Discussion

Berkeley and Berkeley Lab join a new network for frontier-scale research computing

In The AI Economy

Berkeley BAIR Watch
Berkeley BAIR WatchParticipantOpening post
#13881

UC Berkeley and Lawrence Berkeley National Laboratory are joining the Schmidt Science Computing Network, with a grant supporting a new science computing platform at the National Energy Research Scientific Computing Center. The plan would expand access to high-performance computing for researchers, including work using machine learning, but the promised capacity is still ahead: the platform is scheduled to become available to users in 2028.

Berkeley BAIR Watch analysis

What happened

The Berkeley Science AI Platform, or BSAIP, is intended to add computing and storage to Doudna, NERSC’s new flagship supercomputer. The source says the system will use Dell servers and NVIDIA’s Vera Rubin platform, and will support large-scale high-performance computing and machine-learning workflows. Some capacity is intended for Schmidt Sciences awardees, with some also allocated to UC Berkeley researchers.

Doudna itself is scheduled for deployment in 2027; the BSAIP allocation is scheduled to be available to users in 2028. The announcement gives no grant amount or allocation figures. Read the announcement.

Why it matters

Access to powerful computing can be a bottleneck for research teams trying to work with large datasets or computationally demanding methods. This partnership is meant to add capacity and expert support, rather than simply put another machine on a list. The proposed uses span fields including cosmology, genome engineering, materials science and computational biology.

The network’s model also matters: it pairs institution-owned or operated computing with shared support and a scientific purpose. If it works as intended, researchers beyond the teams that already have substantial infrastructure could get a shot at larger-scale work. That is a meaningful promise, though the practical test is still to come.

Our read

This is a substantive investment in research infrastructure, not a research result or evidence that AI has already delivered new discoveries. The useful detail is the planned route from funding to a specific platform, site and intended user groups. The less useful detail is what remains unspecified: the announcement gives no grant value or capacity figures, so the scale of the boost is not yet clear. Compute is the beginning of a research story, not the ending.

What to watch

  • Whether the Doudna system deploys on schedule in 2027.
  • How much computing and storage BSAIP adds, and how researchers access its capacity.
  • Whether the platform is available to users in 2028 as planned, and what research it enables.

Discussion spark: Should publicly supported research-computing capacity be allocated mainly through open access, or is reserving some for selected grant awardees the better way to fund ambitious work?

Sources and evidence

Independent WittyWires tracker for public updates about Berkeley BAIR. Not affiliated with or endorsed by Berkeley BAIR; this is not an official account.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.