Unpacking StarGate AI $500 Billion MegaProject
What It Really Is, And What It Hopes To Be
Oracle CEO Larry Ellison and Open AI CEO Sam Altman joined our favorite Japanese Softbank Banker, Masa Masa Masayoshi, to announce a $500 Billion AI deal for US investment with Donald Trump, beginning with infrastructure in Texas with $100 Billion in immediate investment. Open AI and other AI competitors like Elon Musk’s Grok require huge clusters of GPU chips and energy to train AI models, and the StarGate deal set out the strategy to guarantee that infrastructure on US soil.
The Stargate press conference started well enough, with jobs for Americans, investment in Texas, national security, and leadership in AI. Then Oracle CEO Larry Ellison took to the podium. And Ellison dove into a pitch for personalized mRNA cancer vaccines developed by AI. The StarGate Starship news conference could be seen having an “unscheduled rapid disassembly’ event in a rainbow of colors streaking across the sky.
The press conference ended with one of those “elephant in the room” moments: Larry Ellison pitched AI for personalized mRNA vaccines for cancer. Ellison went from Warp Speed from providing 32,000 GPU clusters for Sam Altman and Open AI to train their AI models on straight to his Deep State agenda project of cancer mRNA vaccines.
Ellison’s float of the “personalized mRNA cancer vaccine by AI” was pilloried by the cognoscenti on X/Twitter, with many commenters immediately drawing parallels to the failure of mRNA vaccines with CoronaVirus. Trump supporters were left agog that Trump could be announcing support for mRNA cancer vaccines so early after winning reelection, with a large swath of his supporters being decidedly anti-mRNA.
The “elephant in the room” question on everyone’s mind was, “Why are you announcing mRNA cancer vaccines so soon?”. Many pundits wondered aloud that the new mRNA cancer vaccines were a fix for the cancer spike caused by the mRNA COVID vaccines. “You mean for the cancers caused by mRNA?” one internet commenter wrote. Still, the AI CEOs plunged ahead, possibly unveiling a decade-long Deep State Cancer Genome Atlas project as the key reason for the new investment in AI infrastructure.
The devil is always in the deals for these mega announcements, but I will try to unpack facts from fiction with what we know. First we know that Open AI was looking for more GPU compute resources than Microsoft could provide since last June when the Oracle - OpenAI compute resources deal was announced.
https://time.com/7209167/stargate-openai-donald-trump/
Trump also hinted at emergency authorization for energy to quickly supply the datacenters in Abilene, Texas with energy. Since the Oracle datacenter are located near the gigantic Hayes oil and gas field, we think this meaning fracking the existing oil fields to provide gas turbines to power the data centers. We just don’t have those details at this time.
Trump’s forthcoming “emergency declarations” in the coming day should shed light on how the US Government plans on helping the effort to enable the Oracle - Open AI datacenters.
My initial reaction to Larry Ellison’s cancer mRNA moment is he was repacking the Abu Dhabi investment firm MGX’s commitment to building AI GPU-based infrastructure and into healthcare, which was recently announced in Davos 2025. Ellison’s deep state connections run deep with AI resources to provide cancer mRNA vaccines, along with former CIA Director John Brennan and Dr. Patrick Soon Shiong of San Diego. We believe MGX from UAE is the front buyer for the CIA interests in mRNA.
https://www.cnbc.com/2025/01/20/abu-dhabi-330-billion-wealth-fund-warns-on-ai-disruption.html
The initial AI data centers for the StarGate MegaProject will be in Abilene, Texas, and Larry Ellison insists the construction has already started (perhaps a few shovelfuls of dirt were turned before his jet flew to Washington, DC, for the Stargate press release).
The chip technology appears to be ARM’s new GPU chip designed in partnership with Nvidia, but British chip maker ARM has many shipped GPUs for mobile devices and doesn’t currently have a server configuration. We will be providing more information as these plans become more concrete. Our researchers do believe a steal project with ARM and GPU clusters has been ongoing at Oak Ridge National Laboratories since 2019.
Journalist George Webb at the Supercomputer Center of Oak Ridge National Lab in 2022.
However, I believe a semi-stealth project was going on at Oak Ridge National Lab in 2019 to develop an ARM GPU server configuration, and that project seems to be more public now.
https://www.hpcwire.com/2019/06/17/jack-wells-joins-openacc-arm-support-coming/
The Oak Ridge project may be the “secret sauce” Larry Ellison brings to the StarGate deal. Cancer vaccines have long been a supercomputing target for Oak Ridge National Labs.
https://www.hpcwire.com/2019/06/17/nvidia-recommits-to-arm-seeks-to-accelerate-all-cpu-architectures/
This doesn’t mean that Grok and Anthropic will be excluded from the new ARM-Nvidia GPU supercomputing technology since both organizations are large customers of Nvidia. However, perhaps Larry Ellison and Oracle have an inside track on the Oak Ridge ARM-Nvidia processor clusters intended for Abilene.
Microsoft maintained it has not been crowded out of the processing supplier for Open AI, reaffirming its 80 Billion dollar commitment to build AI infrastructure this year. Microsoft CEO Nardella firm’s commitment may be the only solid commitment for the coming year for $80 Billion, although the consortium only appeared to have $20 Billion committed from UAE investor group MGX from Davos this week. Ellison seemed anxious to quickly parlay his inside track to ARM-Nvidia GPU cluster technology to MGX UAE investors in Davos this week.
Continued in Paid Section