the scale and speed of their data center and power infrastructure buildouts. Columnist Parmy Olson argued the two sit in a contrasting position of "failure in models" and "success in infrastructure."

May 14, 2026 · AI Infrastructure

Losing the Model Race, Winning the Concrete One

Zuckerberg and Musk trail OpenAI, Google and Anthropic on frontier models — but are pouring billions into the fastest, largest, cheapest-powered AI data centers on earth. Compute and power, not model quality, are becoming the true crux of the race.

$700B+

Projected combined Big Tech AI spending

$115–135B

Meta's 2026 capex, centered on data centers

122 days

To stand up 100,000 GPUs at xAI's Colossus

Power Ambition — to scale (MW / GW)

Columns proportional to peak-power target. Both firms put scale and cheap power first.

300MW+

Colossus (xAI)

current, Memphis TN

2GW

Hyperion initial

Meta, Richland Parish LA

5GW

Hyperion target

Meta full buildout

Colossus · xAI

Memphis, Tennessee

200,000+ GPUs (H100 / H200 / GB200)
Former Electrolux plant, converted
194 PB/s memory bandwidth, 1 EB+ storage
Full capacity leased to Anthropic (Claude)
Gas turbines for instant power — permitting scrutiny

Hyperion · Meta

Richland Parish, Louisiana

~4 million sq ft campus
2GW initial power, 5GW target
Own training / inference clusters
$875M+ in local contracts in year one
Large enough to reshape a regional economy

Praised

Build speed unusual even by industry standards. Leasing Colossus fully to Anthropic proves real demand; Hyperion is an economy-scale bet.

Contested

Gas power and strained grids raise environmental concerns and local opposition. On models, Grok and Llama still trail Claude / GPT / Gemini.

The scramble for power, land and cooling — more than which model wins — is becoming the true crux of the AI race.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.