Pier LaFarge

832 posts

Pier LaFarge

Pier LaFarge

@plafarge

Founder, Sparkfund @sparkfunder

Romney, WV | Birmingham, AL Katılım Aralık 2009
257 Takip Edilen405 Takipçiler
Pier LaFarge retweetledi
Ramez Naam
Ramez Naam@ramez·
There's a key point here many are missing on data center and AI energy use: It's Time to Power, and not Cost of Power, that's the key variable for data center buildout right now. Energy is a small fraction of overall compute cost. While time is everything in the AI race.
Shanu Mathew@ShanuMathew93

Loose thoughts pre take off. The one thing missing from the whole 100% clean vs not holy war going on right now is a debate about i) priorities and ii) bottlenecks. Put aside theoretical preferred options. IF you believe hyperscalers and other invested parties are attempting to build as many data centers or very large clusters, they are spending $ billions of capex (depreciating GPUs cost > land, electrical, thermal, EPC), THEN Time to power is the most determinant factor in your decisions. While decarbonization is a serious goal for most hyperscalers, the race to win AI mandates they’ll pursue the fastest options to power. Gas - 2-3 year lead time on large turbines. Solar - 5 year timeline for interconnection queues Gensets (smaller turbines, reciprocating engines) - Multi-year timeline from what I hear? Islanded/off grid options may be feasible for a select few players (not everyone will do this; my bet is very little tbh but an aside) but if you’re going to spend all that $ and go to all that trouble, you’re going to go with the fastest possible option and be relatively risk averse b/c you want to put the chips to work and faster you’re uptime, the faster the math pencils. Seems like optimal spots are old industrial sites cause you already have infra where you needed (factors include fiber, proximity to gas pipe, transmission hookups, etc.) Of the above options, nat gas has at least a demonstrated ability to scale thru bottlenecks and provide firm baseload power. It becomes the preferred option - mostly out of limited evidence to suggest there could be viable alternatives. So given all that, any options you provide that offer competitive alternatives - speed, cost, ability to handle load, may displace additional gas usage. That is a net climate benefit despite it still relying on gas partially or because it’s displacing additional gas molecules that would otherwise be called upon. If you have a problem with that predicament, then I’d hate to point out the fact that natural gas play a huge role in current electricity markets and fossil fuels in our overall energy/heat usage across the country. Each incremental leg matters.

English
4
9
85
14.9K
Pier LaFarge
Pier LaFarge@plafarge·
@drvolts @duncancampbell @duncancampbell the offgrid.ai paper y'all just put out is amazing work, read every word this morning. I see it as very much a complementary line of thinking to the DCP approach. the reality is we'll need a lot of both to meet challenge of speed and scale.
English
0
1
4
1.7K
Pier LaFarge
Pier LaFarge@plafarge·
Looking forward to #Deploy24 and discussing the Distributed Capacity Procurement model. Let's take distributed energy to grid scale with utility-led planning, siting, dispatch & deployment — with systems designed, built & installed by competitive local ecosystems of DER vendors.
English
0
0
3
324
Pier LaFarge
Pier LaFarge@plafarge·
As for the second, giving utilities the role of planning what assets they need and where they need them most is actually their core role. The DCP model also includes competition in its design by tapping into an ecosystem of local vendors to deliver, build and maintain the DERs.
English
1
0
0
86
Pier LaFarge
Pier LaFarge@plafarge·
Was asked about counterarguments to Distributed Capacity Procurement: 1️⃣ Utility-scale generation is cheaper per MWh than distributed generation. 2️⃣ Utility-led DER deployment gives a regulated monopoly a powerful new role.
English
1
0
1
104
Pier LaFarge
Pier LaFarge@plafarge·
As Ryan Long from @XcelEnergyMN put it, the DCP model brings together developer expertise, utility planning, low-cost capital and accessible space to host the asset, all for more strategic, faster and scaled #DER deployment. Let’s create value at grid scale. #REplus24
English
1
0
0
42
Pier LaFarge
Pier LaFarge@plafarge·
Had a great time discussing distributed capacity procurement at #REplus24. A theme of collaboration across utilities, building owners and DER companies stood out.
Pier LaFarge tweet mediaPier LaFarge tweet mediaPier LaFarge tweet media
English
2
0
1
60
Pier LaFarge
Pier LaFarge@plafarge·
I joined @EngelsAngle on the 'Factor This' podcast to discuss Distributed Capacity Procurement — a new model that's accelerating DER deployment in utility territories. It's something we're laser focused on at @Sparkfund. Check it out: youtube.com/watch?v=W3tB-J…
YouTube video
YouTube
English
0
0
0
200
Pier LaFarge
Pier LaFarge@plafarge·
At #REplus24, I'm joining @JigarShahDC, Ryan Long, and @InfraJosephine to discuss a new utility-led DER deployment model — Distributed Capacity Procurement — that builds on the success of VPPs to deploy capacity faster and at scale. #DERs #VPPs
Pier LaFarge tweet media
English
0
0
2
53
Pier LaFarge
Pier LaFarge@plafarge·
Today we’re proud to announce that Sparkfund has inked a historic partnership with Shell to offer a new bundled energy solution to commercial customers, Shell Energy Inside. This new approach has the potential to redefine how comm…lnkd.in/e2RTzjH lnkd.in/eeUr7NV
English
0
1
4
0
Pier LaFarge
Pier LaFarge@plafarge·
Morgan Simon, a dear friend and investor in Sparkfund, released REAL IMPACT yesterday - a book that addresses impact…lnkd.in/eu6pjRQ
English
0
1
4
0
Pier LaFarge
Pier LaFarge@plafarge·
Companies are re-thinking how they purchase energy technology and Sparkfund made a video about it! lnkd.in/emj4DrC
English
0
1
0
0